JP5840101B2 - Encoding method, encoding apparatus, decoding method, decoding apparatus, program, and recording medium - Google Patents

Description

  The present invention relates to a technique for encoding an audio signal such as voice or music with a small amount of information, and more particularly to an encoding technique for improving quantization accuracy.

  At present, as a technique for efficiently coding a digital input signal obtained by discretizing an acoustic signal such as voice or music, for example, an input signal in each interval (frame) at a fixed interval of about 5 to 200 ms included in the input signal. It is known to encode a frequency domain signal obtained by applying time-frequency conversion to an input signal sequence of one frame with a sequence as a processing target. Among such conventional technologies, an outline of an encoding device and a decoding device disclosed in Non-Patent Document 1 is shown in FIG.

  According to Non-Patent Document 1, the quantization value of the global gain (gain that affects the quantization accuracy of the normalized input signal sequence) is calculated in the time domain. However, since the energy of the signal in the time domain is equal to the energy of the signal in the frequency domain, even if the quantized value of the global gain is obtained in the frequency domain, this result is not different from that in the time domain. Therefore, here, a case where the quantized value of the global gain and the decoded value thereof are calculated in the frequency domain is illustrated.

  Hereinafter, processing in the encoding device will be described.

<Frequency domain conversion unit 101>
The frequency domain transform unit 101 receives an input time domain signal sequence x _F (n) in units of frames, which is composed of a plurality of consecutive samples included in the time domain input signal x (n). The frequency domain transform unit 101 converts the frequency component of L points (L is a positive integer, for example, 256) corresponding to the input time domain signal sequence x _F (n) of one frame to the input frequency domain signal sequence X (ω). Output as [ω∈ {0,..., L-1}]. Here, n represents an index of discrete time, and ω represents an index of discrete frequency. As a time-frequency conversion method, for example, MDCT (Modified Discrete Cosine Transform) or DCT (Discrete Cosine Transform) can be used.

<Normalization unit 102>
The normalization unit 102 includes an input frequency domain signal sequence X (ω) [ω∈ {0,..., L-1}] and an input frequency domain signal sequence X (ω) obtained by the gain control unit 104 described later. A gain (hereinafter referred to as a global gain) g that determines the quantization accuracy of each component of [ω∈ {0,..., L−1}] is input. The normalization unit 102 divides each component of the input frequency domain signal sequence X (ω) [ω∈ {0,..., L-1}] by the global gain g, or the input frequency domain signal sequence X ( ω) [ω∈ {0,..., L-1}] by multiplying each component of [ω∈ {0,..., L-1}] by the reciprocal of the global gain g, respectively. }] Is normalized, and a normalized signal sequence X _Q (ω) [ω∈ {0,..., L−1}] is output.

<Quantization unit 103>
The quantizing unit 103 receives the normalized signal sequence X _Q (ω) [ω∈ {0,..., L−1}]. The quantization unit 103 quantizes the normalized signal sequence X _Q (ω) [ω∈ {0,..., L-1}] by a predetermined method, and the normalized signal sequence X _Q ( ω) Quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1] which is a series of quantized values of each component of [ω∈ {0,. }], And a normalized signal code that is a code corresponding to the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}], and the bits of the normalized signal code Number (hereinafter referred to as the number of consumed bits). Further, when receiving from the gain control unit 104 command information for outputting a quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] and a normalized signal code Then, the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] and the normalized signal code are output.

<Gain control unit 104>
The gain control unit 104 receives the number of consumed bits. The gain control unit 104 adjusts the global gain g so that the number of consumed bits approaches a maximum value that is less than or equal to the number of bits allocated in advance to the normalized signal code (hereinafter referred to as the specified number of bits). The global gain g is output as a new global gain g. As an example of the adjustment of the global gain g, a process of increasing the global gain g when the number of consumed bits is larger than the specified number of bits and decreasing the global gain g otherwise can be exemplified. When the number of consumed bits reaches the maximum value less than the specified number of bits, the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0, ..., L-1}] and the normalized signal code are Command information to be output is output to the quantization unit 103.

<Global Gain Encoding Unit 105>
The global gain encoding unit 105 includes an input frequency domain signal sequence X (ω) [ωε {0,..., L-1}] and a quantized normalized signal sequence X ^ _Q (ω) [ωε {0. ,..., L-1}] is input. The global gain encoding unit 105 includes an input frequency domain signal sequence X (ω) [ω∈ {0,..., L-1}] and a quantum among a plurality of preset global gain quantization values. Correlation between each component of the normalized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] and the global gain quantized value candidate sequence The code corresponding to the global gain quantized value g ^ having the maximum or minimum error is output as the global gain code.

  The normalized signal code and the global gain code, which are output codes of the encoding device, are transmitted to the decoding device and input to the decoding device.

  Hereinafter, processing in the decoding apparatus will be described.

<Global Gain Decoding Unit 106>
A global gain code is input to the global gain decoding unit 106. The global gain decoding unit 106 applies a decoding process corresponding to the encoding process performed by the global gain encoding unit 105 to decode the global gain code, and outputs a decoded global gain g ^.

<Normalized signal decoding unit 107>
A normalized signal code is input to the normalized signal decoding unit 107. The normalized signal decoding unit 107 applies a decoding method corresponding to the encoding method performed by the quantization unit 103 of the encoding device, decodes the normalized signal code, and generates a decoded normalized signal sequence X ^ _Q ( ω) [ω∈ {0,..., L-1}] is output.

<Decoding Frequency Component Calculation Unit 108>
The decoded frequency component calculation unit 108 receives the decoded global gain g ^ and the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}]. The decoded frequency component calculation unit 108 is obtained by multiplying each component of the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] and the decoded global gain g ^. The sequence is output as a decoded frequency domain signal sequence X ^ (ω) [ωε {0,..., L−1}].

<Time domain conversion unit 109>
Decoded frequency domain signal sequence X ^ (ω) [ω∈ {0,..., L−1}] is input to time domain transform section 109. The time domain transform unit 109 applies a frequency-time transform to the decoded frequency domain signal sequence X ^ (ω) [ω∈ {0,..., L-1}], and outputs an output time domain signal sequence in units of frames. Output z _F (n). The frequency-time conversion method is an inverse conversion corresponding to the time-frequency conversion method used in the frequency domain conversion unit 101. In the above example, the frequency-time conversion method here is IMDCT (Inverse Modified Cosine Transform) or IDCT (Inverse Discrete Cosine Transform).

Guillaume Fuchs, Markus Multrus, Max Neuendorf and Ralf Geiger, “MDCT-BASED CODER FOR HIGHLY ADAPTIVE SPEECH AND AUDIO CODING,” 17th European Signal Processing Conference (EUSIPCO 2009), Glasgow, Scotland, August 24-28, 2009.

  In the coding method as described above, the global gain is adjusted to appropriately control the coarseness of quantization of the normalized signal sequence, so that the number of bits consumed, which is the code amount of the normalized signal code, is the specified number of bits. Control is performed so that the following maximum value is obtained. For this reason, when the number of bits consumed is smaller than the specified number of bits, there is a problem in that the encoding process that makes full use of the number of bits allocated in advance for the normalized signal sequence cannot be performed.

  In view of such circumstances, an object of the present invention is to provide an encoding technique that improves the quantization accuracy of a normalized signal sequence with a small increase in code amount and a decoding technique thereof.

  An encoding method according to an aspect of the present invention is an encoding method for encoding an input signal sequence in units of frames, which includes a plurality of input signal samples, wherein each input signal sample included in the input signal sequence is normalized. A normalized signal encoding step for obtaining a normalized signal code obtained by encoding a sequence based on the received signal and a quantized normalized signal sequence corresponding to the normalized signal code, and a gain corresponding to the input signal sequence A global gain encoding step for obtaining a quantized global gain and a global gain code corresponding to the quantized global gain, and a number of samples in which the magnitude of a sample of a quantized normalized signal sequence is greater than a predetermined value Alternatively, the magnitude of the sample value of the sequence obtained by normalizing the input signal samples corresponds to the number of samples larger than a predetermined value. Then, the quantized global normalization is obtained by correcting the value obtained by correcting the quantized global gain with the gain correction amount and correcting the value of each sample of the quantized normalized signal sequence or the quantized normalized signal sequence. A gain correction amount code generation step for generating a gain correction amount code for specifying a gain correction amount that minimizes an error between the signal sequence obtained by multiplying the value of each sample of the completed signal sequence and the input signal sequence; and Quantized and normalized signal sequence obtained by dividing each sample value of the input signal sequence by the gain obtained by correcting the quantized global gain or the quantized global gain by the gain correction amount corresponding to the gain correction amount code A control unit that preferentially performs one step of an error encoding step that generates an error code corresponding to an error with the signal sequence over the other step. .

  The decoding method according to one aspect of the present invention is a decoding method for obtaining an output signal sequence by decoding a frame-unit code, and obtaining a decoded normalized signal sequence by decoding a normalized signal code included in the code A decoding step; a global gain decoding step for obtaining a decoded global gain by decoding a global gain code included in the code; and a number of samples in which a sample value of a decoded normalized signal sequence is larger than a predetermined value. Accordingly, a decoding global gain correction step for correcting the decoding global gain based on the gain correction amount code included in the code, and a decoded normalized signal sequence for correcting the decoding normalized signal sequence based on the error code included in the code A control unit that preferentially performs one step of the correction step over the other step, and a corrected decoding group Has a Barugein, a decoding step of obtaining a signal sequence obtained by multiplying the samples of the modified decoded normalized signal sequence as an output signal sequence, the.

  Prioritizing either the correction of the quantized global gain or the correction of the quantized normalized signal sequence according to the characteristics of the signal sequence improves the quantization accuracy of the input signal with a small increase in code amount, Sound quality degradation caused by musical noise or quantization noise can be reduced.

The block diagram which shows the function structural example of the encoding apparatus and decoding apparatus in connection with a prior art. The block diagram which shows the function structural example of the encoding apparatus which concerns on 1st Embodiment. The figure which shows the processing flow of the encoding process which concerns on 1st Embodiment. The figure which shows the processing flow of the specific example 1 of the 1st example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 2 of the 1st example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 3 of the 1st example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 1 of the 3rd example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 2 of the 3rd example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 3 of the 3rd example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 1 of the 5th example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 2 of the 5th example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 3 of the 5th example of the division process by a 1st reference | standard. The figure which shows the processing flow of the specific example 1 of the 1st example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 2 of the 1st example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 3 of the 1st example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 1 of the 3rd example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 2 of the 3rd example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 3 of the 3rd example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 1 of the 5th example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 2 of the 5th example of the division process by a 2nd reference | standard. The figure which shows the processing flow of the specific example 3 of the 5th example of the division process by a 2nd reference | standard. The block diagram which shows the function structural example of the decoding apparatus which concerns on 1st Embodiment. The figure which shows the processing flow of the decoding process which concerns on 1st Embodiment. FIG. 3 is a block diagram illustrating an example of a functional configuration of a gain correction amount encoding unit 140. The figure for demonstrating the example of the range which put together the divided range and the divided range. The figure for demonstrating the example of a process of the control part 170 and the control part 205. FIG. The block diagram which shows the function structural example of the encoding apparatus which concerns on 2nd Embodiment. The block diagram which shows the function structural example of the decoding apparatus which concerns on 2nd Embodiment.

  Embodiments of the present invention will be described with reference to the drawings. The same components or the same processes may be assigned the same reference numerals and redundant description may be omitted. In addition, the acoustic signal handled in each embodiment is a signal such as a sound, a sound such as a musical sound, and a video. Here, it is assumed that the acoustic signal is a time domain signal. However, the time domain signal may be converted into a frequency domain signal or a frequency domain signal may be converted into a time domain signal by a known technique as necessary. You can also. Therefore, the signal to be encoded may be a time domain signal or a frequency domain signal (in the following description, the frequency domain signal is treated for the sake of concrete explanation). The signal input as the target of the encoding process is a sequence (sample sequence) composed of a plurality of samples, and the encoding process is normally executed in units of frames. I will call it.

For example, referring to the technique shown in FIG. 1, each component included in the input signal sequence X (ω) [ω∈ {0,..., L-1}], the quantized global gain g ^, and the quantized normalized signal The relationship between the components included in the sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] can be expressed by Expression (1). Where e _g is the quantization error between the global gain g and the quantized global gain g ^, and e _XQ is the normalized input signal sequence X _Q (ω) [ω∈ {0, ..., L-1}]. Quantization normalized signal sequence X ^ _Q (ω) represents a quantization error between corresponding components (components having the same value of ω) included in [ω∈ {0, ..., L-1}] .

In normal quantization, the number of bits consumed by a normalized signal code that is a code corresponding to a quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1}] is Depending on the input signal sequence, a part of the predetermined number of bits predetermined for the normalized signal code often remains as unused bits. Therefore, the excess one or more bits (hereinafter, referred to as surplus bits) utilizing the reduction of the quantization error e _g and e _XQ. Furthermore, not only the surplus bits, but also one or a plurality of bits prepared in advance for reducing the quantization error may be used. Hereinafter referred to available bits in reducing the quantization error e _g a "gain correction bits". _Let U _g be the number of gain correction bits. A bit that can be used to reduce the quantization error e _XQ is referred to as an “error expression bit”. _Let U _XQ be the number of error expression bits.

Incidentally, the "to reduce the quantization error e _g" is, in other words, "to correct the quantization global gain" especially none other. Regarding the correction of the quantized global gain, there is a method for correcting the quantized global gain common to the entire discrete frequency index ω∈ {0, 1, 2,..., L-1} of one frame, that is, the entire sequence. Conceivable. However, in consideration of the characteristics of the acoustic signal, the range of the entire sequence B is N (where N is a predetermined integer equal to or greater than 2) rather than correcting the quantization global gain common to the entire sequence { B _n } _{n = 1} ^N = {B ₁ ,..., B _n ,..., B _N }, and then obtaining the gain corresponding to each range by correcting the quantized global gain is better than the voice quality. Can be expected to improve. From this point of view, in the adaptive quantization in the embodiment, when correcting the quantized global gain, the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {0,..., L-1} ] Is divided into a plurality of ranges.

  An easily conceivable method for dividing the same signal sequence B into N ranges by the encoding device and the decoding device is to specify a range such as the boundary position of adjacent ranges and the number of components included in each range. In this method, information is output from the encoding device. However, a large number of bits are required to output information specifying the range. The coding apparatus and the decoding apparatus perform classification according to the same standard without using the information specifying the range as the output of the coding apparatus, that is, without consuming bits. Further, assuming that the gain correction bits, that is, the amount of information for correcting the quantized global gain, are given to each range as evenly as possible, the components of the quantized normalized signal sequence included in each range It is desirable that the amount of information be as uniform as possible. Therefore, as a criterion for series division, “criteria for classifying so that the energy in each range is as equal as possible” or “criteria for classifying so that the number of significant samples included in each range is as equal as possible” is adopted. Specific classification methods based on these criteria will be described in detail later.

  Details of the embodiment will be described below.

<< First Embodiment >>
The encoding apparatus 1 (see FIG. 2) according to the first embodiment includes a normalized signal encoding unit 120, a global gain encoding unit 105, a gain correction amount encoding unit 140, a sorting unit 150, a control unit 170, and error encoding. Part 150 and adding part 190. The encoding device 1 may include a frequency domain transform unit 101 and a synthesis unit 160 as necessary.

  First, an encoding process performed by the encoding device 1 (encoder) will be described (see FIG. 3).

Here, the input signal sequence of the encoding device 1 is an input signal sequence X (ω that is a frequency component of L points (L is a positive integer, for example, 256) corresponding to the acoustic signal x (n) in units of frames. ) [Ω∈ {L _min ,..., L _max }] Here, n is an index of discrete time, ω is an index of discrete frequency, L _min is an index of minimum discrete frequency among frequency components at L point, and L _max is a maximum discrete frequency among frequency components at L point. Represents the index. However, the acoustic signal x (n) in units of frames may be used as the input signal sequence of the encoding device 1, or a residual signal obtained by performing linear prediction analysis on the acoustic signal x (n) in units of frames is encoded. 1 may be used as the input signal sequence, or a frequency component at L point (L is a positive integer, for example, 256) corresponding to the residual signal may be used as the input signal sequence.

<Frequency domain conversion unit 101>
The encoding device 1 may include a frequency domain transform unit 101 as a preprocessing unit of the encoding device 1 or in the encoding device 1. In this case, the frequency domain transform unit 101 generates a frequency component of L points (L is a positive integer, for example, 256) corresponding to the acoustic signal x (n) in the time domain in units of frames, and the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] As the time-frequency conversion method, for example, MDCT (Modified Discrete Cosine Transform) or DCT (Discrete Cosine Transform) can be used. In this case as well, instead of the time domain acoustic signal in units of frames, a residual signal obtained by linear prediction analysis of the time domain acoustic signals in units of frames may be x (n).

<Normalized signal encoding unit 120>
The normalized signal encoding unit 120 includes a sequence (normalized signal sequence X) of signals obtained by normalizing each component of the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] in units of frames. _Q (ω) [ω∈ {L _min ,..., L _max }]) is encoded, and a normalized normalized signal sequence X ^ _Q (ω corresponding to the normalized signal code is obtained. ) [Ω∈ {L _min ,..., L _max }] is output (step S1e).

  When the number U of surplus bits is obtained from the normalized signal encoding unit 120, the obtained number U of surplus bits is transmitted to the control unit 170.

  The normalized signal encoding unit 120 is realized by, for example, the normalization unit 102, the quantization unit 103, and the gain control unit 104 in FIG. Each of the normalization unit 102, the quantization unit 103, and the gain control unit 104 operates as described in the [Background Art] column.

<Global Gain Encoding Unit 105>
The global gain encoding unit 105 supports a quantized global gain g ^ that is a gain corresponding to the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] and a quantized global gain g ^. The global gain code to be obtained is obtained (step S2e). The global gain encoding unit 105 also obtains a quantization step width corresponding to the quantized global gain g ^ as necessary.

  The global gain encoding unit 105 operates, for example, as described in the “Background art” column.

  In addition, for example, the global gain encoding unit 105 includes a table storing a plurality of sets of quantized global gain candidates and global gain codes corresponding to the candidates, and the global gain obtained by the normalized signal encoding unit 120 The candidate for the quantized global gain closest to the gain g may be set as the quantized global gain g ^, and the global gain code corresponding to the candidate may be output.

In short, the global gain coding unit 105 multiplies each component of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] and the gain and obtains a signal. Quantized global gain g ^ obtained on the basis of the maximum correlation or minimum error between the sequence and the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] and this quantization A global gain code corresponding to the global gain may be obtained and output.

  When the gain correction amount encoding unit 140 performs processing using the quantization step width corresponding to the quantized global gain 量子, the quantization step width corresponding to the quantized global gain ＾ is also the gain correction amount code. Is output to the conversion unit 140.

<Control unit 170>
The control unit 170 responds to the number of samples in which the value of the sample of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is larger than a predetermined value. Thus, one of the process of generating the gain correction amount code and the process of generating the error code is preferentially performed over the other process (step S3e). The predetermined value is 0, for example. Priority information (for example, U _XQ , U _g , which will be described later), which is information regarding which process is preferentially performed, is transmitted to the error encoding unit 180 and the gain correction amount encoding unit 140.

As an example, the error code is a signal obtained by dividing each sample value of the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] by the quantized global gain g ^ as described later. This is a code corresponding to an error between the sequence and the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. In this example, the process of generating an error code is the process of step S4e by an error encoding unit 180 described later.

In this example, as will be described later, the gain correction amount code is a corrected quantization obtained by correcting a quantized normalized signal sequence with a gain obtained by correcting the quantized global gain g ^ with a gain correction amount. The signal sequence obtained by multiplying the value of each sample of the normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] and the input signal series X (ω) [ω∈ { L _min ,..., L _max }] is a code for specifying a gain correction amount that minimizes an error. In this example, the process of generating the gain correction amount code is the process of steps S6e and S7e by the sorting unit 150 and the gain correction amount encoding unit 140 described later. Note that the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is modified to provide a corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω The process of obtaining ε {L _min ,..., L _max }] is the process of step S5e by the adder 190 described later.

As another example, the gain correction amount code is the quantized global gain g ^ for each sample of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min , ..., L _max }]. A code for specifying a gain correction amount that minimizes an error between the signal sequence obtained by multiplying the value and the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] may be used. In this example, the process of generating the gain correction amount code is the “modified post-correction normalized signal sequence X ^ ′ _Q (ω in steps S6e and S7e by the segmentation unit 150 and the gain correction amount encoding unit 140 described later. ) ”Is replaced with“ quantized normalized signal sequence X ^ _Q (ω) ”.

In this example, the error code is a gain obtained by correcting each sample value of the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] with a quantized global gain g ^ using a gain correction amount. May be a code corresponding to an error between the signal sequence obtained by dividing by and the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. In this example, the process for generating the error code is obtained by correcting “quantized global gain g ^” in step S4e by “error quantizing unit 180, which will be described later”, with “quantized global gain g ^ being corrected by the gain correction amount. It is the processing replaced with “Gain”.

The number of samples in which the value of the sample of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is larger than a predetermined value is a so-called quantization normal. This is the number of non-zero samples of the digitized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. A non-zero sample of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is also referred to as a non-zero coefficient.

  Hereinafter, an example of processing of the control unit 170 will be described with reference to FIG.

In the example of FIG. 26, the control unit 170 sets ω = L _min and count = 0, and initializes the variable ω and the variable count (step S3e1).

The controller 170 determines whether ω is equal to or less than L _max (step S3e2). If ω is equal to or less than L _max , the process proceeds to step S3e3. If ω is not less than or equal to L _max , the process proceeds to step S3e6.

The controller 170 determines whether the value of X ^ _Q (ω) is larger than a predetermined value αg ^ (step S3e3). α is a predetermined real number of 0 or more. For example, α = 0. In the example of FIG. 26, it is determined whether the value of X ^ _Q (ω) is larger than a predetermined value αg ^ by determining whether the value of | X ^ _Q (ω) | ² is larger than αg ^. ing. Of course, this is only an example, and other determination methods may be used as long as it can be determined whether the magnitude of X ^ _Q (ω) is larger than a predetermined value αg ^.

If the value of | X ^ _Q (ω) | ² is larger than αg ^, the control unit 170 increments the variable count by 1 (step S3e4) and increments the variable ω by 1 (step S3e5). Then, it returns to step S3e1.

If the value of | X ^ _Q (ω) | ² is not larger than αg ^, the control unit 170 increments the variable ω by 1 (step S3e5). Then, it returns to step S3e1.

For example, in this way, the control unit 170 determines that the magnitude of the sample value of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is larger than a predetermined value. First count the number of large samples.

  The control unit 170 determines whether the variable count is larger than Th (step S3e6). Th is a predetermined positive integer equal to or less than the L point.

If the variable count is larger than Th, the control unit 170 assigns a predetermined number of bits among the U-bit surplus bits to the gain correction bits (step S3e7). For example, U _yes is a predetermined number, and U _g = min (U _yes , U). min is a function for selecting the minimum value. Therefore, U _g is the smaller of U _yes and U. U _yes is 6, for example. Thereafter, the control unit 170 determines whether there are extra bits, in other words, whether UU _g is greater than 0 (step S3e9). If there are more bits, the control unit 170 assigns the remaining bits to the error expression bits (step S3e10). For example, U _XQ = UU _g . If there are no more bits, the process in step S3e ends.

Note that C ₁ may be a predetermined constant such as 0.75, floor may be a floor function, and U _yes = floor (C ₁ U). Also, the C ₂ to a predetermined constant such as 0.25, the ceil as a ceiling function may be _{U yes = ceil (C 1 U} ).

If the variable count is not greater than Th, the control unit 170 assigns a predetermined number of bits among the U-bit surplus bits to the error expression bits (step S3e7). For example, _let U _no be U and U _XQ = min (U _no , U). Thereafter, the control unit 170 determines whether there are extra bits, in other words, whether UU _XQ is greater than 0 (step S3e11). If there are more bits, the controller 170 assigns the remaining bits to the gain correction bits (step S3e12). For example, U _g = UU _XQ . If there are no more bits, the process in step S3e ends. U _no may be a predetermined value other than U.

U _XQ is transmitted to the error encoding unit 180, and U _g is transmitted to the gain correction amount encoding unit 140.

  In this way, the control unit 170 allocates a predetermined number of bits among the surplus bits to the code generated in the step that is preferentially performed, and the surplus is added to the code generated in the step that is not preferentially performed. Of the bits, bits other than the assigned bits are assigned. Thereby, an efficient quantization error can be corrected.

If the number of non-zero samples of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min , ..., L _max }] is small, quantization by correcting the quantized global gain g ^ The contribution to accuracy improvement is small. This is because the number of samples of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] affected by the correction of the quantized global gain g ^ is small. Therefore, when the number of non-zero samples of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min , ..., L _max }] is small, a gain correction amount code is generated. Giving priority to the process of generating an error code over the process of making a greater contribution to the improvement of quantization accuracy.

<Error encoding unit 180>
The error encoding unit 180 uses the error code bit of U _XQ bits to correct an error for correcting the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. A code and a decoded quantization error q (ω) corresponding to the error code are generated (step S4e).

The input signal sequence X (ω) [ω∈ {L _min ,..., L _max }], the quantized global gain g ^, and the quantized normalized signal series X ^ _Q (ω) [ω∈ {L _min , .., L _max }] is input to the error encoding unit 180.

First, the error encoding unit 180 calculates a quantization error r (ω) = X (ω) / g ^ −X ^ _Q (ω) [ω∈ {L _min ,..., L _max }].

  Since the quantization error r (ω) is a rounding error due to quantization, it is distributed almost uniformly from −0.5 to +0.5. In order to encode all samples with an arbitrary number of bits, the encoding method and the rule of the position of the target sample are determined by the error expression bits.

  Next, the error encoding unit 180 encodes the sequence of quantization errors r (ω) with the number of error expression bits, and obtains the obtained error code and the decoded quantization error q (ω) corresponding to the error code. Output (step S4e). This error code is transmitted to the synthesis unit 160. In addition, error encoding section 180 transmits decoded quantization error q (ω) to adding section 190.

<Specific example 1 of error encoding>
When the quantization error value is encoded, a plurality of samples may be collected and vector quantization may be performed. However, generally, a code sequence is stored in a table (codebook), and a distance calculation between the input and the code sequence is necessary, which increases the amount of memory and the amount of calculation. Also, the configuration becomes complicated, for example, an individual codebook is required to cope with an arbitrary number of bits.

  The operation of specific example 1 is as follows.

  A code book for each possible value of the error expression bit number is stored in advance in the code book storage unit in the error coding unit 180. Each codebook stores in advance a vector having the same number of samples as the number of quantization error sequences that can be expressed by the number of error expression bits corresponding to each codebook, and a code corresponding to the vector. ing.

  Error coding section 180 selects a codebook corresponding to the number of error expression bits determined by control section 170 from the codebook stored in the codebook storage section, and performs vector quantization using the selected codebook. . The encoding process after selecting a codebook is the same as general vector quantization. That is, a vector that minimizes the distance between each vector of the selected codebook and the input quantization error sequence or maximizes the correlation between them is output as a sequence of decoded quantization error q (ω) At the same time, a code corresponding to this vector is output as an error code.

  In the above description, the vector stored in the codebook has the same number of samples as the quantization error sequence, but the number of vector samples stored in the codebook is 1 / integer of the quantization error sequence. The quantization error sequence may be vector quantized for each of a plurality of portions, and the obtained plurality of codes may be error codes, and the vector corresponding to each code may be a sequence of the decoded quantization error q (ω) for each portion. .

<Specific Example 2 of Error Encoding Unit 180>
When encoding the quantization error included in the quantization error sequence one sample at a time, the priority order of the quantization error samples included in the quantization error sequence is determined, and the error from the quantization error sample having the higher priority is determined. Only those that can be encoded with the number of expression bits are encoded. For example, encoding is performed preferentially from quantization error samples having a large absolute value or energy of quantization error.

  When determining the priority order, for example, the power spectrum envelope value can be referred to. Of course, as with the power spectrum envelope value, an approximate value of the power spectrum envelope value, an estimated value of the power spectrum envelope value, a value obtained by smoothing any of these values in the frequency direction, and a plurality of samples of any of these values Although an average value of the above and a value having the same magnitude relationship as at least one of these values can be referred to, only the case where the power spectrum envelope value is used will be described below. When the input signal sequence X (ω) is a very small value, that is, when the input signal sequence X (ω) is less than half the step width, the result of dividing the input signal sequence X (ω) by the quantized global gain g ^ is At 0, the quantization error r (ω) is also much smaller than 0.5. That is, when the power spectrum envelope value is small to some extent, the effect on the auditory quality is small when the quantization error r (ω) is encoded in addition to the input signal sequence X (ω). May be excluded from the encoding target. If the power spectrum envelope value is large to some extent, it is not known which quantization error of the sample is large. For example, the order of the position of the original sample on the frequency axis in ascending order (in order of decreasing frequency) or the power spectrum envelope The quantization error sample r (ω) is encoded by 1 bit for each error expression bit in descending order. Moreover, it is sufficient to exclude the case where the power spectrum envelope value is below a certain level.

  Note that, as the power spectrum envelope value, a value obtained by denormalizing the normalized power spectrum envelope value calculated inside the quantization unit 103 using the quantized global gain g ^ obtained from the global gain encoding unit 105 Can be used. Of course, values obtained by other methods may be used. For example, the power spectrum envelope value is a value representing a power spectrum of a quantized linear prediction coefficient obtained by quantizing a linear prediction coefficient obtained by linear prediction analysis of an input time domain acoustic signal x (n). Also good. In this case, the code of the quantized linear prediction coefficient is transmitted to the decoding device 2.

The normalized power spectrum envelope value is a parameter that represents the outline of the normalized signal sequence X _Q (ω). In general, the quantizing unit 103 divides the normalized signal sequence X _Q (ω) into a rough shape (normalized power spectrum envelope value) and a fine structure from the viewpoint of quantization efficiency, and sets each parameter. Are encoded. The normalized signal code output from the quantization unit 103 includes these encoded parameters. The normalized power spectrum envelope value is calculated from the energy of each subband by dividing the normalized input signal sequence X _Q (ω) by the number of subbands that are predetermined positive integers such as 8 and 16. .

In encoding a quantization error sequence, a value r (ω) = x of a certain quantization error sample is set, and distortion due to the quantization is E = ∫ ₀ ^0.5 f (x) (x−μ) ² dx. Here, f (x) is a probability distribution function, and μ is an absolute value of the reconstruction value. In order to minimize the distortion E due to quantization, μ may be determined so that dE / dμ = 0. That is, μ may be the center of gravity of the probability distribution of the quantization error r (ω).

If the quantized normalized signal sequence X ^ _Q (ω) is divided by the quantized global gain g ^ and rounded, that is, if the corresponding quantized normalized signal sequence X ^ _Q (ω) is not zero The distribution of the quantization error r (ω) is almost uniform, and μ = 0.25.

If the quantized normalized signal sequence X ^ _Q (ω) is divided by the quantized global gain g ^ and rounded, that is, if the corresponding quantized normalized signal sequence X ^ _Q (ω) is zero, Since the distribution of the quantization error r (ω) tends to concentrate on 0, it is necessary to use the center of gravity of the distribution as the value of μ.

In this case, for each of a plurality of quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) is not 0, a quantization error sample to be encoded is selected and selected. The position of the quantized error sample in the plurality of quantized error samples and the value of the selected quantized error sample may be encoded and transmitted to the decoding device 2 as an error code. For example, select the quantization error sample with the largest absolute value among the four quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) is not 0, and select The quantized error sample value is quantized (for example, whether it is + or-) and the information is sent in 1 bit, and the position of the selected quantization error sample is sent in 2 bits. The decoded value of the selected quantization error sample is a value obtained by adding + or − to μ. Since the quantization error sample that has not been selected is not sent to the decoding device 2, the decoded value is set to zero. In general, q bits are required to inform the decoding device of which position of 2 ^q samples is selected.

  At this time, the value of the center of gravity of the distribution of only the sample having the largest absolute value of the quantization error value may be used as μ.

The number of error expression bits is U _XQ , and the number of quantization error samples constituting the quantization error sequence, the value of the corresponding quantized normalized signal sequence X ^ _Q (ω) is not zero, When the number is T and the number of quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) is 0 is S, encoding can be performed in the following procedure. .

(A) In the case of U _XQ ≦ T, the error encoding unit 180 includes T quantization error samples in which the value of the corresponding quantized normalized signal sequence X ^ _Q (ω) in the quantization error sequence is not 0. Among the selected quantization error samples, U _XQ are selected from the corresponding power spectrum envelope values that are large, and a 1-bit code that is information representing the positive / negative of the quantization error samples is generated. Then, the generated U _XQ bit code is output as an error code. If the corresponding power spectrum envelope values are the same, for example, select according to a predetermined rule such as selecting a quantization error sample with a smaller position on the frequency axis (a quantization error sample with a lower frequency). To do.

  Also, the error encoding unit 180 gives information indicating the sign of each quantization error sample to the absolute value 0.25 of the reconstruction value for each selected quantization error sample, and gives the reconstruction value +0.25 or -0.25. To obtain a decoded quantization error q (ω) corresponding to each quantization error sample.

  Further, error encoding section 180 outputs 0 as a decoded quantization error q (ω) corresponding to each quantization error sample for each quantization error sample that has not been selected.

(B) When T <U _XQ ≦ T + S The error encoding unit 180 includes T number of non-zero values of the corresponding quantized normalized signal sequence X ^ _Q (ω) in the quantization error sequence. For each quantization error sample, a 1-bit code that is information representing the sign of the quantization error sample is generated.

The error encoding unit 180 also encodes quantization error samples in which the value of the corresponding quantized normalized signal sequence X ^ _Q (ω) in the quantization error sequence is 0 with U _XQ -T bits. . When there are a plurality of quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) has a value of 0, encoding is performed with priority from the corresponding power spectrum envelope value. Specifically, from the quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) has a value of 0, the corresponding U _XQ -T number from the largest power spectrum envelope value For 1, a 1-bit code representing the sign of the quantization error sample is generated. Alternatively, a plurality of quantization error samples having a corresponding power spectrum envelope value out of quantization error samples whose corresponding quantized normalized signal sequence X ^ _Q (ω) is 0 are extracted, and a plurality of quantizations are obtained. Vector quantization is performed for each error sample to generate a U _XQ -T bit code. If the corresponding power spectrum envelope values are the same, for example, select according to a predetermined rule such as selecting a quantization error sample with a smaller position on the frequency axis (a quantization error sample with a lower frequency). To do.

The error encoding unit 180 further outputs a combination of the generated U _XQ bit code and the U _XQ -T bit code as an error code.

Further, the error encoding unit 180, for each quantization error sample for which the value of the quantized normalized signal sequence X ^ _Q (ω) is not 0, information indicating the sign of each quantization error sample is represented as a reconstructed value. An absolute value of 0.25 is given to obtain a reconstructed value +0.25 or -0.25, which is output as a decoded quantization error q (ω) corresponding to each quantization error sample.

Also, the error encoding unit 180 selects U _XQ −T from the quantized error samples whose quantized normalized signal sequence X ^ _Q (ω) has a value of 0 corresponding to a large power spectrum envelope value. For each quantization error sample of, information representing the sign of each quantization error sample is given to the reconstruction value A (0 <A <0.25) to obtain the reconstruction value + A or -A, and each quantization error The decoded quantization error q (ω) corresponding to the error sample is output. Alternatively, the error encoding unit 180 may perform quantization from a quantization error sample whose quantization normalized signal sequence X ^ _Q (ω) has a value of 0 to a corresponding one having a large power spectrum envelope value. For an error sample, a sequence of decoded quantization error sample values corresponding to the U _XQ -T bit code obtained by vector quantization included in the error code is obtained, and decoding corresponding to the plurality of quantization error samples is obtained. Output as quantization error q (ω).

  Further, the error encoding unit 180 outputs 0 as a decoded quantization error q (ω) corresponding to each quantization error sample for each quantization error sample that has not been encoded.

(C) In the case of T + S <U _XQ The error encoding unit 180 generates a 1-bit first-round code representing the positive / negative of the quantization error sample for each of all the quantization error samples included in the quantization error sequence. To do.

In addition, the error encoding unit 180 further encodes the quantization error sample using the remaining U _XQ- (T + S) bits in the procedure (A) or (B). That is, U _XQ- (T + S) is set as a new U _XQ and the second round (A) is executed for the encoding error in the first round. That is, as a result, at least some quantization error samples are quantized by 2 bits per quantization error sample. In the first round coding, the quantization error r (ω) was uniform within the range of −0.5 to +0.5, but the first round error value to be coded in the second round Is in the range of -0.25 to +0.25.

Specifically, the error encoding unit 180 has a value of the corresponding quantized normalized signal sequence X ^ _Q (ω) that is not 0 among the quantization error samples constituting the quantization error sequence, and For a quantization error sample having a positive quantization error r (ω), a 1-bit value representing the positive or negative value is obtained by subtracting the reconstruction value 0.25 from the quantization error sample value. A second round code is generated.

Further, the error encoding unit 180 has a value of the corresponding quantized normalized signal sequence X ^ _Q (ω) that is not 0 among the error samples constituting the quantization error sequence, and the quantization error r ( For a quantization error sample with a negative value of ω), a 1-bit second-round code representing the positive or negative value is obtained by subtracting the reconstruction value −0.25 from the quantization error sample value. Is generated.

Further, the error encoding unit 180 has a value of the corresponding quantized normalized signal sequence X ^ _Q (ω) among error samples constituting the quantization error sequence, and a quantization error r ( A quantization error sample having a positive value of ω) is obtained by subtracting the reconstruction value A (A is a predetermined positive value smaller than 0.25) from the quantization error sample value. For the value, a 1-bit second-round code representing the sign is generated.

Further, the error encoding unit 180 has a value of the corresponding quantized normalized signal sequence X ^ _Q (ω) among error samples constituting the quantization error sequence, and a quantization error r ( A quantization error sample having a negative value of ω) is obtained by subtracting the reconstruction value −A (A is a predetermined positive value smaller than 0.25) from the quantization error sample value. For each value, a 1-bit second-round code representing the sign is generated.

  Furthermore, the error encoding unit 180 outputs a combination of the generated first and second cycle codes as an error code.

The error encoding unit 180 performs the following processing for each quantization error sample whose quantization normalized signal sequence X ^ _Q (ω) is not 0, and decodes the quantized quantum corresponding to each quantization error sample. Generate and output the error q (ω). First, information representing positive / negative corresponding to the first-round code is given to the absolute value 0.25 of the reconstructed value to obtain a reconstructed value +0.25 or −0.25, which is defined as a first-round decoded quantization error q ₁ (ω). Also, the reconstruction value +0.125 or −0.125 obtained by giving positive / negative information corresponding to the second round code to the absolute value 0.125 of the reconstruction value is set as the second round decoding quantization error q ₂ (ω). Then, the first round decoding quantization error q ₁ (ω) and the second round decoding quantization error q ₂ (ω) are added, and the decoding quantization error q (ω) corresponding to each quantization error sample is added. And

In addition, the error encoding unit 180 performs the following processing for each quantization error sample in which the value of the quantized normalized signal sequence X ^ _Q (ω) is 0, and corresponds to each quantization error sample. The decoding quantization error q (ω) to be generated is generated and output. First, positive / negative information corresponding to the first round code is given to the absolute value A (0 <A <0.25) of the reconstructed value to obtain the reconstructed value + A or -A, and the first round decoding quantization error q ₁ (ω). Also, positive / negative information corresponding to the second round code is given to the absolute value A / 2 of the reconstructed value to obtain the reconstructed value + A / 2 or -A / 2, and the second round decoding quantization error q ₂ ( ω). Then, the first round decoding quantization error q ₁ (ω) and the second round decoding quantization error q ₂ (ω) are added, and the decoding quantization error q (ω) corresponding to each quantization error sample is added. And If there is no second-round code corresponding to each quantization error sample, the first-round decoded quantization error q ₁ (ω) is converted to the decoded quantization error q (ω corresponding to each quantization error sample. ) Is output.
If not all of the T + S quantization error samples in the quantization error sequence are encoded, or the corresponding quantization normalized signal sequence X ^ _Q (ω) has a value of 0 If multiple samples are encoded together with 1 bit or less per sample, the quantization error sequence is encoded with fewer UU bits than U _XQ bits, so the condition of (C) is the case when T + S <UU. That's fine.

  Instead of the power spectrum envelope values (A) and (B) described above, an approximate value of the power spectrum envelope value or an estimated value of the power spectrum envelope value may be used.

  Further, instead of the power spectrum envelope values of (A) and (B) above, a value obtained by smoothing the power spectrum envelope value, the approximate value of the power spectrum envelope value, or the estimated value of the power spectrum envelope value in the frequency direction May be used. As a value obtained by smoothing, the weighted spectrum envelope coefficient obtained by the weighted envelope normalization unit 15 may be input to the error encoding unit 180 and used, or may be calculated by the error encoding unit 180.

Further, instead of the power spectrum envelope values (A) and (B) described above, a value obtained by averaging a plurality of power spectrum envelope values may be used. Instead of the power spectrum envelope value W (ω) [L _min ≦ ω ≦ L _max ], an average value of the approximate value of the power spectrum envelope value and an average value of the estimated value of the power spectrum envelope value may be used. Furthermore, an average value of values obtained by smoothing the power spectrum envelope value, the approximate value of the power spectrum envelope value, or the estimated value of the power spectrum envelope value in the frequency direction may be used. The average value here is a value obtained by averaging the target values for a plurality of samples, that is, a value obtained by averaging the target values for the plurality of samples.

Further, instead of the power spectrum envelope values of (A) and (B) above, a power spectrum envelope value, an approximate value of the power spectrum envelope value, an estimated value of the power spectrum envelope value, and any of these values A value having the same magnitude relationship as at least one of a value obtained by smoothing and a value obtained by averaging any one of these values for a plurality of samples may be used. In this case, the error encoding unit 180 calculates and uses values that have the same magnitude relationship. The value having the same magnitude relationship is a square value or a square root. For example, the power spectrum envelope value W (ω) [L _min ≦ ω ≦ L _max ] is the same value as the square of the power spectrum envelope value (W (ω)) ² [L _min ≦ ω ≦ L _max ] or the square root of the power spectrum envelope value (W (ω)) ^1/2 [L _min ≦ ω ≦ L _max ].

<Adding unit 190>
The adder 190 receives a sequence of decoded quantization errors q (ω) and a quantized normalized signal sequence X ^ _Q (ω). The adder 190 adds the sequence of the decoded quantization error q (ω) and the quantized normalized signal sequence X ^ _Q (ω), and the corrected quantized normalized signal sequence X ^ ′ _Q (ω ) (Step S5e). The post-correction quantization normalized signal sequence X ^ ′ _Q (ω) is transmitted to the segmentation unit 150 and the gain correction amount encoding unit 140.

<Division section 150>
The segmentation unit 150 classifies the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] so that the energy in each range is as equal as possible. Or “Criteria for classifying so that the number of significant samples included in each range is as equal as possible”, N ranges (where N = 2 ^D and D is a predetermined integer greater than or equal to 2 (Step S3e). When aligning the above description, it modified quantized normalized signal sequence _{X ^ 'Q (ω) [} ω∈ {L min, ..., L max}] {L min the set of discrete frequency index, ... , L _max }, the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] corresponds to the entire sequence B, and the segmentation unit 150 Quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] is included in each range as “criteria for classifying energy in each range to be as equal as possible” Is divided into N ranges {B _n } _{n = 1} ^N = {B ₁ ,..., B _n ,..., B _N }. Details of the sorting process for sorting by “criteria for sorting so that the energy of each range is as equal as possible” or “criteria for sorting so that the number of significant samples included in each range are as equal as possible” will be described later. Information relating to the division into N ranges obtained by this division processing (hereinafter referred to as division information) is output from the division unit 150 and provided to the gain correction amount encoding unit 140.

  Details of the sorting process performed by the sorting unit 150 will be described later.

<Gain Correction Amount Encoding Unit 140>
As shown in FIG. 24, the gain correction amount encoding unit 140 includes, for example, a storage unit 141, a bit allocation unit 142, and an encoding unit 143. The gain correction amount encoding unit 140 may include a multiplication unit 144 indicated by a broken line in FIG. 24 as necessary. The gain correction amount encoding unit 140 receives the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }], the quantized global gain g ^, and the corrected quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min ,..., L _max }] and classification information are input. The gain correction amount encoding unit 140 classifies the quantized global gain into a plurality of gain correction amounts using a plurality of gain correction amount codebooks stored in the storage unit 141 of the gain correction amount encoding unit 140. Multiplying the correction gain obtained by correcting for each range and the value of each sample of the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min , ..., L _max }] Gain that is a code for specifying a gain correction amount that maximizes the correlation or minimizes the error between the signal sequence obtained in this way and the input signal sequence X (ω) [ω∈ {L _min ,..., L _max }] The correction amount code is output. The gain correction amount encoding unit 140 outputs a U _g bit gain correction amount code based on the input gain correction bit number U _g .

At this time, the bit allocation unit 142 of the gain correction amount encoding unit 140 has a range divided into gain correction amounts corresponding to each range divided into N pieces (N = 2D, where D is an integer of 2 or more). the (k is each integer from 1 to D-1) 2 ^k pieces in preference to the gain correction amount corresponding to the range number of segmented range is often within the scope of the gain correction amount in the summary range Assign bits. Specifically, a range of 2 ^N-1 divided ranges, a range of 2 ^N-2 divided ranges, ..., a range of 2 divided ranges, a divided range Prioritize bits in order of.

  A range having a large number of divided ranges included in the range may be abbreviated as a large range. In addition, a range having a small number of divided ranges included in the range may be abbreviated as a small range. Furthermore, a range in which the number of divided ranges included in the range is the same may be abbreviated as a range of the same size.

  For the gain correction amount corresponding to the same size range, bits may be assigned in preference to the gain correction amount corresponding to any range, but for the gain correction amount corresponding to the range where auditory importance is high It is preferable to assign bits preferentially. For example, it is assumed that information on auditory importance in each range is input from a unit (not shown) in the encoding device 100 or from the outside of the encoding device 100. In this case, for the gain correction amount corresponding to the same size range, the bit is given priority over the gain correction amount corresponding to the high auditory importance range according to the auditory importance information of each range. Assign. In other words, for the gain correction amount corresponding to the range where the number of divided ranges included in the range is the same, a bit may be assigned in preference to the gain correction amount corresponding to the range having high auditory importance.

  In general, the lower frequency band is often higher in auditory importance than the higher frequency band. For this reason, as for the gain correction amount corresponding to the same size range, a bit may be assigned in preference to the gain correction amount corresponding to the low frequency range. That is, for the gain correction amount corresponding to the range where the number of the divided ranges included in the range is the same, a bit may be assigned in preference to the gain correction amount corresponding to the low frequency range.

[First example of gain correction amount encoding process]
The first example of the gain correction amount encoding process is an example in which the gain obtained by adding the quantized global gain g ^ and the gain correction amount is used as the correction gain. Here, with respect to the gain correction amount corresponding to the same size range, the bit is assigned with priority over the gain correction amount corresponding to the low frequency range, and the vector correction is performed for the gain correction amounts in the two adjacent ranges. explain. A case where the post-correction quantization normalized signal sequence is divided into four ranges will be described.

In this example, the corrected quantized normalized signal sequence is divided into a first range R1, a second range R2, a third range R3, and a fourth range R4. For example, as shown in FIG. 25, the first range R1 is a section [L _min , L ₍₁₎ -1], and the second range R2 is a section [L ₍₁₎ , L ₍₂₎ -1]. Yes, the third range R3 is the section [L ₍₂₎ , L ₍₃₎ -1], and the fourth range R4 is the section [L ₍₃₎ , L _max ]. The horizontal axis in FIG. 25 represents the sample number. These ranges R1, R2, R3, R4 is (although k is each integer generally from 1 to D-1, k = 1 in this example) ^{2 k} pieces are summarized by. A range in which the range R1 and the range R2 are combined is referred to as a range R12, and a range in which the range R3 and the range R4 are combined is referred to as a range R34.

These ranges R1, R2, R3, R4, R12, and R34 are divided into groups each composed of a ranges for each range of the same size. In general, a is an integer of 2 or more, but in this example, a = 2. In this example, the range R1 and the range R2 constitute a group G12, the range R3 and the range R4 constitute a group G34, and the range R12 and the range R34 constitute a group G1234. That is, the range constituting each group is as follows.
Group G12 = {Range R1, Range R2}
Group G34 = {Range R3, Range R4}
Group G1234 = {range R12, range R34}
Vector quantization is performed in each of these groups G12, G34, and G1234.

  Specifically, the following three vector quantizations are performed. The first vector quantization is vector quantization for the group G12, that is, vector quantization of a gain correction amount corresponding to the range R1 and a gain correction amount corresponding to the range R2. This is hereinafter referred to as “first VQ”. The second vector quantization is vector quantization for the group G34, that is, vector quantization of a gain correction amount corresponding to the range R3 and a gain correction amount corresponding to the range R4. This is hereinafter referred to as “second VQ”. The third vector quantization is vector quantization for the group G1234, that is, vector quantization of a gain correction amount corresponding to the range R12 and a gain correction amount corresponding to the range R34. This is hereinafter referred to as “third VQ”.

<Storage unit 141>
The storage unit 141 of the gain correction amount encoding unit 140 stores a gain correction amount candidate Δ ₁ (ma) corresponding to the range R1, a gain correction amount candidate Δ ₂ (ma) corresponding to the range R2, and these 2 ^Ma pairs (2 Ma powers, Ma is an integer of 1 or more, ma∈ {1,..., 2 ^Ma }]) are stored with the code idx ₁₂ (ma) that identifies gain correction amount candidates. Yes. Specifically, a set of Δ ₁ (1), Δ ₂ (1) and idx ₁₂ (1), a set of Δ ₁ (2), Δ ₂ (2) and idx ₁₂ (2), ... , Δ ₁ (2 ^Ma ), Δ ₂ (2 ^Ma ), and idx ₁₂ (2 ^Ma ) are stored in the storage unit 141 as a first VQ gain correction amount codebook. The number of bits of the code idx ₁₂ (ma) is Ma bits. That is, the code (hereinafter referred to as the first VQ code) idx ₁₂ output by the first VQ is Ma bits.

If a vector composed of a number (a = 2 in this example) of gain correction amount candidates is called a gain correction amount candidate vector, the gain correction amount codebook of the first VQ has Δ ₁ (1) and delta ₂ (1) gain correction amount candidate vectors constructed by, Δ ₁ (2) and delta ₂ gain correction amount candidate vectors constructed by _{(2), ···, Δ 1} (2 Ma) and delta ₂ (2 ^Ma) and a total of 2 ^Ma pieces of gain correction amount candidate vectors gain correction amount candidate vectors comprised of a total of 2 ^Ma pieces of gain correction amount candidate vector and the corresponding total 2 ^Ma number of code delta ₁₂ (1 ), Δ ₁₂ (2),..., Idx ₁₂ (2 ^Ma ).

Further, the storage unit 141 stores gain correction amount candidates Δ ₃ (mb) corresponding to the range R3, gain correction amount candidates Δ ₄ (mb) corresponding to the range R4, and gain correction amount candidates. set of 2 ^Mb pieces of a code idx ₃₄ (mb) for identifying (2 Mb th power, Mb is an integer of 1 or more, mb∈ {1, ..., 2 Mb}])) are stored. Specifically, a set of Δ ₃ (1), Δ ₄ (1) and idx ₃₄ (1), a set of Δ ₃ (2), Δ ₄ (2) and idx ₃₄ (2), ... , Δ ₃ (2 ^Mb ), Δ ₄ (2 ^Mb ), and idx ₃₄ (2 ^Mb ) are stored in the storage unit 141 as a second VQ gain correction amount codebook. Mb may be the same value as Ma or a different value. The number of bits of the code idx ₃₄ (mb) is Mb bits. That is, a code (hereinafter referred to as a second VQ code) idx ₃₄ output by the second VQ is Mb bits.

The second VQ gain correction amount codebook includes a gain correction amount candidate vector composed of Δ ₃ (1) and Δ ₄ (1), and a gain correction amount composed of Δ ₃ (2) and Δ ₄ (2). Candidate vectors, 2 ^Mb gain correction amount candidate vectors, and a total of 2 ^Mb gain correction amounts of gain correction amount candidate vectors composed of Δ ₃ (2 ^Mb ) and Δ ₄ (2 ^Mb ) It can be considered that a total of 2 ^Mb codes Δ ₃₄ (1), Δ ₃₄ (2),..., Idx ₃₄ (2 ^Mb ) corresponding to the candidate vectors are stored.

Further, the storage unit 141 stores a gain correction amount candidate Δ ₁₂ (mc) in the range R12, a gain correction amount candidate Δ ₃₄ (mc) in the range R34, and a code idx for specifying these gain correction amount candidates. ₁₂₃₄ (mc) set of the can 2 ^Mc pieces (2 Mc th power, Mc is an integer of 1 or more, mc∈ {1, ..., 2 Mc}])) are stored. Specifically, a set of Δ ₁₂ (1), Δ ₃₄ (1) and idx ₁₂₃₄ (1), a set of Δ ₁₂ (2), Δ ₃₄ (2) and idx ₁₂₃₄ (2), ... , Δ ₁₂ (2 ^Mc ), Δ ₃₄ (2 ^Mc ), and idx ₁₂₃₄ (2 ^Mc ) are stored in the storage unit 141 as a third VQ gain correction amount codebook. Mc may be the same value as Ma or a different value. Further, Mc may be the same value as Mb or a different value. The number of bits of the code idx ₁₂₃₄ (mc) is Mc bits. A code (hereinafter referred to as a third VQ code) idx ₁₂₃₄ output by the third VQ is Mc bits.

The third VQ gain correction amount codebook includes a gain correction amount candidate vector composed of Δ ₁₂ (1) and Δ ₃₄ (1), and a gain correction amount composed of Δ ₁₂ (2) and Δ ₃₄ (2). candidate ^{_{vectors, ···, Δ 12 (2 Mc}} ) and delta ₃₄ and a total of 2 ^Mc number of gain correction amount candidate vectors constructed gain correction amount candidate vector (2 ^Mc), a total of 2 ^Mc number of gain correction amount It may be considered that a total of 2 ^Mc codes Δ ₁₂₃₄ (1), Δ ₁₂₃₄ (2),..., Idx ₁₂₃₄ (2 ^Mc ) corresponding to the candidate vectors are stored.

Thus, a plurality of gain correction amount candidates are associated with each of the divided ranges and the ranges obtained by collecting the divided ranges by 2 ^k pieces (k is an integer from 1 to D-1). Has been. In this example, Δ ₁ (1), ..., Δ ₁ (2 ^Ma ) is associated with the range R1, and Δ ₂ (1), ..., Δ ₂ (2 ^Ma ) is associated with the range R2. It is, in the range _{R3 Δ 3 (1), ...} , Δ 3 (2 Mb) have been associated, in the range _{R4 Δ 4 (1), ...} , Δ 4 (2 Mb) is correlated Δ ₁₂ (1),..., Δ ₁₂ (2 ^Mc ) is associated with the range R12, and Δ ₃₄ (1),..., Δ ₃₄ (2 ^Mc ) is associated with the range R34. Has been.

  The gain correction amount candidates have a relationship that the absolute value of the gain correction amount candidate corresponding to the large range is larger than the absolute value of the gain correction amount candidate corresponding to the smaller range. May be. That is, the absolute value of the gain correction amount candidate corresponding to a range having a large number of divided ranges included in the range is included in the range than the number of the divided ranges included in the range. Alternatively, there may be a relationship that the absolute value of gain correction amount candidates corresponding to a range with a small number of ranges is larger.

  In this example, the range R12 and the range R34 are larger than the range R1, the range R2, the range R3, and the range R4.

_{Thus, Δ 12 (1), ...} , the absolute value of Δ ₁₂ (2 ^Mc) _{is, Δ 1 (1), ...} , the absolute value of ^{_{Δ 1 (2 Ma), Δ}} 2 (1), ..., Δ 2 ( 2 ^Ma ) absolute value, Δ ₃ (1), ..., Δ ₃ (2 ^Mb ) absolute value and Δ ₄ (1), ..., Δ ₄ (2 ^Mb ) absolute value .

_{Similarly, Δ 34 (1), ...} , the absolute value of Δ ₃₄ (2 ^Mc) _{is, Δ 1 (1), ...} , the absolute value of ^{_{Δ 1 (2 Ma), Δ}} 2 (1), ..., Δ 2 The absolute value of (2 ^Ma ), the absolute value of Δ ₃ (1), ..., Δ ₃ (2 ^Mb ) and the absolute value of Δ ₄ (1), ..., Δ ₄ (2 ^Mb ) Good.

  For example, a gain correction amount candidate vector can be generated as follows.

First, 2 ^Md number of normalized gain correction amount candidate vectors composed of a values are stored in the storage unit 141. For example, Md = Ma = Mb = Mc. When a value constituting the normalized gain correction amount candidate vector is expressed as Δ ¹ (m),..., Δ ^a (m), the normalized gain correction amount candidate vector is (Δ ¹ (m),. ^a (m)). The storage unit 141, 2 ^Md pieces of normalized gain correction amount candidate vectors, i.e. ^{(Δ 1 (1), ...} , Δ a (1)), ..., (Δ 1 (2 Md), ..., Δ a ( 2 ^Md )) is stored.

  In addition, it is assumed that a predetermined coefficient is determined for each size of the range. This coefficient is larger as the corresponding range is larger. In other words, this coefficient is larger as the number of divided ranges included in the range is larger.

In the above example, the ranges R12, R34 are larger than the ranges R1, R2, R3, R4. For this reason, the coefficient step ₁₂₃₄ corresponding to the ranges R12 and R34 is larger than the coefficient step ₁₂ corresponding to the ranges R1 and R2. Similarly, the coefficient step ₁₂₃₄ corresponding to the ranges R12 and R34 is larger than the coefficient step ₃₄ corresponding to the ranges R3 and R4.

The gain correction amount corresponding to the range R12 and the gain correction amount corresponding to the range R34 are corrected within the quantization step width of the quantized global gain g ^. The gain correction amount corresponding to the range R1 and the gain correction amount corresponding to the range R2 are corrected within the range of the quantization step width of the gain correction amount corresponding to the range R12 × the coefficient step ₁₂ . The gain correction amount corresponding to the range R3 and the gain correction amount corresponding to the range R4 are corrected within the range of the quantization step width of the gain correction amount corresponding to the range R34 × the coefficient step ₃₄ .

At this time, a vector obtained by multiplying the normalized gain correction amount candidate vector by a coefficient corresponding to the size of the range is set as the gain correction amount candidate vector of the range. In other words, the normalized gain correction amount candidate vector ^{(Δ 1 (m), ...} , Δ a (m)) a number of values delta ¹ constituting the (m), ..., to each of the delta ^a (m), A vector (stepΔ ¹ (m),..., StepΔ ^a composed of ^a values stepΔ ¹ (m),..., StepΔ ^a (m) obtained by multiplying the coefficient step corresponding to the size of the range. (m)) is a gain correction amount candidate vector in that range. This multiplication is performed by the multiplication unit 144 of the gain correction amount encoding unit 140. Since there are 2 ^Md normalized gain correction amount candidate vectors (Δ ¹ (m),..., Δ ^a (m)), by performing this multiplication for each of m = 1,..., 2 ^Md , 2 ^Md Gain correction amount candidate vectors (step Δ ¹ (m),..., Step Δ ^a (m)) are obtained.

In the above example of a = 2, when Md = Ma, the gain correction amount candidate vectors (Δ ₁ (m), Δ ₂ (m)) corresponding to the ranges R1, R2 constituting the group G12 are (Δ ₁ (m), Δ ₂ (m)) = (step ₁₂ Δ ¹ (m), step ₁₂ Δ ² (m)) [m = 1,..., 2 ^Ma ]. When Md = Mb, the gain correction amount candidate vectors (Δ ₃ (m), Δ ₄ (m)) corresponding to the ranges R3, R4 constituting the group G34 are (Δ ₃ (m), Δ ₄ (m )) = (step ₃₄ Δ ¹ (m), step ₃₄ Δ ² (m)) [m = 1,..., 2 ^Mb ]. When Md = Mb, the gain correction amount candidate vectors (Δ ₁₂ (m), Δ ₃₄ (m)) corresponding to the ranges R 12, R 34 constituting the group G 1234 are (Δ ₁₂ (m), Δ ₃₄ (m )) = (step ₁₂₃₄ Δ ¹ (m), step ₁₂₃₄ Δ ² (m)) [m = 1,..., 2 ^Mc ].

Note that, as described in [Specific Example 3 of Encoding Processing] below, the encoding unit 143 performs gain correction on at least one of the first VQ code idx ₁₂ , the second VQ code idx _34, and the third VQ code idx _1234. In some cases, only some of the bits included in the code that specifies the quantity candidates are output as the code. In this case, the codes included in the gain correction amount code book are set as follows, for example.

An example of the third VQ gain correction amount codebook in the case of Mc = 2 will be described. The storage unit 141 includes a set of Δ ₁₂ (1), Δ ₃₄ (1) and idx ₁₂₃₄ (1), a set of Δ ₁₂ (2), Δ ₃₄ (2) and idx ₁₂₃₄ (2), Δ ₁₂ (3), Δ ₃₄ (3) and idx ₁₂₃₄ (3), and Δ ₁₂ (4), Δ ₃₄ (4), and idx ₁₂₃₄ (4) Stored as a codebook. Here, idx ₁₂₃₄ (1) is 2 bits of {0,0}, idx ₁₂₃₄ (2) is 2 bits of {1,0}, idx ₁₂₃₄ (3) is 2 bits of {0,1}, idx ₁₂₃₄ Let (2) be 2 bits of {1,1}.

<Bit allocation unit 142>
The bit allocation unit 142 of the gain correction amount encoding unit 140 includes a gain correction amount corresponding to the range R1, a gain correction amount corresponding to the range R2, a gain correction amount corresponding to the range R3, a gain correction amount corresponding to the range R4, Bits are assigned in preference to the gain correction amount corresponding to a large range among the six gain correction amounts of the gain correction amount corresponding to the range R12 and the gain correction amount corresponding to the range 34. That is, bits are assigned with priority over the gain correction amount corresponding to the range R12 and the gain correction amount corresponding to the range R34.

  In other words, bits are allocated in preference to the third VQ code corresponding to a larger range among the first VQ code, the second VQ code, and the third VQ code. For the first VQ code and the second VQ code, bits are assigned with priority over the first VQ code corresponding to a lower frequency range. A specific bit allocation method is as follows.

When the input gain correction bit number U _g is equal to or smaller than Mc, bits are assigned to gain correction amounts Δ ₁₂ (mc) and Δ ₃₄ (mc) corresponding to the ranges R12 and R34, respectively, but the ranges R1 and R2 , R3, and R4, no bits are assigned to the gain correction amounts Δ ₁ (ma), Δ ₂ (ma), Δ ₃ (mb), and Δ ₄ (mb), respectively. Therefore, in this case, the bit allocation unit 142 instructs the encoding unit 143 to perform only the third VQ and output the third VQ code idx ₁₂₃₄ as the gain correction amount code.

When the input gain correction bit number U _g is greater than Mc and less than or equal to Ma + Mc, the gain correction amounts Δ ₁₂ (mc) and Δ ₃₄ (mc) corresponding to the ranges R12 and R34, respectively, and the ranges R1 and R2, respectively. Bits are assigned to the gain correction amounts Δ ₁ (ma) and Δ ₂ (ma) to be performed, but bits are assigned to the gain correction amounts Δ ₃ (mb) and Δ ₄ (mb) corresponding to the ranges R3 and R4, respectively. Absent. Therefore, in this case, the bit allocation unit 142 performs the third VQ and the first VQ, and instructs the encoding unit 143 to output the third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ as gain correction amount codes.

When the input gain correction bit number U _g is larger than Ma + Mc, gain correction amounts Δ ₁ (ma), Δ ₂ (ma), Δ ₃ corresponding to the ranges R1, R2, R3, R4, R12, R34, respectively. Bits are assigned to (mb), Δ ₄ (mb), Δ ₃ (mb), and Δ ₄ (mb). In this case, the bit allocation unit 142 performs the third VQ, the first VQ, and the second VQ, and encodes an instruction to output the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ as gain correction amount codes. To the unit 143.

When the input gain correction bit number U _g is 0 or less, no bit is allocated to any range, and the bit allocation unit 142 does not give an instruction to the encoding unit 143.

<Encoding unit 143>
The encoding unit 143 includes an instruction from the bit allocation unit 142, an input signal sequence X (ω) [ω∈ {L _min ,..., L _max }], a quantized global gain g ^, and a corrected quantized signal. Normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] and classification information are input.

  The encoding unit 143 selects a gain correction amount that minimizes a predetermined error from among a plurality of gain correction amount candidates corresponding to the divided ranges. Further, the encoding unit 143 outputs a gain correction amount code for specifying the selected gain correction amount.

  Since the a ranges constituting each group are respectively associated with a gain correction amount candidates constituting the gain correction amount candidate vector, the encoding unit 143 includes a plurality of gain correction amount candidate vectors. It may be considered that a gain correction amount code for specifying a gain correction amount candidate vector that minimizes a predetermined error for each group is output.

The predetermined error is a quantized normalized signal sequence X after correction to a gain obtained by correcting the quantized global gain for each divided range with a plurality of gain correction amounts for each divided range. ^ ' _Q (ω) The signal sequence obtained by multiplying the values of each sample of [ω∈ {L _min , ..., L _max }] and the input signal series X (ω) [ω∈ {L _min , ..., L _max }]. Specifically, the predetermined error is an addition value defined by the equations (D1), (D3), and (D5).

[Specific Example of Encoding Process 1: Example Using Different Addition Formulas in Three Cases]
Specific example 1 is an example in which the input gain correction bit number U _g is either Mc, Mc + Ma, or Mc + Ma + Mb.

(a) When the input gain correction bit number U _g is Mc When the input gain correction bit number U _g is Mc, only the third VQ is performed and the third VQ code idx ₁₂₃₄ is output as the gain correction amount code An instruction to do this is given from the bit allocation unit 142. In this case, the encoding unit 143 first calculates an addition value defined by Expression (D1) for each mc of 1 to 2 ^Mc . In the equation (D1), the section [L _min , L ₍₂₎ −1] corresponds to the range R12, and the section [L ₍₂₎ , L _max ] corresponds to the range R34.

The added value defined by the equation (D1) is a value obtained by adding the quantized global gain g ^ and the gain correction amount candidate Δ ₁₂ (mc) in the range R12 and the corrected quantization normalization in the range R12. Signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₂₎ −1}] multiplied by the value of each sample and the input signal sequence X ( ω) [ω∈ {L _min ,..., L ₍₂₎ −1}] corresponding to the sum of squares of the difference between samples, the quantized global gain g ^ and the gain correction amount candidate Δ ₃₄ (in the range R34) mc) and the value of each sample of the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₂₎ ,..., L _max }] in the range 34 And the sum of squares of the difference between corresponding samples of the signal sequence obtained by multiplying and the input signal sequence X (ω) [ω∈ {L ₍₂₎ ,..., L _max }] in the range R34 Value.

Next, the encoding unit 143 selects mc that minimizes the addition value, and outputs the code idx ₁₂₃₄ (mc) corresponding to the selected mc as the third VQ code idx ₁₂₃₄ . In this example, the third VQ code idx ₁₂₃₄ becomes the gain correction amount code idx. The third VQ code idx ₁₂₃₄ is obtained by the equation (D2).

(b) When the input gain correction bit number U _g is Mc + Ma When the input gain correction bit number U _g is Mc + Ma, the third VQ and the first VQ are performed, and the third VQ code idx ₁₂₃₄ The bit allocation unit 142 instructs to output the first VQ code idx ₁₂ as a gain correction amount code. In this case, the coding unit 143, first, for each set from mc and 1 is any one of 1 to 2 ^Mc and ma is either 2 ^Ma (mc, ma), the formula (D3) Calculate the added value to be defined. In Formula (D3), the section [L _min , L ₍₁₎ -1] corresponds to the range R1, the section [L ₍₁₎ , L ₍₂₎ -1] corresponds to the range R2, and the section [ L _min , L ₍₂₎ −1] corresponds to the range R12, and the section [L ₍₂₎ , L _max ] corresponds to the range R34.

The added value defined by the equation (D3) is obtained by adding the quantized global gain g ^, the gain correction amount candidate Δ ₁₂ (mc) in the range R12, and the gain correction amount candidate Δ ₁ (ma) in the range R1. Is multiplied by the value of each sample of the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1. Sum of squares of differences between corresponding samples of the signal sequence obtained in this way and the input signal sequence X (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1, and the quantized global gain The value obtained by adding g ^, the gain correction amount candidate Δ ₁₂ (mc) of the range R12 and the gain correction amount candidate Δ ₂ (ma) of the range R2, and the corrected quantized normalized signal of the range R2 The sequence X ^ ′ _Q (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] multiplied by the value of each sample and the input signal sequence X (( _{ω) [ω∈ {L (1} ), ..., L (2) -1}] corresponding to the And square sum of the difference between the sample, the quantization global gain g ^ and the gain correction amount candidate delta ₃₄ (mc) and the corrected quantized normalized signal sequence of values obtained by adding the range R34 ranging R34 X ^ ′ _Q (ω) [ω∈ {L ₍₂₎ ,..., L _max }] multiplied by the value of each sample and the input signal series X (ω) [ω∈ {L ₍₂₎ ,..., L _max }] and the sum of squares of the differences between the corresponding samples.

Next, the encoding unit 143 sets the code idx ₁₂₃₄ (mc) corresponding to the pair of mc and ma that minimizes the added value as the third VQ code idx ₁₂₃₄ and sets the code idx ₁₂ (ma) as the first VQ code idx _12. As the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the gain correction amount code idx. The third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ are obtained by the equation (D4).

(c) When the input gain correction bit number U _g is Mc + Ma + Mb When the input gain correction bit number U _g is Mc + Ma + Mb, the third VQ, the first VQ, and the second VQ are The bit allocation unit 142 instructs to output the third VQ code idx ₁₂₃₄ , the first VQ code idx ₁₂ and the second VQ code idx ₃₄ as gain correction amount codes. In this case, the encoding unit 143 firstly sets a pair (mc, 1) of mc that is any one of 1 to 2 ^Mc , ma that is any one of 1 to 2 ^Ma , and mb that is any one of 1 to 2 ^Mb . For each of ma, mb), the addition value defined by equation (D5) is calculated. In the equation (D5), the section [L _min , L ₍₁₎ -1] corresponds to the range R1, the section [L ₍₁₎ , L ₍₂₎ -1] corresponds to the range R2, and the section [ L ₍₂₎ , L ₍₃₎ -1] corresponds to range R3, section [L ₍₃₎ , L _max ] corresponds to range R4, and section [L _min , L ₍₂₎ -1] ranges Corresponding to R12, the section [L ₍₂₎ , L _max ] corresponds to the range R34.

The added value defined by the equation (D5) is obtained by adding the quantized global gain g ^, the gain correction amount candidate Δ ₁₂ (mc) in the range R12, and the gain correction amount candidate Δ ₁ (ma) in the range R1. Is multiplied by the value of each sample of the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1. Sum of squares of differences between corresponding samples of the signal sequence obtained in this way and the input signal sequence X (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1, and the quantized global gain The value obtained by adding g ^ and the gain correction amount candidate Δ ₁₂ (mc) in the range R12 and the gain correction amount candidate Δ ₂ (ma) in the range R1 and the corrected quantized normalized signal in the range R2 The sequence X ^ ′ _Q (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] multiplied by the value of each sample and the input signal sequence X (( _{ω) [ω∈ {L (1} ), ..., L (2) -1}] corresponding to the And square sum of the differences between the sample obtained by adding the quantized global gain g ^ and scope gain correction amount candidate delta ₃₄ of R34 (mc) and the gain correction amount candidate delta ₃ range R3 (mb) Multiplying the value by the value of each sample of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1}] in the range R3 The sum of squares of the difference between corresponding samples of the obtained signal sequence and the input signal sequence X (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ -1}] in the range R3, and the quantization global A value obtained by adding the gain g ^, the gain correction amount candidate Δ ₃₄ (mc) of the range R34, and the gain correction amount candidate Δ ₄ (mb) of the range R4 and the corrected quantization normalized value of the range R4 Signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max }] multiplied by the value of each sample and the input signal sequence X (ω) [ The corresponding sun with ω∈ {L ₍₃₎ ,…, L _max }] The sum of squared differences between pulls.

Next, the encoding unit 143 sets the code idx ₁₂₃₄ (mc) corresponding to the set of mc, ma, and mb that minimizes the added value as the third VQ code idx ₁₂₃₄ and sets the code idx ₁₂ (ma) as the first VQ code idx. ₁₂ , the code idx ₃₄ (mb) is output as the second VQ code idx ₃₄ , and the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ are output as the gain correction amount code idx. The third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ are obtained by the equation (D6).

[Specific example of encoding process 2: Example using the same addition formula in three cases]
Specific example 2 is also an example in the case where the input gain correction bit number U _g is either Mc, Mc + Ma, or Mc + Ma + Mb, as in specific example 1.

  In the first specific example, the addition value is obtained by any one of the equations (D1), (D3), and (D5). In the second specific example, the addition value is obtained only by the equation (D5).

(a) When the input gain correction bit number U _g is Mc When the input gain correction bit number U _g is Mc, the gain correction amount Δ ₁₂ (mc), corresponding to the ranges R12 and R34, respectively. Although bits are assigned to Δ ₃₄ (mc), gain correction amounts Δ ₁ (ma), Δ ₂ (ma), Δ ₃ (mb), Δ ₄ (mb) corresponding to ranges R1, R2, R3, and R4, respectively. No bits are assigned to. In this case, an instruction to perform only the third VQ and output the third VQ code idx ₁₂₃₄ as a gain correction amount code is issued from the bit allocation unit 142.

The encoding unit 143 defines Δ ₁ (ma), Δ ₂ (ma), Δ ₃ (mb), and Δ ₄ (mb) as 0, and defines mc from 1 to 2 ^Mc by equation (D5). Find the added value. Then, the encoding unit 143 selects mc that minimizes the addition value defined by the equation (D5), and outputs the code idx ₁₂₃₄ (mc) corresponding to the selected mc as the third VQ code idx ₁₂₃₄ . . In this example, the third VQ code idx ₁₂₃₄ becomes the gain correction amount code idx.

(b) When the input gain correction bit number U _g is Mc + Ma When the input gain correction bit number U _g is Mc + Ma, the gain correction amount Δ ₁₂ corresponding to the ranges R12 and R34, respectively. Bits are assigned to gain correction amounts Δ ₁ (ma) and Δ ₂ (ma) corresponding to (mc), Δ ₃₄ (mc) and ranges R1 and R2, respectively, but gain correction amounts corresponding to ranges R3 and R4, respectively. No bits are assigned to Δ ₃ (mb) and Δ ₄ (mb). In this case, an instruction to perform the third VQ and the first VQ and output the third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ as a gain correction amount code is issued from the bit allocation unit 142.

Encoding unit 143, in this _{case, Δ 3 (mb), Δ} 4 a (mb) as 0, ma set from mc and 1 is any one of 1 to 2 ^Mc are either 2 ^Ma (mc , ma), an addition value defined by the equation (D5) is obtained. Then, the encoding unit 143 sets the code idx ₁₂₃₄ (mc) corresponding to the minimum mc and ma that minimizes the addition value as the third VQ code idx ₁₂₃₄ and the code idx ₁₂ (ma) as the first VQ code idx _12. The third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ are output as the gain correction amount code idx.

(c) When the input gain correction bit number U _g is Mc + Ma + Mb When the input gain correction bit number U _g is Mc + Ma + Mc, all ranges R1, R2, R3 A bit is assigned to each of R4, R12, and R34. In this case, an instruction to perform the third VQ, the first VQ, and the second VQ and output the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ as the gain correction amount code is issued from the bit allocation unit 142.

The encoding unit 143 obtains the addition value by the equation (D5) as in the first specific example, and sets the code idx ₁₂₃₄ (mc) corresponding to the minimum mc that minimizes the addition value as the third VQ code idx _1234. idx ₁₂ (ma) is the first VQ code idx ₁₂ and code idx ₃₄ (ma) is the second VQ code idx ₃₄ , and the third VQ code idx ₁₂₃₄ , the first VQ code idx ₁₂ and the second VQ code idx ₃₄ are the gain correction amount code idx. Output as.

As described above, the gain correction amount to which no bit is assigned is not stored in the storage unit 141 but is set to 0. Therefore, it can be considered as a gain correction amount corresponding to the fact that no correction is performed. For example, in the above “(a) When the number of input gain correction bits U _g is Mc”, Δ ₁ (ma), Δ ₂ (ma), Δ ₃ which are gain correction amounts to which no bits are assigned. (mb) and Δ ₄ (mb) are gain correction amounts corresponding to not performing correction.

[Specific example 3 of encoding process: an example including a case where the number of gain correction bits is odd]
In the third example, when the input gain correction bit number U _g includes other than Mc, Mc + Ma, and Mc + Ma + Mb, that is, the input gain correction bit number U _g is any one of 1 or more. It is an example in the case of.

(a) When the number of input gain correction bits U _g is greater than 0 and less than or equal to Mc (0 <U ≦ Mc)
If the input gain correction bit number U _g is greater than 0 and less than or equal to Mc, the bit allocation unit 142 instructs to perform only the third VQ and output the third VQ code idx ₁₂₃₄ as the gain correction amount code.

In this case, the encoding unit 143 sets Δ ₁₂ (mc) and Δ ₃₄ (mc) to 0 for all ^{mcs in} the range of 2 ^Ug +1 to 2 ^Mc , and Δ ₁ for all ma from 1 to Ma. (ma) and Δ ₂ (ma) are set to 0, and Δ ₃ (mb) and Δ ₄ (mb) are set to 0 for all mb from 1 to Mb, and the added value is obtained by the equation (D5).

Then, the encoding unit 143 sets, as the third VQ code idx ₁₂₃₄ , a portion of the U _g bit that can distinguish all mcs of 1 to 2 ^Ug from the code idx ₁₂₃₄ (mc) corresponding to mc that minimizes the addition value. The third VQ code idx ₁₂₃₄ is output as the gain correction amount code idx.

For example, in the case of U _g = 1 and Mc = 2, {0} which is the first bit of the {0,0} bits of idx ₁₂₃₄ (1) or {x of idx ₁₂₃₄ (2) {1}, which is the first bit of the two bits of 1,0}, is set as the third VQ code idx ₁₂₃₄ .

(b) When the number of input gain correction bits U _g is greater than Mc and less than or equal to Mc + Ma (Mc <U ≦ Mc + Ma)
When the input gain correction bit number U _g is greater than Mc and less than or equal to Mc + Ma, an instruction to perform the third VQ and the first VQ and output the third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ as a gain correction amount code is issued. This is performed from the bit allocation unit 142.

In this case, the encoding unit 143 sets Δ ₁ (ma) and Δ ₂ (ma) to 0 for all ma in the range of 2 ^Ug-Mc +1 to 2 ^Ma , and sets all mb of 1 to 2 ^Mb. And Δ ₃ (mb) and Δ ₄ (mb) are set to 0, and the added value is obtained by equation (D5).

Then, the encoding unit 143 sets the code idx ₁₂₃₄ (mc) corresponding to mc and ma having the minimum addition value as the third VQ code idx _1234, and 1 to 2 ^{Ug-Mc of the} codes idx ₁₂ (ma) A portion of U _g -Mc bits that can distinguish all of ma is output as the first VQ code idx ₁₂ , and the third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ are output as the gain correction amount code idx.

(c) When the input gain correction bit number U _g is larger than Mc + Ma (Mc + Ma <U)
When the input gain correction bit number U _g is larger than Mc + Ma, the third VQ, the first VQ, and the second VQ are performed, and the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ are gain correction amount codes. Is output from the bit allocation unit 142.

In this case, the encoding unit 143 sets Δ ₃ (mb) and Δ ₄ (mb) to 0 for all ^{mbs in} the range of 2 ^Ug-Mc-Ma +1 to 2 ^Mb , and sets the addition value as an expression (D5 )

Then, the encoding unit 143 sets the code idx ₁₂₃₄ (mc) corresponding to the minimum mc, ma, and mb that minimizes the added value as the third VQ code idx _1234, and sets the code idx ₁₂ (ma) as the first VQ code. and idx _12, a U _g -Mc-Ma-bit part that can distinguish all mc 1 to 2 ^Ug-Mc-Ma of the code idx ₃₄ (mc) and the 2VQ code idx _34, the 3VQ code idx ₁₂₃₄ And the first VQ code idx ₁₂ , the second VQ code idx _34, and the gain correction amount code idx.

  Note that Equation (D2), Equation (D4), and Equation (D6) correspond to the vector quantization based on the criterion that minimizes the error, but the vector quantization and error based on the criterion that maximizes the correlation Of course, a technique such as scalar quantization based on the criterion that minimizes or maximizes the correlation may be applied.

  Note that gain correction amount candidates used in each vector quantization may be stored in one gain correction amount codebook to generate a gain correction amount code.

It is assumed that the number of divided ranges is ^2D . 2 ^D number of segmented 2 ^k pieces by summarizing the number of ranges range is ^{2 D / 2 k = 2 Dk} . Therefore, the number of the divided ranges and the ranges obtained by collecting the divided ranges by 2 ^k pieces (k is an integer from 1 to D-1) is 2 ^D + Σ _{d = 1} ^D-1 2 ^Dd , In total, Σ _{d = 1} ^D 2 ^d = 2 ^D + Σ _{d = 1} ^D−1 2 ^Dd . Hereinafter, A = Σ _{d = 1} ^D 2 ^d .

In this case, it is assumed that the gain correction amount candidate vector includes A gain correction amount candidates. 2 ^D number of segmented range and by collectively range of 2 ^D number of segmented range of 2 ^k pieces (each integer from k 1 to D-1), respectively constituting the gain correction amount candidate vectors Assume that these are associated with A gain correction amount candidates.

In the example in which D = 2 and k = 1 and the addition value is obtained using the equation (D5), A = Σd _{= 1} ² 2 ^d = 2 + 4 = 6, and the gain correction amount of the index idx (m) Candidate vectors (Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m)) are six gain correction amount candidates Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m). Candidate gain correction amounts Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m) are ranges R12, R34, R1, respectively. It corresponds to R2, R3, R4.

The gain correction amount code book stores a plurality of gain correction amount candidate vectors. In the above example, for example, 2 ^Me gain correction amount candidate vectors (Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m )) [m = 1,... 2 ^Me ] is stored in the gain correction amount code book. Me is an integer of 2 or more.

  In this case, the encoding unit 143 obtains a gain correction amount code that identifies a gain correction amount candidate vector that minimizes a predetermined error from among a plurality of gain correction amount candidate vectors stored in the gain correction amount codebook. . Here, the predetermined error is an added value defined by, for example, the equation (D5).

  The synthesizing unit 160 receives a normalized signal code, a global gain code, an error code, and a gain correction amount code. The error code and the gain correction amount code are collectively set as a correction code. The synthesis unit outputs a bit stream in which the normalized signal code, the global gain code, and the correction code are collected. The bit stream is transmitted to the decoding device 2.

[Modification of Encoding Unit 143]
The encoding unit 143 of the gain correction amount encoding unit 140 obtains a gain correction amount code for specifying the gain correction amount that minimizes the addition value defined by the equation (D13) instead of the equation (D1). Also good.

s ₁₂ and s _34, for example, is defined as the following equation.

  The encoding unit 143 may obtain a gain correction amount code for specifying a gain correction amount that minimizes the addition value defined by the equation (D14) instead of the equation (D3).

For example, s ₁ and s ₂ are defined as in the following equations.

  The encoding unit 143 may obtain a gain correction amount code for specifying a gain correction amount that minimizes the addition value defined by the equation (D15) instead of the equation (D5).

For example, s ₃ and s ₄ are defined as follows.

In this way, the quantized global gain g ^ is divided into each of a plurality of gain correction amounts for each divided range, and the corrected quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] are divided into values obtained by multiplying the sum of squares of all sample values by the value obtained by dividing the sum of squares of all sample values within the range corresponding to the respective gain correction amounts. You may correct for every range.

  Also, the encoding unit 143 may obtain a gain correction amount code for specifying a gain correction amount that minimizes the addition value defined by the equation (D16) instead of the equation (D5).

In this way, the quantized global gain g ^ is a value obtained by adding a plurality of gain correction amounts for each divided range, and the corrected quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min ,..., L _max }] each divided range with a value obtained by multiplying the sum of squares of all sample values by the value obtained by dividing the sum of squares of the values of all samples within the above divided ranges. You may correct every.

Note that s ₁₂ , s ₃₄ , s ₁ , s ₂ , s ₃ , and s ₄ may be defined as in the following equations, respectively.

c _12, the energy of the samples in the range R12 is the number of sample that is larger than a predetermined value. c _34, the energy of the samples in the range R34 is the number of sample that is larger than a predetermined value. c ₁₂₃₄ is the number of samples in which the energy of the samples in the range R1234 is larger than a predetermined value. c ₁ is the number of samples in which the energy of the samples in the range R1 is larger than a predetermined value. c ₂ is the energy of the samples in the range R2 is the number of sample that is larger than a predetermined value. c ₃ is the number of samples in which the energy of the sample in the range R3 is larger than a predetermined value. c ₄ is the energy of the samples in the range R4 is the number of sample that is larger than a predetermined value.

In this case, the quantized global gain g ^ includes each of a plurality of gain correction amounts for each divided range and the corrected quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min ,..., L _max }] is obtained by dividing the number of samples whose sample energy is greater than a predetermined value by the number of samples whose sample energy within a range corresponding to each gain correction amount is greater than the predetermined value. It is corrected by the value obtained by multiplying. Alternatively, the quantized global gain g ^ is a value obtained by adding a plurality of gain correction amounts for each divided range, and a corrected quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., _Lmax }] is multiplied by the number of samples whose energy is greater than a predetermined value divided by the number of samples whose energy within each of the divided ranges is greater than the predetermined value. The corrected value is corrected for each divided range.

<Details of Sorting Process Performed by Sorting Unit 150>
First, the classification process in “Criteria for classifying energy in each range to be as equal as possible” will be explained, and then classification in “Criteria for classifying so that the number of significant samples included in each range will be as equal as possible” Processing will be described.

  Hereinafter, the “criteria for classifying so that the energy of each range is as equal as possible” is abbreviated as the first standard, and the “criteria for classifying so that the number of significant samples included in each range is as equal as possible” is abbreviated as the second standard. Sometimes.

For example, the classification process based on the “criteria for classifying the energy of each range to be as equal as possible” is, for example, the nth range of the corrected quantized normalized signal sequence (n is each integer from 1 to N−1). The
(a) The sum of squares of the values of all samples included in the first range to the n-th range of the corrected quantized normalized signal sequence and all the samples of the corrected quantized normalized signal sequence So that n of N of the sum of squares of values is closest.
Or
(b) The sum of absolute values of the values of all samples included in the first range to the nth range of the corrected quantized normalized signal sequence and all the samples of the corrected quantized normalized signal sequence So that n / N of the absolute value sum of the values of
Or
(c) The total number of samples from the first range to the n-th range of the corrected quantized normalized signal sequence is from the first range to the n-th range of the corrected quantized normalized signal sequence. So that the sum of squares of the values of all samples included in is a minimum number of samples that is not less than n / N of the sum of squares of the values of all samples of the corrected quantized normalized signal sequence.
Or
(d) The sum of the number of samples from the first range to the nth range of the corrected quantized normalized signal sequence is from the first range to the nth range of the corrected quantized normalized signal sequence. So that the absolute value sum of the values of all the samples included in is a minimum number of samples that is not less than n / N of the absolute value sum of the values of all the samples of the corrected quantized normalized signal sequence.
Or
(e) The sum of the number of samples from the first range to the nth range of the corrected quantized normalized signal sequence is from the first range to the nth range of the corrected quantized normalized signal sequence. So that the sum of squares of the values of all samples included in is the maximum number of samples that is n or less of N / N of the sum of squares of the values of all samples of the corrected quantized normalized signal sequence,
Or
(f) The total number of samples from the first range to the nth range of the corrected quantized normalized signal sequence is from the first range to the nth range of the corrected quantized normalized signal sequence. So that the absolute value sum of the values of all the samples included in is the maximum number of samples that is not more than n / N of the absolute value sum of the values of all the samples of the corrected quantized normalized signal sequence,
Seeking
A range other than the first range to the (N−1) th range in the corrected quantized normalized signal sequence is set as the Nth range of the corrected quantized normalized signal sequence, whereby the corrected quantum This is done by dividing the normalized normalized signal sequence into N ranges.

  The classification process exemplified above is realized by a method of sequentially determining the classification based on the “standard for classifying so that the energy in each range is as equal as possible” sequentially from the first range. According to the classification process exemplified above, it is possible to realize the classification based on the “standard for classifying the energy in each range so as to be as equal as possible” with a small amount of calculation processing.

[First example of classification processing based on the first standard]
A first example of the sorting process based on the first reference will be described with reference to FIGS. 4, 5, and 6. FIG. The sorting process of the first example corresponds to the above (a). The division processing of the first example is performed using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence, and the first quantized normalized signal sequence of the corrected first sequence. The sum of squares of the values of all the samples included in the range to the nth range and the n of N of the squares of the values of all the samples of the corrected quantized normalized signal sequence are closest to each other. Obtaining and correcting a range other than the range from the first range to the (N-1) th range in the corrected quantized normalized signal sequence as the Nth range of the corrected quantized normalized signal sequence This is a process of dividing the post-quantization normalized signal sequence into N ranges.

[[Specific example 1 of the first example of the sorting process based on the first standard: Example of dividing into two ranges]]
FIG. 4 shows an example of dividing into two ranges, that is, an example where N = 2. The modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, formula corrected quantized normalized signal sequence obtained by _{(2) X ^ 'Q [} ω∈ {L min, ..., L max}] of all samples _{X ^' Q (L min)} , ..., 1/2 of the sum of squares of X ^ ' _Q (L _max ) and all samples X ^' _Q (L _min ), ..., X included in the first range of the modified quantized normalized signal sequence ^ 'such that the difference between the square sum of the values of _Q (L _mid -1) is minimized, obtaining the L _mid is a sample number in the lowest frequency side of the second range. That is, L _mid is obtained by the equation (3). Accordingly, the first range is determined as X ^ _Q [ω∈ {L _min ,..., L _mid −1}].

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, number of samples L _mid -L _min of the first range The number of samples in the second range may be L _max −L _mid +1, or anything insofar as the information can identify the first range and the second range.

[[Example 2 of first example of sorting process based on first standard: Example of sorting into four ranges]]
FIG. 5 shows an example in which a modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is divided into four ranges, specifically , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, formula corrected quantized normalized signal sequence obtained by _{(2) X ^ 'Q [} ω∈ {L min, ..., L max}] of all samples _{X ^' Q (L min)} , ..., A quarter sum of squares of X ^ ' _Q (L _max ) and all samples X ^' _Q (L _min ), ..., X included in the first range of the modified quantized normalized signal sequence ^ 'and _Q (L ₍₁₎ -1) sample number in the difference between the square sum of the values of determined so as to minimize L ₍₁₎ to the lowest frequency side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, all samples X ^ ′ _Q (L _min ),..., X of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by the equation (2). ^ ' _Q (L _max ) 2/4 of the sum of squares (that is, 1/2) and all samples included in the first range to the second range of the modified quantized normalized signal sequence _{X ^ 'Q (L min)} , ..., X ^' Q (L (2) -1) of the third range L ₍₂₎ the difference is determined so as to minimize the sum of squared values most The sample number is on the low frequency side. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Further, all samples X ^ ′ _Q (L _min ),..., X of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by the equation (2). ^ ' _Q (L _max ) 3/4 of the sum of squares and all samples X ^' _Q (L _min ) included in the first to third ranges of the modified quantized normalized signal sequence , ..., X ^ ' _Q (L ₍₃₎ -1) L ₍₃₎ obtained so that the difference from the sum of squares is minimized is the sample number on the lowest side of the fourth range. To do. Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L ₍₃₎ ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the first example of classification processing based on the first standard: Example of dividing into N ranges]]
FIG. 6 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified into N ranges, specifically, In this example, L _(n) , which is the sample number on the lowest side of the _(n + 1) th range, is determined as information indicating the boundary between the nth range and the (n + 1) th range. In the following description, L _min is assumed to be L ₍₀₎ .

First, modified quantized normalized signal sequence _{X ^ 'Q (ω) [} ω∈ {L min, ..., L max}] All samples X ^ _Q (L _{min) of, ..., X ^ Q (L} max ) Is calculated. The sum of squares pow is obtained by equation (2).

Next, for each n from n = 1 to N−1, the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by the equation (2) is used. All samples X ^ ' _Q ( _Lmin ), ..., X ^' _Q ( _Lmax ) n-th of the square sum and the first range to the nth range of the corrected quantized normalized signal sequence L _(n) obtained so that the difference from the sum of squares of the values of all the samples X ^ ' _Q (L _min ), ..., X ^' _Q (L _(n) -1 _{) included} in It is assumed that the sample number is on the lowest side of the (n + 1) th range. Thereby, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _(N−1) ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

[Second example of classification processing based on the first standard]
The second example of the sorting process based on the first standard corresponds to the above (b). The sorting process of the second example is the same method as the sorting process of the first example, except that “sum of squares” in the sorting process of the first example is replaced with “sum of absolute values”. According to the sorting process of the second example, it is possible to perform the sorting process with a smaller amount of calculation processing than the sorting process of the first example because the square calculation performed in the sorting process of the first example can be omitted.

[Third example of classification processing based on the first standard]
A third example of the sorting process based on the first reference will be described with reference to FIGS. 7, 8, and 9. FIG. The classification process of the third example corresponds to the above (c). The division processing of the third example is performed by using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence and the first quantized normalized signal sequence of the corrected first sequence. The sum of the number of samples from the range to the nth range is the corrected quantized sum of the squares of the values of all samples included in the first range to the nth range of the corrected quantized normalized signal sequence The minimum number of samples that is equal to or greater than n / N of the square sum of the values of all samples of the normalized signal sequence is obtained, and the first range of the corrected quantized normalized signal sequence is changed from the first range. This is a process of dividing the corrected quantized normalized signal sequence into N ranges by setting a range other than the N−1 range as the Nth range of the corrected quantized normalized signal sequence.

[[Specific example of the third example of the sorting process based on the first standard 1: Example of dividing into two ranges]]
FIG. 7 shows an example of dividing into two ranges, that is, an example in the case of N = 2. The modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, the square sum p _low is p _low ≧ pow from L _min to the index of the number of index omega discrete frequency L _min from the modified quantized normalized signal sequence while increasing in the order X ^ _'Q (ω) / 2 is satisfied, the first range up to the discrete frequency index ω when p _low ≧ pow / 2 is satisfied for the first time, and the value obtained by adding 1 to the index ω is the second range. Output as the index L _mid which is the sample number on the lowest side. As a result, the first range is determined as X ^ ′ _Q [ω∈ {L _min ,..., L _mid −1}].

FIG. 7 is a flowchart for realizing the above processing. The initial value of the discrete frequency index ω is set to L _min , and the initial value of the low frequency energy p _low is set to | X ^ ′ _Q (L _min ) | ² . Then, it is determined whether or not p _low ≧ pow / 2 is satisfied. p _low ≧ if the pow / 2 does not meet the will, the one plus the index of discrete frequency ω as a new ω, 'energy of _Q (ω) | X ^' to p _low X ^ _Q (ω) | Add ² to make new p _low . When p _low ≧ pow / 2 is satisfied, a value obtained by adding 1 to the index ω of the discrete frequency at that time is output as the index L _mid .

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, number of samples L _mid -L _min of the first range The number of samples in the second range may be L _max −L _mid +1, or anything insofar as the information can identify the first range and the second range.

[[Example 2 of the third example of the sorting process based on the first standard: Example of sorting into four ranges]]
FIG. 8 shows an example in which the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is divided into four ranges, specifically , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, the squares of the values of all the samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L ₍₁₎ −1) included in the first range of the corrected quantized normalized signal sequence sum formula corrected quantized normalized determined by (2) signal sequence _{X ^ 'Q [ω∈ {L} min, ..., L max}] of all samples _{X ^' Q (L min)} , ..., X ^ ' _Q (L _max ) is not less than one quarter of the sum of squares, and the highest of the first range from all samples included in the first range of the corrected quantized normalized signal sequence Correction that the sum of squares of the value of the signal sequence X ^ ' _Q (L _min ), ..., X ^' _Q (L ₍₁₎ -2), excluding one sample on the band side, is obtained by equation (2) Square sum of all samples X ^ ' _Q (L _min ), ..., X ^' _Q (L _max ) of post-quantized normalized signal sequence X ^ ' _Q [ω∈ {L _min , ..., L _max }] L _(1), which is smaller than one-fourth of L, is obtained as the sample number on the lowest side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

In addition, all samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L ₍₂₎ −1) included in the first range to the second range of the corrected quantized normalized signal sequence The sum of squares of the values of all the samples X ^ ′ _Q (L _min ) of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by Equation (2) ), ..., X ^ ' _Q (L _max ) is greater than or equal to two-quarters of the sum of squares (ie, half), and from the first range of the modified quantized normalized signal sequence Signal sequence X ^ ' _Q (L _min ), ..., X ^' _Q (L _{(2) -excluding} one sample at the highest side of the second range from all samples included in range 2 The sum of squares of the values of 2) is the total sample X ^ ′ _Q () of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by Equation (2). L _min ), ..., X ^ ' _Q (L _max ) less than two-quarters of the sum of squares (ie, half), L ₍₂₎ is the third range It is obtained as the sample number on the lowest side of. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

In addition, all samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L ₍₃₎ −1) included in the first to third ranges of the corrected quantized normalized signal sequence The sum of squares of the values of all the samples X ^ ′ _Q (L _min ) of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by Equation (2) ), ..., X ^ ' _Q (L _max ), which is not less than three-quarters of the sum of squares, and all the signals included in the first range to the third range of the corrected quantized normalized signal sequence The sum of squares of the values of the signal sequence X ^ ' _Q (L _min ), ..., X ^' _Q (L ₍₃₎ -2), which is obtained by removing one sample at the highest side of the third range from the sample, is , All samples X ^ ′ _Q (L _min ),..., X ^ of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by the equation (2) 'L _(3), which is smaller than three-quarters of the sum of squares of _Q (L _max ), is determined as the sample number on the lowest side of the fourth range. Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

  Specifically, these processes can be realized by, for example, the following.

  First, the energy pow is obtained by equation (2).

Next, it is determined whether or not the condition of the expression (4) is satisfied while increasing i sequentially by 1 from L _min, and the value obtained by adding 1 to i satisfying the condition of the expression (4) is expressed as L ₍₁₎ Asking. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, it is determined whether or not the condition of the expression (5) is satisfied while increasing i by ₁ sequentially from L _(1), and a value obtained by adding 1 to i satisfying the condition of the expression (5) is expressed as L _{(2 )} . Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Further, it is determined whether or not the condition of Expression (6) is satisfied while i is incremented by 1 in order from L ₍₂₎ , and the result of adding 1 to i satisfying the condition of Expression (6) is L _{(3 )} . Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L ₍₃₎ ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the third example of classification processing based on the first standard: Example of dividing into N ranges]]
FIG. 9 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified into N ranges, specifically, In this example, L _(n) , which is the sample number on the lowest side of the _(n + 1) th range, is determined as information indicating the boundary between the nth range and the (n + 1) th range.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, for each n of n = 1 to N−1, all samples X ^ ′ _Q (L _min ), included in the first range to the n-th range of the modified quantized normalized signal sequence ..., X ^ ' _Q (L _(n) -1) is the sum of squares of the corrected quantized normalized signal sequence X ^' _Q [ω∈ {L _min , ..., L _max }] of all samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L _max ) is not less than n / N of the sum of squares, and the corrected quantized normalized signal sequence Signal sequence X ^ ' _Q (L _min ), ..., X ^' _{Q obtained} by excluding one sample at the highest side of the nth range from all samples included in the first range to the nth range The sum of squares of the values of (L _(n) -2) is the total of the corrected quantized normalized signal sequence X ^ ' _Q [ω∈ {L _min , ..., L _max }] obtained by equation (2). sample _{X ^ 'Q (L min)} , ..., X ^' Q (L max) N content of less than n, L _(n) to the (n + 1) of the range of the sum of the squares of the In seeking the most sample number in the low frequency side. Thus, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is X ′ ′ _Q [ω∈ for each n in the first range, n = 2 to N−1. {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

Specifically, this processing can be realized by, for example, the following. First, the energy pow is obtained by equation (2). Next, for n = 1, it is determined whether or not the condition of Expression (7) is satisfied while i is incremented by 1 from L _min, and 1 is added to i satisfying the condition of Expression (7). Is determined as L ₍₁₎ . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, for each n from n = 2 to N−1, it is determined whether or not the condition of Expression (7) is satisfied while i is increased by 1 from L _{(n−1) in} order, and Expression (7) The process for obtaining L _(n) by adding 1 to i satisfying the condition is repeated in ascending order of n. With the above processing, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _(n) −1}] is determined as the nth range for each n from n = 2 to N−1. .

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _(N−1) ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

[Fourth example of classification processing based on the first standard]
The fourth example of the sorting process based on the first standard corresponds to the above (d). The sorting process of the fourth example is the same method as the sorting process of the third example, except that “sum of squares” in the sorting process of the third example is replaced with “sum of absolute values”. According to the classification process of the fourth example, it is possible to perform the classification process with a smaller calculation processing amount than the classification process of the third example because the square calculation performed in the classification process of the third example can be omitted.

[Fifth example of classification processing based on the first standard]
A fifth example of the sorting process based on the first reference will be described with reference to FIGS. 10, 11, and 12. The classification process of the third example corresponds to the above (e). The division processing of the fifth example is performed using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence, and the first quantized normalized signal sequence of the corrected first sequence. The sum of the number of samples from the range to the nth range is the corrected quantized sum of the squares of the values of all samples included in the first range to the nth range of the corrected quantized normalized signal sequence The maximum number of samples that is equal to or less than n / N of the square sum of the values of all samples of the normalized signal sequence is obtained, and the first range of the corrected quantized normalized signal sequence is changed from the first range. This is a process of dividing the corrected quantized normalized signal sequence into N ranges by setting a range other than the N−1 range as the Nth range of the corrected quantized normalized signal sequence.

[[Example 1 of the fifth example of the sorting process based on the first standard 1: Example of dividing into two ranges]]
FIG. 10 shows an example of dividing into two ranges, that is, an example where N = 2. Quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified after modification is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Then, the number of the index omega discrete frequency L _min quantized normalized signal while increasing in order from the sequence X ^ _'Q square sum p _low from L _min to the index (omega) is p _low ≦ pow / 2 The first range is a discrete frequency obtained by subtracting 1 from the index ω of the discrete frequency when p _low ≦ pow / 2 is not satisfied for the first time, and the index ω is set to the second range. _Is output as the index L _mid , which is the sample number on the lowest side of. As a result, the first range is determined as X ^ ′ _Q [ω∈ {L _min ,..., L _mid −1}].

FIG. 10 is a flowchart for realizing the above processing. The initial value of the discrete frequency index ω is set to L _min , and the initial value of the low frequency energy p _low is set to | X ^ ′ _Q (L _min ) | ² . Then, it is determined whether or not p _low ≦ pow / 2 is satisfied. When satisfying p _low ≦ pow / 2 is the one plus the index of discrete frequency omega as a new omega, 'energy _Q (ω) | X ^' the _{p low X ^ Q (ω)} | ^The sum of ² is taken as the new p _low . If p _low ≦ pow / 2 is not satisfied, the discrete frequency index ω at that time is output as the index L _mid .

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, number of samples L _mid -L _min of the first range The number of samples in the second range may be L _max −L _mid +1, or anything insofar as the information can identify the first range and the second range.

[[Example 2 of the fifth example of the sorting process based on the first standard: Example of sorting into four ranges]]
FIG. 11 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be divided into four ranges, specifically, , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, modified quantized normalized signal sequence _{X ^ 'Q (ω) [} ω∈ {L min, ..., L max}] All samples X ^ _Q (L _{min) of, ..., X ^ Q (L} max ) Is calculated. The sum of squares pow is obtained by equation (2).

Next, the values of all the samples X ^ ' _Q (L _min ), ..., X ^' _Q (L _{b (1)} -1) included in the first range of the corrected quantized normalized signal sequence are The sum of squares represents all samples X ^ ' _Q (L _min ), ... of the modified quantized normalized signal sequence X ^' _Q [ω∈ {L _min , ..., L _max }] obtained by equation (2). , X ^ ' _Q (L _max ) is less than or equal to ¼ of the sum of squares, and all samples included in the first range of the corrected quantized normalized signal sequence have the second range most After correction, the sum of squares of the value of the signal sequence X ^ ' _Q (L _min ), ..., X ^' _Q (L ₍₁₎ ) plus one sample on the low-frequency side is obtained by equation (2) Quantized normalized signal sequence X ^ ' _Q [ω∈ {L _min , ..., L _max }] of all samples X ^' _Q (L _min ), ..., X ^ ' _Q (L _max ) L _(1), which is larger than a quarter, is obtained as the sample number on the lowest side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

In addition, all samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L ₍₂₎ −1) included in the first range to the second range of the corrected quantized normalized signal sequence The sum of squares of the values of all the samples X ^ ′ _Q (L _min ) of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by Equation (2) ), ..., X ^ ' _Q (L _max ) is less than or equal to two-quarters of the sum of squares (ie, half), and the first range of the corrected quantized normalized signal sequence is Signal sequence X ^ ' _Q ( _Lmin ), ..., X ^' _Q (L ₍₂₎ ), which is obtained by adding one sample at the lowest side of the third range to all samples included in the range 2 sum of squares of the values, corrected quantized normalized signal sequence was determined by the equation _{(2) X ^ 'Q [} ω∈ {L min, ..., L max}] of all samples X ^' _Q (L _min ), ..., X ^ _'Q (2 quarters of the sum of the squares of the L _max) (i.e., one-half) is greater than, L ₍₂₎ of the third range Also determined as a sample number in the low frequency side. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

In addition, all samples X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L ₍₃₎ −1) included in the first to third ranges of the corrected quantized normalized signal sequence The sum of squares of the values of all the samples X ^ ′ _Q (L _min ) of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] obtained by Equation (2) ), ..., X ^ ' _Q (L _max ) is less than or equal to three-quarters of the sum of squares, and all the signals included in the first to third ranges of the corrected quantized normalized signal sequence The sum of squares of the values of the signal sequence X ^ ' _Q ( _Lmin ), ..., X ^' _Q (L ₍₃₎ ), which is obtained by adding one sample at the lowest side of the fourth range to the sample, is corrected quantized normalized determined by (2) signal sequence _{X ^ 'Q [ω∈ {L} min, ..., L max}] All samples X ^ _{of' Q (L min), ...} , X ^ 'Q L _(3), which is greater than three-quarters of the sum of squares of (L _max ), is determined as the sample number on the lowest side of the fourth range. Thereby, X ^ ′ _Q [ω∈ {L _{b (2)} ,..., L ₍₃₎ −1}] is determined as the third range.

  Specifically, this processing can be realized by, for example, the following.

  First, the energy pow is obtained by equation (2).

Next, it is determined whether or not the condition of Expression (8) is satisfied while i is incremented by 1 from L _{min in} order, and a value obtained by adding 1 to i satisfying the condition of Expression (8) is L _{(1 )} . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, it is determined whether or not the condition of Expression (9) is satisfied while i is incremented by 1 in order from L _(1), and a value obtained by adding 1 to i satisfying the condition of Expression (9) is L _{( Find} as ₂₎ . Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Furthermore, while increasing the i from L ₍₂₎ one by one in sequence will determine whether or not the condition of formula (10), a material obtained by adding 1 to satisfy the condition i of formula (10) L _{(3 )} . Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L ₍₃₎ ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the fifth example of classification processing based on the first standard: Example of dividing into N ranges]]
FIG. 12 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified into N ranges, specifically, In this example, L _(n) , which is the sample number on the lowest side of the _(n + 1) th range, is determined as information indicating the boundary between the nth range and the (n + 1) th range.

First, all samples X ^ ' _Q (L _min ), ..., X ^' _Q () of the modified quantized normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] determine the sum of squares pow of L _max). The sum of squares pow is obtained by equation (2).

Next, for each n of n = 1 to N−1, all samples X ^ ′ _Q (L _min ), included in the first range to the n-th range of the modified quantized normalized signal sequence ..., X ^ ' _Q (L _(n) -1) is the sum of squares of the quantized normalized signal sequence X ^' _Q [ω∈ {L _min , ..., L _max }] Of all the samples X ^ ' _Q (L _min ),..., X ^' _Q (L _max ) is less than n / N of the sum of squares and is the first of the corrected quantized normalized signal sequence X ^ ′ _Q (L _min ),..., X ^ ′ _Q (L, which is obtained by adding one sample at the lowest side of the (n + 1) th range to all samples included in the nth to nth ranges _(n) ) is the sum of squares of all samples X ^ ′ _Q [ω∈ {L _min ,..., L _max }] of the modified quantized normalized signal sequence X ^ ′ _Q obtained by the equation (2) _{Q (L min), ...,} X ^ 'Q N content greater than that of the n of the sum of the squares of the (L _max), L _{(n) is} of the (n + 1) of the range top Obtaining a sample number in the low frequency side. Thus, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is X ′ ′ _Q [ω∈ for each n in the first range, n = 2 to N−1. {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

Specifically, this processing can be realized by, for example, the following. First, the energy pow is obtained by equation (2). Next, for n = 1, it is determined whether or not the condition of Expression (11) is satisfied while i is incremented by 1 from L _min, and 1 is added to i satisfying the condition of Expression (11). Is determined as L ₍₁₎ . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, for each n from n = 2 to N−1, it is determined whether or not the condition of equation (11) is satisfied while i is incremented by 1 sequentially from L _(n−1 ). The process for obtaining L _(n) by adding 1 to i satisfying the condition is repeated in ascending order of n. With the above processing, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _(n) −1}] is determined as the nth range for each n from n = 2 to N−1. .

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _(N−1) ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

[Sixth example of classification processing based on the first standard]
The sixth example of the sorting process corresponds to the above (f). The sorting process of the sixth example is the same method as the sorting process of the fifth example, except that “sum of squares” in the sorting process of the fifth example is replaced with “sum of absolute values”. According to the sorting process of the sixth example, it is possible to perform the sorting process with a smaller calculation processing amount than the sorting process of the fifth example, because the square calculation performed in the sorting process of the fifth example can be omitted.

  Next, the classification process according to the second standard “standard for classifying so that the number of significant samples included in each range is as equal as possible” will be described.

For example, the classification process in the “criteria for classifying so that the number of significant samples included in each range is as equal as possible” is performed by, for example, the nth range (n is 1 to N−) of the corrected quantized normalized signal sequence. Each integer up to 1)
(a) out of all samples included in the first range to the n-th range of the corrected quantized normalized signal sequence, the number of samples whose sample energy is greater than or equal to a predetermined value; , N of N samples of the number of samples whose sample energy is greater than or equal to or greater than a predetermined value among all samples included in the corrected quantized normalized signal sequence,
Or
(b) The number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the n-th range of the corrected quantized normalized signal sequence And n / N of the number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the corrected quantized normalized signal sequence is closest to each other. ,
Or
(c) The number of samples whose sample energy is greater than or equal to a predetermined value among all samples included in the first range to the n-th range of the corrected quantized normalized signal sequence is , Among all the samples included in the corrected quantized normalized signal sequence, the sample number is the minimum number of samples that is n or more of N / N of the number of samples that are greater than or equal to a predetermined value. like,
Or
(d) Number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected quantized normalized signal sequence Is the minimum number of samples in which the absolute value of the samples among all the samples included in the corrected quantized normalized signal sequence is greater than or equal to a predetermined value and not less than n / N of the number of samples So that
Or
(e) The number of samples whose sample energy is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected quantized normalized signal sequence is , Among all the samples included in the post-correction quantized normalized signal sequence, the maximum number of samples is equal to or less than n / N of the number of samples whose sample energy is greater than a predetermined value or greater than a predetermined value. like,
Or
(f) Number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected quantized normalized signal sequence Is the maximum number of samples in which the absolute value of the samples among all the samples included in the corrected quantized normalized signal sequence is less than or equal to n / N of the number of samples that are greater than or equal to the predetermined value So that
Seeking
A range other than the first range to the (N−1) th range in the corrected quantized normalized signal sequence is set as the Nth range of the corrected quantized normalized signal sequence. This is done by dividing the digitized signal sequence into N ranges.

  The classification process exemplified above realizes the classification based on the “criteria for classifying so that the number of significant samples included in each range is as equal as possible” by a method of sequentially determining each range. . According to the classification process exemplified above, it is possible to realize classification according to “a criterion for classifying so that the number of significant samples included in each range is as equal as possible” with a small amount of calculation processing.

[First example of classification processing based on the second standard]
A first example of the sorting process based on the second reference will be described with reference to FIGS. 13, 14, and 15. The sorting process of the first example corresponds to the above (a). The division processing of the first example is performed using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence, and the first quantized normalized signal sequence of the corrected first sequence. The number of samples whose sample energy is greater than or equal to a predetermined value or more than all the samples included in the range to the nth range, and all samples included in the corrected quantized normalized signal sequence The first range of the corrected quantized normalized signal sequence is determined such that n of the number of samples whose sample energy is greater than or equal to or greater than a predetermined value is closest to N. To a range other than the (N−1) -th range as the Nth range of the corrected quantized normalized signal sequence, so that the corrected quantized normalized signal sequence is divided into N ranges. is there.

[[Example 1 of the first example of the sorting process based on the second standard 1: Example of dividing into two ranges]]
FIG. 13 shows an example of dividing into two ranges, that is, an example where N = 2. The modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, f _count (ω) is determined for each index ω by the equation (B2). The f _count (ω) for each index ω includes a sample corresponding to the index ω of the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }]. energy _{| X ^ 'Q (ω)} | 2 is for large samples than the predetermined value "sample energy | X ^' | set 1 as information representing that _Q (omega) is larger than the predetermined value", it 0 is set as information indicating that “the sample energy | X ^ ′ _Q (ω) | is not larger than a predetermined value” for samples other than. In this example, the predetermined value is arbitrarily set to a minute amount ε (ε is a value of 0 or more).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value Half the number f _count (L _min ) +... + F _count (L _max ) and all samples X ^ ′ _Q (ω) included in the first range of the corrected quantized normalized signal sequence Among [ω∈ {L _min , ..., L _mid -1}], the difference between the number of samples whose sample energy is larger than a predetermined value f _count (L _min ) + ... + f _count (L _mid -1) is the smallest _Lmid , which is the sample number on the lowest side of the second range, is obtained so that That is, L _mid is obtained by the equation (B3). Accordingly, the first range is determined as X ^ ′ _Q [ω∈ {L _min ,..., L _mid −1}].

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, sample number of the first range L _mid -1-L _It may be _min +1, may be the number of samples in the second range L _max -L _mid +1, or anything insofar as it is information that can identify the first range and the second range. Good.

[[Example 2 of first example of sorting process based on second standard: Example of sorting into four ranges]]
FIG. 14 shows an example in which the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is divided into four ranges, specifically , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) and all samples X ^ ′ _Q (ω included in the first range of the corrected quantized normalized signal sequence ) The number of samples of [ω∈ {L _min , ..., L ₍₁₎ -1}] whose sample energy is greater than a predetermined value f _count (L _min ) + ... + f _count (L ₍₁₎ -1) L ₍₁₎ obtained so as to minimize the difference between is the sample number on the lowest side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Also, the number of samples whose sample energy is greater than a predetermined value among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence From f _count (L _min ) +... + f _count (L _max ) to two- _fourths (ie, half) and from the first range to the second range of the corrected quantized normalized signal sequence L samples of energy of all the samples included are determined as the difference between the number of larger samples than the predetermined value _{f count (L min) + ...} + f count (L (2) -1) is minimum _{Let (2)} be the sample number on the lowest side of the third range. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Also, the number of samples whose sample energy is greater than a predetermined value among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence f _count (L _min ) +... + f _count (L _max ) of three quarters and all samples included in the first range to the third range of the corrected quantized normalized signal sequence The number of samples whose sample energy is greater than a predetermined value f _count (L _min ) +… + f _count (L ₍₃₎ −1) L ₍₃₎ obtained so as to minimize the difference from the fourth range The sample number on the lowest side of Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L ₍₃₎ ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the first example of classification processing based on the second criterion: Example of dividing into N ranges]]
FIG. 15 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified into N ranges, specifically In this example, L _(n) , which is the sample number on the lowest side of the _(n + 1) th range, is determined as information indicating the boundary between the nth range and the (n + 1) th range. In the following description, L _min is assumed to be L ₍₀₎ .

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, for each n from n = 1 to N−1, all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max } included in the corrected quantized normalized signal sequence ] Of the number of samples f _count (L _min ) +... + F _count (L _max ) of samples whose sample energy is greater than a predetermined value, and the first range of the corrected quantized normalized signal sequence To the nth range of samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _(n) −1}], the number f of samples whose sample energy is greater than a predetermined value f L _(n) obtained so that the difference from _count (L _min ) +... + f _count (L _(n) −1) is minimized is set to the sample number on the lowest side of the ₍ n + 1 ₎ th range. Thereby, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _(N−1) ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

  Note that “the number of samples whose sample energy is equal to or greater than a predetermined value among all samples included in the first range to the nth range of the post-correction quantization normalized signal sequence, and the post-correction quantization normalization “N = 1 to N−1” so that n / N of the number of samples whose sample energy is equal to or greater than a predetermined value among all the samples included in the digitized signal sequence In order to define the n-th range, “<” in the formula (B2) may be replaced with “≦”.

[Second example of classification processing based on the second standard]
A second example of the sorting process based on the second standard corresponds to the above (b). The sorting process of the second example is the same method as the sorting process of the first example, except that “sum of squares” in the sorting process of the first example is replaced with “sum of absolute values”. According to the sorting process of the second example, it is possible to perform the sorting process with a smaller amount of calculation processing than the sorting process of the first example because the square calculation performed in the sorting process of the first example can be omitted.

[Third example of sorting process based on second standard]
A third example of the sorting process based on the second reference will be described with reference to FIGS. 16, 17, and 18. FIG. The classification process of the third example corresponds to the above (c). The division processing of the third example is performed by using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence and the first quantized normalized signal sequence of the corrected first sequence. All samples in which the number of samples whose sample energy is greater than or equal to or greater than or equal to a predetermined value among all samples included in the range to the nth range is included in the corrected quantized normalized signal sequence Out of the corrected quantized normalized signal sequence, the sample energy is determined to be a minimum number of samples equal to or greater than n / N of the number of samples whose energy is greater than or equal to a predetermined value. The range other than the first range to the (N−1) th range is set as the Nth range of the corrected quantized normalized signal sequence, so that the corrected quantized normalized signal sequence is reduced to N ranges. It is a process to classify.

[[Specific example 1 of the third example of the sorting process based on the second standard 1: Example of dividing into two ranges]]
FIG. 16 shows an example of dividing into two ranges, that is, an example where N = 2. The modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, the energy of the samples of all the samples included a number k of the index ω discrete frequency corrected quantized normalized signal sequence from L _min from L _min to the index k, while increasing the order of the predetermined value It is determined whether or not the number of large samples f _count (L _min ) + ... + f _count (k) is (f _count (L _min ) + ... + f _count (L _max )) / 2 or more. _count (L _min ) +… + f _count (k) is the first range up to index k of the discrete frequency where (f _count (L _min ) +… + f _count (L _max )) / 2 or more A value obtained by adding 1 to the index k is output as an index L _mid which is a sample number on the lowest side of the second range. Accordingly, the first range is determined as X ^ ′ _Q [ω∈ {L _min ,..., L _mid −1}].

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, number of samples L _mid -L _min of the first range The number of samples in the second range may be L _max −L _mid +1, or anything insofar as the information can identify the first range and the second range.

[[Specific example 2 of the third example of classification processing based on the second standard: Example of dividing into four ranges]]
FIG. 17 shows an example in which the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is divided into four ranges, specifically , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] included in the first range of the corrected quantized normalized signal sequence The number of samples f _count (L _min ) +… + f _count (L ₍₁₎ -1) whose energy is greater than a predetermined value is all samples X ^ ' _Q included in the corrected quantized normalized signal sequence _{(ω) [ω∈ {L min} , ..., L max}] number f _count (L _min) of the sample energy is greater than a predetermined value of a sample of the _{+ ... + f count (L max} ) 4 over one of And all the samples included in the signal sequence excluding the one sample at the highest frequency side of the first range from all the samples included in the first range of the corrected quantized normalized signal sequence. Number of samples in which the sample energy is greater than a predetermined value among the samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₁₎ -2}] f _count (L _min ) + ... + f _count ( L _{(1) -2)} is All samples X ^ _'Q contained in the corrected quantized normalized signal sequence _{(ω) [ω∈ {L min} , ..., L max}] number f _count energy sample is larger samples than the predetermined value among the L _(1), which is smaller than a quarter of (L _min ) +... + F _count (L _max ), is determined as the sample number on the lowest side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Next, all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₂₎ −1}] included in the first and second ranges of the corrected quantized normalized signal sequence samples of energy is large sample number f _count than a predetermined value among the _{(L min) + ... + f} count (L (2) -1) , all the samples X contained in the revised quantization normalized signal sequence ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] 2 minutes, the number of samples f _count (L _min ) + ... + f _count (L _max ) And a signal obtained by excluding one sample at the highest side of the second range from all samples included in the first and second ranges of the corrected quantized normalized signal sequence Number of samples f _count (L _min ) of all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₂₎ −2}] included in the sequence whose sample energy is greater than a predetermined value + ... + f _{co unt} (L ₍₂₎ -2) is the sample of all samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] included in the modified quantized normalized signal sequence Number of samples whose energy is greater than a predetermined value f _count (L _min ) + ... + f _count (L _max ) less than one half, L ₍₂₎ is the sample number at the lowest side of the third range Asking. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Next, all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₃₎ − included in the first, second, and third ranges of the corrected quantized normalized signal sequence 1}], the number of samples f _count (L _min ) +... + F _count (L ₍₃₎ −1) in which the sample energy is greater than the predetermined value is included in the corrected quantized normalized signal sequence. Of samples X ^ ' _Q (ω) [ω∈ {L _min ,..., L _max }], the number of samples whose sample energy is greater than a predetermined value f _count (L _min ) +... + F _count (L _max ) 1 that is equal to or higher than 4/3 of all the samples included in the first, second, and third ranges of the corrected quantized normalized signal sequence and that is the highest in the third range Number of samples whose sample energy is greater than a predetermined value among all samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₃₎ -2}] included in the signal sequence excluding two samples f _count ( L _min ) + ... + f _count (L ₍₃₎ -2) is the sum of all samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L included in the modified quantized normalized signal sequence _max}] of sample energy number of larger samples than the predetermined value _{f count (L min) + ...} + f count 3 min 4 smaller than (L _max), most L a ₍₃₎ of the fourth range Obtained as the sample number on the low frequency side. Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

  Specifically, these processes can be realized by, for example, the following.

First, f _count (ω) is determined for each index ω by the equation (B2). Then, let f _count (L _min ) +... + F _count (L _max ) be F.

Next, it is determined whether or not the formula (B4) is satisfied while i is incremented by 1 sequentially from L _min, and a value obtained by adding 1 to i satisfying the formula (B4) is obtained as L ₍₁₎ . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, it is determined whether or not the formula (B5) is satisfied while increasing i sequentially by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B5) is obtained as L ₍₂₎ . Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Further, it is determined whether or not the formula (B6) is satisfied while increasing i sequentially by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B6) is obtained as L ₍₃₎ . Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L ₍₃₎ ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the third example of classification processing based on the second criterion: Example of dividing into N ranges]]
FIG. 18 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be divided into N ranges, specifically, In this example, L _(n) , which is the sample number on the lowest side of the _(n + 1) th range, is determined as information indicating the boundary between the nth range and the (n + 1) th range.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, for each n from n = 1 to N−1, all samples X ^ ′ _Q (ω) [ω included in the first range to the n-th range of the corrected quantized normalized signal sequence ∈ {L _min , ..., L _(n) -1}], the number of samples f _count (L _min ) + ... + f _count (L _(n) -1) is corrected. The number of samples f _count (where the sample energy is greater than a predetermined value among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the post-quantization normalized signal sequence f _count ( L _min ) +... + F _count (L _max ) n times N or more, and the first through the nth range of the corrected quantized normalized signal sequence The energy of the sample in the signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _(n) −2}] excluding one sample at the highest frequency in the range of n is a predetermined value. Greater than Sample number _{f count (L min) + ...} + f count (L (n) -2) are all included in the corrected quantized normalized signal sequence samples _{X ^ 'Q (ω) [} ω∈ { _{L min, ..., L max}} ] samples of energy number of larger samples than the predetermined value f _count of _{(L min) + ... + f} count n content less than the n of (L _max), L _{(n) is} Obtained as the sample number at the lowest side of the (n + 1) th range. Thus, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is X ′ ′ _Q [ω∈ for each n in the first range, n = 2 to N−1. {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

  Specifically, this processing can be realized by, for example, the following.

First, f _count (ω) is determined for each index ω by the equation (B2). Then, let f _count (L _min ) +... + F _count (L _max ) be F.

Next, for n = 1, it is determined whether or not the formula (B7) is satisfied while increasing i sequentially by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B7) is represented by L _{(1 )} . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, for each n from n = 2 to N−1, it is determined whether or not Expression (B7) is satisfied while i is increased by 1 sequentially from L _(n−1) , and Expression (B7) is satisfied. The process of obtaining the value obtained by adding 1 to i as L _(n) is repeated in the order of increasing n. With the above processing, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _{(n) −1} }] is determined as the nth range for each n from n = 2 to N−1. .

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _(n) ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

[Fourth example of classification processing based on the second standard]
The fourth example of the sorting process based on the second standard corresponds to the above (d). The sorting process of the fourth example is the same method as the sorting process of the third example, except that “sum of squares” in the sorting process of the third example is replaced with “sum of absolute values”. According to the classification process of the fourth example, it is possible to perform the classification process with a smaller calculation processing amount than the classification process of the third example because the square calculation performed in the classification process of the third example can be omitted.

[Fifth example of sorting process based on second criterion]
A fifth example of the sorting process based on the second reference will be described with reference to FIGS. The classification process of the fifth example corresponds to the above (e). The division processing of the fifth example is performed using the nth range (n is an integer from 1 to N−1) of the corrected quantized normalized signal sequence, and the first quantized normalized signal sequence of the corrected first sequence. All samples in which the number of samples whose sample energy is greater than or equal to or greater than or equal to a predetermined value among all samples included in the range to the nth range is included in the corrected quantized normalized signal sequence Out of the corrected quantized normalized signal sequences, the energy of the samples is determined to be the maximum number of samples that is less than n / N of the number of samples that is greater than or equal to the predetermined value. The range other than the first range to the (N−1) th range is set as the Nth range of the corrected quantized normalized signal sequence, so that the corrected quantized normalized signal sequence is reduced to N ranges. It is a process to classify.

[[Specific example of the fifth example of the sorting process based on the second standard 1: Example of dividing into two ranges]]
FIG. 19 shows an example of dividing into two ranges, that is, an example where N = 2. The modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is changed to X ^ ′ _Q (ω) [ω∈ {L _min ,. _mid −1}] and X ^ ′ _Q (ω) [ω∈ {L _mid ,..., L _max }] An example of dividing into two ranges, specifically, the first range is the low range and the second range. This is an example in which _Lmid , which is the sample number on the lowest frequency side of the second range, is determined as information representing the boundary with the high frequency range of 2.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, the number k of the discrete frequency index ω is increased from L _{min in} order, and the sample energy is greater than a predetermined value among all samples included in the quantized normalized signal sequence from L _min to the index k. number _{f count (L min) + ...} + f count (k) is _{(f count (L min) +} ... + f count (L max)) / 2 than it is determined whether or not large, for the first time f _count (L _min ) + ... + f _count (k) is k-1 less than the index k of the discrete frequency where (f _count (L _min ) + ... + f _count (L _max )) / 2 is greater than The index k is output as the index L _mid which is the sample number on the lowest side of the second range. As a result, the first range is determined as X ^ ′ _Q [ω∈ {L _min ,..., L _mid −1}].

Then, a range other than the first range of the modified quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }], that is, X ^ ′ _Q [ω∈ {L _mid , ..., L _max }] is the second range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into two ranges.

Sorting information identifying circuit 150 outputs may be the L _mid, may be a value obtained by calculating the predetermined value in the L _mid, number of samples L _mid -L _min of the first range The number of samples in the second range may be L _max −L _mid +1, or anything insofar as the information can identify the first range and the second range.

[[Example 2 of the fifth example of the sorting process based on the second standard: Example of sorting into four ranges]]
FIG. 20 shows an example in which the corrected quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be classified is divided into four ranges, specifically , L ₍₁₎ , which is the sample number on the lowest side of the second range, is determined as information representing the boundary between the first range and the second range, and the second range and the third range are determined. L ₍₂₎ which is the sample number on the lowest side of the third range is determined as information indicating the boundary of the third range, and information of the fourth range is determined as information indicating the boundary between the third range and the fourth range. This is an example of determining L ₍₃₎ which is the sample number on the lowest side.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] included in the first range of the corrected quantized normalized signal sequence The number of samples f _count (L _min ) +… + f _count (L ₍₁₎ -1) whose energy is greater than a predetermined value is all samples X ^ ' _Q included in the corrected quantized normalized signal sequence _{(ω) [ω∈ {L min} , ..., L max}] number f _count (L _min) of the energy is larger samples than a predetermined value of a sample of the + ... + than a quarter of the f _count (L _max) All samples included in the signal sequence that are large and exclude all samples included in the first range of the corrected quantized normalized signal sequence from one sample located on the highest side of the first range The number of samples f _count (L _min ) +... + F _count (L) in which X ^ ' _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −2}] ₍₁₎ -2) is All samples X ^ _'Q contained in the corrected quantized normalized signal sequence _{(ω) [ω∈ {L min} , ..., L max}] number f _count energy sample is larger samples than the predetermined value among the (L _min ) +... + F _count (L _max ), which is equal to or less than a quarter, L ₍₁₎ is obtained as the sample number on the lowest side of the second range. Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Next, all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₂₎ −1}] included in the first and second ranges of the corrected quantized normalized signal sequence samples of energy is large sample number f _count than a predetermined value among the _{(L min) + ... + f} count (L (2) -1) , all the samples X contained in the revised quantization normalized signal sequence ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] 2 minutes, the number of samples f _count (L _min ) + ... + f _count (L _max ) And a signal sequence obtained by excluding one sample at the highest frequency side of the second range from all samples included in the first and second ranges of the corrected quantized normalized signal sequence Of all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₂₎ −2}] included in the sample f _count (L _min ) + … + F _{co unt} (L ₍₂₎ -2) is the sample of all samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max }] included in the modified quantized normalized signal sequence Number of samples whose energy is greater than a predetermined value f _count (L _min ) + ... + f _count (L _max ) or less, L ₍₂₎ is the lowest sample in the third range Ask as a number. Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Next, all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₃₎ − included in the first, second, and third ranges of the corrected quantized normalized signal sequence 1}], the number of samples f _count (L _min ) +... + F _count (L ₍₃₎ −1) in which the sample energy is greater than the predetermined value is included in the corrected quantized normalized signal sequence. Of samples X ^ ' _Q (ω) [ω∈ {L _min ,..., L _max }], the number of samples whose sample energy is greater than a predetermined value f _count (L _min ) +... + F _count (L _max ) One of the highest frequencies in the third range from all samples included in the first, second, and third ranges of the corrected quantized normalized signal sequence. The number f of samples whose sample energy is greater than a predetermined value among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₃₎ −2}] included in the signal sequence excluding the samples f _count ( L _min ) + ... + f _count (L ₍₃₎ -2) is the sum of all samples X ^ ' _Q (ω) [ω∈ {L _min , ..., L included in the modified quantized normalized signal sequence _max}] energy of the sample becomes 4 or less thirds greater than the predetermined value sample number _{f count (L min) + ...} + f count (L max) of the, L ₍₃₎ a fourth range Obtained as the sample number on the lowest side. Thereby, X ^ ′ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

  Specifically, this processing can be realized by, for example, the following.

First, f _count (ω) is determined for each index ω by the equation (B2). Then, let f _count (L _min ) +... + F _count (L _max ) be F.

Next, it is determined whether or not the formula (B8) is satisfied while increasing i sequentially by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B8) is obtained as L ₍₁₎ . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, it is determined whether or not the formula (B9) is satisfied while i is incremented by 1 sequentially from L _min, and a value obtained by adding 1 to i satisfying the formula (B9) is obtained as L ₍₂₎ . Thereby, X ^ ′ _Q [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] is determined as the second range.

Further, it is determined whether or not the formula (B10) is satisfied while i is sequentially increased by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B10) is obtained as L ₍₃₎ . Thereby, X ^ _Q [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] is determined as the third range.

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is a range other than the first range to the third range, that is, X ^ ′ _Q [ω Let ∈ {L _{b (3)} ,..., L _max }] be the fourth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into four ranges.

Sorting information identifying circuit 150 outputs may be a L ₍₁₎ and L ₍₂₎ and L _(3), predetermined to each of the L ₍₁₎ and L ₍₂₎ and L ₍₃₎ The calculated value may be the number of samples in each range, or anything insofar as it is information that can identify all four ranges.

[[Generalization of the fifth example of classification processing based on the second standard: Example of dividing into N ranges]]
FIG. 21 shows an example of dividing the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] to be divided into N ranges, specifically, In this example, L _{b (n)} which is the sample number on the lowest side of the _(n + 1) th range is determined as information indicating the boundary between the nth range and the (n + 1) th range.

First, f _count (ω) is determined for each index ω by the equation (B2).

Next, among all the samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the corrected quantized normalized signal sequence, the samples whose energy is greater than a predetermined value The number f _count (L _min ) +... + F _count (L _max ) is obtained.

Next, for each n from n = 1 to N−1, all samples X ^ ′ _Q (ω) [ω included in the first range to the n-th range of the corrected quantized normalized signal sequence ∈ {L _min , ..., L _(n) -1}], the number of samples f _count (L _min ) + ... + f _count (L _(n) -1) is corrected. The number of samples f _count (where the sample energy is greater than a predetermined value among all samples X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] included in the post-quantization normalized signal sequence f _count ( L _min ) +... + F _count (L _max ) n greater than n of N and included in all the samples included in the first range to the nth range of the corrected quantized normalized signal sequence. Of the signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _(n) −2}] excluding one sample on the highest side of the range of large Sample number _{f count (L min) + ...} + f count (L (n) -2) are all included in the corrected quantized normalized signal sequence samples _{X ^ 'Q (ω) [} ω∈ { L _min, ..., L _max} number f _count (L _min) of the sample energy is greater than a predetermined value of a sample of] + ... + f n content becomes n following the _{count (L max), L (} n) As the sample number on the lowest side of the (n + 1) th range. Thus, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is X ′ ′ _Q [ω∈ for each n in the first range, n = 2 to N−1. {L _(n−1) ,..., L _(n) −1}] is determined as the nth range.

  Specifically, this processing can be realized by, for example, the following.

First, f _count (ω) is determined for each index ω by the equation (B2). Then, let f _count (L _min ) +... + F _count (L _max ) be F.

Next, for n = 1, it is determined whether or not the formula (B11) is satisfied while increasing i sequentially by 1 from L _min, and a value obtained by adding 1 to i satisfying the formula (B11) is represented by L _{(1 )} . Thereby, X ^ ′ _Q [ω∈ {L _min ,..., L ₍₁₎ −1}] is determined as the first range.

Further, for each n from n = 2 to N−1, it is determined whether or not the equation (B11) is satisfied while increasing i by ₁ from L _{(n−1) in} order, and the equation (B11) is satisfied. The process of obtaining the value obtained by adding 1 to i as L _(n) is repeated in the order of increasing n. With the above processing, X ^ ′ _Q [ω∈ {L _(n−1) ,..., L _{(n) −1} }] is determined as the nth range for each n from n = 2 to N−1. .

Then, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is in a range other than the first range to the (N−1) th range, that is, X ^ ′ _Q Let [ω∈ {L _{b (n)} ,..., L _max }] be the Nth range.

As described above, the corrected quantized normalized signal sequence X ^ ′ _Q [ω∈ {L _min ,..., L _max }] is divided into N ranges.

Sorting information identifying circuit 150 is output, L _(n) (n is the integer from 1 to N-1) may be a, L _(n) (each n is an integer from 1 to N-1 ) May be a value obtained by calculating a predetermined value, may be the number of samples in each range, or may be anything as long as it is information that can specify all N ranges.

[Sixth example of classification processing based on the second standard]
The sixth example of the sorting process based on the second standard corresponds to the above (f). The sorting process of the sixth example is the same method as the sorting process of the fifth example, except that “sum of squares” in the sorting process of the fifth example is replaced with “sum of absolute values”. According to the sorting process of the sixth example, it is possible to perform the sorting process with a smaller calculation processing amount than the sorting process of the fifth example, because the square calculation performed in the sorting process of the fifth example can be omitted.

  In addition, it is not limited to the implementation structure which transmits a bit stream from the encoding apparatus 1 to the decoding apparatus 2, For example, the information obtained by the synthesis | combination part 160 is recorded on a recording medium, and the said information read from the said recording medium Is also allowed to be input to the decoding device 2.

  The decoding apparatus 2 (see FIG. 22) according to the first embodiment includes a normalized signal decoding unit 107, a global gain decoding unit 106, a control unit 205, a modified code separation unit 220, an error decoding unit 230, an addition unit 240, and a decoded signal sequence. A generation unit 250 and a sorting unit 260 are included. The decoding device 2 may include a separation unit 210 and a time domain conversion unit 270 as necessary.

  Hereinafter, the process in the decoding apparatus 2 (decoder) is demonstrated (refer FIG. 23).

  The bit stream transmitted from the encoding device 1 is input to the decoding device 2. The separation unit 210 extracts the normalized signal code, the global gain code, and the correction code from the bit stream.

<Normalized signal decoding unit 107>
A normalized signal code is input to the normalized signal decoding unit 107. The normalized signal decoding unit 107 applies a decoding method corresponding to the encoding method performed by the normalized signal encoding unit 120 of the encoding device 1 to decode the normalized signal code and decode the normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is obtained (step S1d). In this example, in order to explain corresponding to the encoding device 1, ω represents an index of discrete frequency, and each component of the discrete frequency at point L is represented by ω = L _min to L _max respectively. To do. The normalized signal decoding unit 107 performs the same operation as the normalized signal decoding unit 107 of FIG. 1 described in the “Background Art” column.

For example, the normalized signal decoding unit 107 first decodes the rough shape (normalized power spectrum envelope value) and the fine structure of the normalized signal sequence X _Q (ω) from the code included in the normalized signal code. To do. Then, the normalized signal decoding unit 107 obtains a decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] from the normalized power spectrum envelope value and the fine structure. .

<Global Gain Decoding Unit 106>
A global gain code is input to the global gain decoding unit 106. The global gain decoding unit 160 decodes the global gain code and outputs a decoded global gain g ^ (step S2d). The decoding process performed by the global gain decoding unit 106 is a process corresponding to the encoding process performed by the global gain encoding unit 105, and the global gain decoding unit in the [Background Technology] column. This is a well-known technique as described in 106.

<Control unit 205>
The control unit 205 receives the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. The control unit 205 determines the magnitude of the sample value of the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] according to the number of samples larger than a predetermined value. , Correction of decoding global gain g ^ based on gain correction amount code and correction normalized decoding signal sequence X ^ _Q (ω) [ω∈ {L _min , ..., L _max }] One of the processes to be performed is prioritized over the other process (step S3d). The predetermined value is 0, for example.

  The process of correcting the quantized global gain g ^ is the process of steps S7d and S8d by the sorting unit 260 and the decoded signal sequence generation unit 250 described later.

The process of correcting the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is the process of steps S5d and S6d by the error decoding unit 230 and the adding unit 240 described later. That is.

The number of samples whose value of the sample of the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is larger than a predetermined value is the so-called decoding normalized This is the number of non-zero samples of the signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }]. A non-zero sample of the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is also referred to as a non-zero coefficient.

The processing performed by the control unit 205 on the decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] is quantized and normalized by the control unit 170 of the encoding device 1. This is similar to the processing performed for the signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }].

That is, the control unit 205 determines the gain correction bit number U _g and the error expression bit number U _XQ by performing the process of FIG. 26, for example. The control unit 205 transmits the gain correction bit number U _g and the error expression bit number U _XQ to the correction code separation unit 220.

For example, in this way, the control unit 205 determines priority information that is information about which process is to be preferentially performed, and transmits the priority information to the correction code separation unit 220. In the above example, the priority information is the number of gain correction bits U _g and the number of error expression bits U _XQ .

<Modified code separation unit 220>
Priority information is input to the correction code separation unit 220. The correction code separation unit 220 separates the error code and the gain correction amount code from the correction code based on the priority information (step S4d).

The decoded normalized signal sequence X ^ _Q (ω) [ω∈ {L _min , ..., L _max }] is converted into the quantized normalized signal series X ^ _Q (ω) [ω∈ {L _min , ..., L _max }]. For this reason, the priority information determined by the control unit 170 of the encoding device 1 matches the priority information determined by the control unit 205. In other words, the values of U _g and U _XQ determined by the control unit 170 of the encoding device 1 coincide with the values of U _g and U _XQ determined by the control unit 205 of the decoding device 2. Therefore, the correction code separation unit 220 can separate the error code of U _XQ bits and the gain correction amount code of U _g bits from the correction code.

  The error code is transmitted to the error decoding unit 230, and the gain correction amount code is transmitted to the decoded signal sequence generation unit 250.

<Error decoding unit 230>
An error code is input to the error decoding unit 230. The error decoding unit 230 obtains a decoded quantization error q (ω) by decoding the error code by a decoding method corresponding to the error encoding unit 180 of the encoding device 1 (step S5d). Since the sample and the procedure are determined by the encoding device 1 and the decoding device 2 for each error expression bit number by code and decoding, it can be uniquely decoded.

<Specific example 1 of error decoding> (corresponding to <specific example 1 of error encoding> of encoding apparatus 1)
A code book for each possible value of the error expression bit number is stored in advance in the code book storage unit in the error decoding unit 230. In each codebook, a vector having the same number of samples as the number of decoded quantization error sequences that can be expressed by the number of error expression bits corresponding to each codebook and a code corresponding to the vector are associated in advance. Stored.

  After calculating the error expression bit number, the error decoding unit 230 selects a code book corresponding to the calculated error expression bit number from the code book stored in the code book storage unit, and uses the selected code book Perform vector inverse quantization. The decoding process after selecting the codebook is the same as general vector inverse quantization. That is, out of each vector of the selected codebook, a vector corresponding to the input error code is output as a sequence of decoded quantization error q (ω).

  In the above description, the vector stored in the codebook has the same number of samples as the sequence of the decoded quantization error, but the number of samples of the vector stored in the codebook is the integer number of the sequence of the decoded quantization error. It is also possible to perform vector inverse quantization on each of a plurality of codes included in an error code input for each of a plurality of portions of a decoding quantization error sequence.

<Specific Example 2 of Error Decoding Unit 230> (corresponding to <Specific Example 2 of Error Coding> of Encoding Device 1)
The number of error expression bits is U _XQ , the number of samples of the decoded normalized signal sequence X ^ _Q (ω) output by the decoding unit 21 is not 0, the number of samples is T, and the decoded normalized signal sequence X output by the decoding unit 21 When the number of samples whose value of ^ _Q (ω) is 0 is S, the following decoding procedure is preferable.

(A) U If error decoding unit 230 of the _XQ ≦ T, of the T sample value of the decoded normalized signal sequence X ^ _Q (ω) is not 0, U from having a large corresponding power spectral envelope values _XQ For each selected sample, 1-bit code included in the input error code is decoded to obtain positive / negative information of the sample, and the obtained positive / negative information is reconstructed. The reconstructed value +0.25 or −0.25 obtained by giving the absolute value 0.25 is output as a decoded quantization error q (ω) corresponding to the decoded normalized signal sequence X ^ _Q (ω). If the corresponding power spectrum envelope values are the same, for example, select according to a predetermined rule such as selecting a quantization error sample with a smaller position on the frequency axis (a quantization error sample with a lower frequency). To do. For example, a rule corresponding to the rule used in the error encoding unit 180 of the encoding device 1 is held in the error decoding unit 230 in advance.

  As a power spectrum envelope value, a value obtained by denormalizing the normalized power spectrum envelope value decoded inside the normalized signal decoding unit 107 using the decoded global gain g ^ decoded by the global gain decoding unit 106 is obtained. Can be used. Of course, values obtained by other methods may be used. For example, the power spectrum envelope value may be a value representing the power spectrum of the decoded linear prediction coefficient decoded from the code of the quantized linear prediction coefficient.

  The error decoding unit 230 also outputs 0 as a decoded quantization error q (ω) corresponding to each sample for each sample that has not been selected.

(B) When T <U _XQ ≦ = T + S The error decoding unit 230 is included in the input error code for samples whose decoding normalized signal sequence X ^ _Q (ω) is not 0. Decode the 1-bit code to obtain the positive / negative information of the decoded quantization error sample, and give the obtained positive / negative information to the absolute value 0.25 of the reconstructed value, the reconstructed value +0.25 or −0.25, A decoded quantization error q (ω) corresponding to the decoded normalized signal sequence X ^ _Q (ω) is output.

The error decoding unit 230 is also input with respect to each of U _XQ -T samples having a large power spectrum envelope value from among samples whose decoding normalized signal sequence X ^ _Q (ω) is 0. By decoding the 1-bit code included in the error code, the positive and negative information of the decoded quantization error sample is obtained, and the obtained positive and negative information is converted into a reconstructed value that is a predetermined positive value smaller than 0.25. The reconstruction value + A or −A obtained by giving the absolute value A is output as a decoded quantization error q (ω) corresponding to the decoded normalized signal sequence X ^ _Q (ω).

_{Alternatively,} U _XQ -T bits included in the error code for a plurality of samples having a corresponding power spectrum envelope value among samples whose decoding normalized signal sequence X ^ _Q (ω) is 0 Vector is dequantized to obtain a sequence of corresponding decoded quantization error sample values, and each obtained decoded quantization error sample value corresponds to its decoded normalized signal sequence X ^ _Q (ω) Is output as a decoded quantization error q (ω).

  Also, the error decoding unit 230 outputs 0 as a decoded quantization error q (ω) corresponding to each sample for each sample to which no sign representing positive / negative information is assigned.

Thus, the absolute value of the reconstructed value when the value of the quantized normalized signal sequence X ^ _Q (ω) and the value of the decoded normalized signal sequence X ^ _Q (ω) is not 0 is set to 0.25, for example. The absolute value of the reconstructed value when the value of the quantized normalized signal sequence X ^ _Q (ω) and the decoded normalized signal sequence X ^ _Q (ω) is 0 is A (0 <A <0.25). To do. The absolute values of these reconstructed values are examples, and reconstruction is performed when the value of the quantized normalized signal sequence X ^ _Q (ω) and the value of the decoded normalized signal sequence X ^ _Q (ω) are not zero. The absolute value of the value is greater than the absolute value of the reconstructed value when the value of the quantized normalized signal sequence X ^ _Q (ω) and the value of the decoded normalized signal sequence X ^ _Q (ω) are zero. It only needs to be large.

  When the corresponding power spectrum envelope values are the same, for example, selection is made according to a predetermined rule such as selecting a sample having a smaller position on the frequency axis (a sample having a lower frequency).

(C) In the case of T + S <U _XQ The error decoding unit 230 performs the following process for samples whose decoded normalized signal sequence X ^ _Q (ω) is not 0.

The 1-bit first-cycle code included in the input error code is decoded to obtain positive / negative information, and the obtained positive / negative information is given to the absolute value 0.25 of the reconstructed value + reconstructed value + 0.25 Alternatively, −0.25 is set as the first round decoding quantization error q ₁ (ω) corresponding to the decoded normalized signal sequence X ^ _Q (ω). Furthermore, a reconstructed value obtained by decoding the 1-bit second-round code included in the input error code to obtain positive / negative information, and giving the obtained positive / negative information to the absolute value 0.125 of the reconstructed value Let +0.125 or -0.125 be the second round decoding quantization error q ₂ (ω). The first round decoding quantization error q ₁ (ω) and the second round decoding quantization error q ₂ (ω) are added to obtain a decoding quantization error q (ω).

In addition, the error decoding unit 230 performs the following processing for samples whose value of the decoded normalized signal sequence X ^ _Q (ω) is 0.

The 1-bit first-cycle code included in the input error code is decoded to obtain positive / negative information, and the obtained positive / negative information is given to the absolute value A of the reconstructed value which is a positive value smaller than 0.25. The reconstructed value + A or −A obtained in this way is defined as the first-round decoded quantization error q ₁ (ω) corresponding to the decoded normalized signal sequence X ^ _Q (ω). Further, the 1-bit second-round code included in the input error code is decoded to obtain positive / negative information, and the obtained positive / negative information is given to the absolute value A / 2 of the reconstructed value. The configuration value + A / 2 or −A / 2 is set as the second-round decoding quantization error q ₂ (ω). The first round decoding quantization error q ₁ (ω) and the second round decoding quantization error q ₂ (ω) are added to obtain a decoding quantization error q (ω).

As described above, whether the value of the corresponding quantized normalized signal sequence X ^ _Q (ω) and the value of the decoded normalized signal sequence X ^ _Q (ω) is 0 or not is the second round. The absolute value of the reconstruction value corresponding to the code is set to ½ of the absolute value of the reconstruction value corresponding to the first cycle code.

If there is no second-round code corresponding to each sample, the first-round decoding quantization error q ₁ (ω) corresponding to each sample is converted to the decoding quantization error q (ω corresponding to each sample. ).

  Further, instead of the power spectrum envelope values of the above (A) and (B), an approximate value of the power spectrum envelope value, an estimated value of the power spectrum envelope value, a value obtained by smoothing any of these values, Either a value obtained by averaging any of these values for a plurality of samples or a value having the same magnitude relationship as any of these values may be used. However, it is necessary to use the same type of value as the error encoding unit 180 of the encoding device 1.

<Adding unit 240>
The adder 240 receives a sequence of decoded quantization errors q (ω) and a decoded normalized signal sequence X ^ _Q (ω). The adding unit 240 adds the sequence of the decoded quantization error q (ω) and the decoded normalized signal sequence X ^ _Q (ω) to obtain a corrected decoded normalized signal sequence X ^ ′ _Q (ω) and (Step S6d). The corrected decoded normalized signal sequence X ^ ′ _Q (ω) is transmitted to the decoded signal sequence generation unit 250 and the sorting unit 260.

<Division section 260>
The sorting unit 260 classifies the modified decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] so that the energy in each range is as equal as possible. Or, according to the “criteria for classifying so that the number of significant samples included in each range is as equal as possible”, N ranges (where N = 2 ^D and D is a predetermined integer of 2 or more) ) (Step S7d). Consistent with the above description, the set of discrete frequency indexes of the corrected decoded normalized signal sequence X ^ ′ _Q (ω) is {L _min ,..., L _max }, and the corrected decoded normalized signal sequence is set. X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] corresponds to the entire sequence B, and the classification unit 260 uses the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω _{∈ {L min, ..., L} max}] of n in the range ^{_{{B n} n = 1 n}} = {B 1, ..., B n, ..., B n} is partitioned into. Details of this sorting process will be described later. The division index b obtained by this division processing is provided to the decoded signal sequence generation unit 250.

  The number of partitions N is set in advance in, for example, the partition unit 150 of the encoding device 1 and the partition unit 260 of the decoding device 2 so as to have a value common to the number of partitions N in the partition unit 150 of the encoding device 1. .

The classification processing performed by the classification unit 260 on the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] is corrected by the classification unit 150 of the encoding device 1. The classification of the coding apparatus 1 is performed so that the same process as the classification process performed on the post-quantization normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] is performed. It is preset between the unit 150 and the sorting unit 260 of the decoding device 2.

<Details of Sorting Process Performed by Sorting Unit 260>
The classification processing performed by the classification unit 260 on the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] is corrected by the classification unit 150 of the encoding device 1. This is the same as the division processing performed for the post-quantization normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }].

For example, the sorting process based on the “criteria for sorting so that the energy in each range is as equal as possible” is performed by, for example, n-th range (n is an integer from 1 to N−1) of the corrected decoded normalized signal sequence. ,
(a) The sum of squares of the values of all samples included in the first range to the nth range of the corrected decoded normalized signal sequence and the values of all the samples of the corrected decoded normalized signal sequence As n of the N sum of squares is closest,
Or
(b) The sum of absolute values of the values of all samples included in the first range to the nth range of the corrected decoded normalized signal sequence and the values of all the samples of the corrected decoded normalized signal sequence N of N of the sum of absolute values of
Or
(c) The total number of samples from the first range to the nth range of the corrected decoded normalized signal sequence is included in the first range to the nth range of the corrected decoded normalized signal sequence So that the sum of squares of all sample values is the minimum number of samples that is not less than n / N of the sum of squares of all sample values of the corrected decoded normalized signal sequence,
Or
(d) The total number of samples from the first range to the n-th range of the corrected decoded normalized signal sequence is included in the first range to the n-th range of the corrected decoded normalized signal sequence So that the absolute value sum of all sample values is the minimum number of samples that is not less than n / N of the absolute value sum of all sample values of the corrected decoded normalized signal sequence,
Or
(e) The total number of samples from the first range to the n-th range of the corrected decoded normalized signal sequence is included in the first range to the n-th range of the corrected decoded normalized signal sequence So that the sum of squares of all sample values is the maximum number of samples that is n or less of N / N of the sum of squares of all sample values of the corrected decoded normalized signal sequence,
Or
(f) The total number of samples from the first range to the nth range of the corrected decoded normalized signal sequence is included in the first range to the nth range of the corrected decoded normalized signal sequence So that the absolute value sum of all sample values is the maximum number of samples that is not more than n / N of the absolute value sum of all sample values of the corrected decoded normalized signal sequence,
Seeking
The corrected decoding normalization is performed by setting a range other than the first range to the (N-1) th range in the corrected decoded normalized signal sequence as the Nth range of the corrected decoded normalized signal sequence. This is done by dividing the completed signal sequence into N ranges.

  The classification process exemplified above is realized by a method of sequentially determining the classification based on the “standard for classifying so that the energy in each range is as equal as possible” sequentially from the first range. According to the classification process exemplified above, it is possible to realize the classification based on the “standard for classifying the energy in each range so as to be as equal as possible” with a small amount of calculation processing.

For example, the classification process based on the “criteria for classifying so that the number of significant samples included in each range is as equal as possible” is performed by, for example, the nth range (n is 1 to N−1) of the corrected decoded normalized signal sequence. Each integer)
(a) the number of samples whose sample energy is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected decoded normalized signal sequence; N of N samples of the number of samples whose sample energy is greater than or equal to or greater than a predetermined value among all samples included in the corrected decoded normalized signal sequence is closest.
Or
(b) the number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected decoded normalized signal sequence; , N of N samples of the number of samples whose absolute value of samples is greater than or equal to or greater than a predetermined value among all samples included in the corrected decoded normalized signal sequence is closest.
Or
(c) The number of samples whose sample energy is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected decoded normalized signal sequence, Among all samples included in the post-correction decoded normalized signal sequence, the minimum sample number is such that the energy of the sample is greater than a predetermined value or not less than n / N of the number of samples greater than or equal to a predetermined value. ,
Or
(d) The number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the n-th range of the corrected decoded normalized signal sequence , Out of all the samples included in the modified decoded normalized signal sequence, the absolute value of the sample is the minimum number of samples that is n or more of N / N of the number of samples that is greater than or equal to the predetermined value. like,
Or
(e) The number of samples whose sample energy is greater than or equal to a predetermined value among all the samples included in the first range to the nth range of the corrected decoded normalized signal sequence, Among all samples included in the post-correction decoded normalized signal sequence, the maximum sample number is such that the sample energy is greater than a predetermined value or less than n / N of the number of samples greater than or equal to a predetermined value. ,
Or
(f) The number of samples whose absolute value of samples is greater than or equal to a predetermined value among all samples included in the first range to the nth range of the corrected decoded normalized signal sequence is The absolute number of samples among all samples included in the corrected decoded normalized signal sequence is the maximum number of samples that is n or less than N / N of the number of samples that are greater than or equal to a predetermined value. like,
Seeking
The corrected decoding normalization is performed by setting a range other than the first range to the (N-1) th range in the corrected decoded normalized signal sequence as the Nth range of the corrected decoded normalized signal sequence. This is done by dividing the completed signal sequence into N ranges.

  The classification process exemplified above realizes the classification based on the “criteria for classifying so that the number of significant samples included in each range is as equal as possible” by a method of sequentially determining each range. . According to the classification process exemplified above, it is possible to realize classification according to “a criterion for classifying so that the number of significant samples included in each range is as equal as possible” with a small amount of calculation processing.

A specific example of the sorting process performed by the sorting unit 260 is a specific example of the sorting process performed by the sorting unit 150 of the encoding device 1 to “a first example of a sorting process based on the first criterion” to “a sorting based on the first criterion”. Modified Quantized Normalized Signals in Specific Examples of “Sixth Example of Processing”, “First Example of Classification Processing Based on Second Criteria” to “Sixth Example of Classification Processing Based on Second Criteria” sequence _{X ^ 'Q (ω) [} ω∈ {L min, ..., L max}] corrected decoded normalized signal sequence _{X ^' Q (ω) [} ω∈ {L min, ..., L max}] It has been replaced with.

<Decoded signal sequence generation unit 250>
The decoded signal sequence generation unit 250 includes a corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }], a gain correction amount code idx, and a decoded global gain g. ^ Is entered. Decoded signal sequence generation unit 250, the correction decoded normalized signal sequence _{X ^ 'Q (ω) [} ω∈ {L min, ..., L max}] predetermined N = 2 ^D number of (D 2 A plurality of gain corrections for each divided range using a plurality of gain correction amount codebooks stored in the storage unit 251 in the decoded signal sequence generation unit 250. Correction gain obtained by correcting the decoded global gain g ^ by a quantity for each divided range and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L _max } ] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L _max }] (step S8d).

  The gain correction amount code idx specifies the gain correction amount assigned by the bit based on the criterion that the bit is assigned with priority over the gain correction amount corresponding to a large range in the encoding device 1. In other words, the gain correction amount code idx is a gain correction in which bits are allocated based on a criterion that bits are preferentially allocated to gain correction amounts corresponding to a range having a large number of divided ranges included in the range. The amount is specified.

  For the gain correction amount corresponding to the same size range, bits may be assigned in preference to the gain correction amount corresponding to any range, but the gain correction amount corresponding to the range of high auditory importance It is preferable that bits are preferentially assigned to. In general, the low frequency band is often higher in auditory importance than the high frequency band. For this reason, with respect to the gain correction amount corresponding to the same size range, bits are assigned in preference to the gain correction amount in the low frequency range.

  Therefore, the gain correction amount code idx is a gain correction amount corresponding to a range of high auditory importance in the encoding apparatus 1 for a gain correction amount corresponding to the same number of divided ranges included in the range. The gain correction amount to which the bit is allocated may be specified based on the criterion that the bit is allocated with priority over the amount.

  Further, the gain correction amount code idx has priority over the gain correction amount corresponding to the low frequency range for the gain correction amount corresponding to the same number of the divided ranges included in the range in the encoding device 1. Further, the gain correction amount to which the bit is assigned may be specified based on the criterion of assigning the bit.

[First Example of Decoded Signal Sequence Generation Processing]
The first example of the decoded signal sequence generation process is an example in which the gain obtained by adding the quantized global gain g ^ and the gain correction amount is used as the correction gain. Here, for the gain correction amount corresponding to the same size range, bits are assigned in preference to the gain correction amount in the low frequency range, and the gain correction amounts in the two adjacent ranges are decoded by vector quantization. An example will be described. A case will be described in which the corrected decoded normalized signal sequence is divided into four ranges.

In this example, the corrected decoded normalized signal sequence is divided into a first range R1, a second range R2, a third range R3, and a fourth range R4. For example, as shown in FIG. 25, the first range R1 is a section [L _min , L ₍₁₎ -1], and the second range R2 is a section [L ₍₁₎ , L ₍₂₎ -1]. Yes, the third range R3 is the section [L ₍₂₎ , L ₍₃₎ -1], and the fourth range R4 is the section [L ₍₃₎ , L _max ]. The horizontal axis in FIG. 25 represents the sample number. These ranges R1, R2, R3, R4 is (although k is each integer generally from 1 to D-1, k = 1 in this example) ^{2 k} pieces are summarized by. A range in which the range R1 and the range R2 are combined is referred to as a range R12, and a range in which the range R3 and the range R4 are combined is referred to as a range R34.

  These ranges R1, R2, R3, R4, R12, and R34 are divided into groups each composed of a ranges for each range of the same size. In general, a is an integer of 2 or more, but in this example, a = 2. In this example, the range R1 and the range R2 constitute a group G12, the range R3 and the range R4 constitute a group G34, and the range R12 and the range R34 constitute a group G1234. That is, the range constituting each group is as follows.

Group G12 = {Range R1, Range R2}
Group G34 = {Range R3, Range R4}
Group G1234 = {range R12, range R34}
Vector quantization decoding is performed in each of these groups G12, G34, and G1234.

  Specifically, the vector quantization decoding of the gain correction amounts in the two adjacent ranges is performed by decoding the following three vector quantizations. The first vector quantization decoding is a vector quantization decoding of a gain correction amount corresponding to the range R1 and a gain correction amount corresponding to the range R2. This is hereinafter referred to as “first VQ decoding”. The second vector quantization decoding is a vector quantization decoding of the gain correction amount corresponding to the range R3 and the gain correction amount corresponding to the range R4. This is hereinafter referred to as “second VQ decoding”. The third vector quantization decoding is a vector quantization decoding of the gain correction amount corresponding to the range R12 and the gain correction amount corresponding to the range R34. This is hereinafter referred to as “third VQ decoding”.

  In the present specification, the decoded gain correction amount may be referred to as a decoded gain correction amount.

  The storage unit 251 in the decoded signal sequence generation unit 250 has the same first VQ gain correction amount codebook, second VQ gain correction amount codebook, and third VQ gain correction amount codebook as the storage unit 141 of the encoding device 1. Is stored.

That is, if a vector composed of a number (a = 2 in this example) of gain correction amount candidates is called a gain correction amount candidate vector, Δ ₁ ( 1) and Δ ₂ (1) gain correction amount candidate vector, Δ ₁ (2) and Δ ₂ (2) gain correction amount candidate vector,..., Δ ₁ (2 ^Ma ) and delta ₂ total 2 ^Ma pieces of gain correction amount candidate vector of a gain correction amount candidate vectors comprised of (2 ^Ma), a total of 2 ^Ma pieces of gain correction amount candidate vector and the corresponding total 2 ^Ma number of code delta ₁₂ ( 1), Δ ₁₂ (2), ..., idx ₁₂ (2 ^Ma ) are stored.

The second VQ gain correction amount codebook includes a gain correction amount candidate vector composed of Δ ₃ (1) and Δ ₄ (1), and a gain composed of Δ ₃ (2) and Δ ₄ (2). correction amount candidate ^{_{vectors, ···, Δ 3 (2 Mb}} ) and delta ₄ and a total of 2 ^Mb pieces of gain correction amount candidate vectors constructed gain correction amount candidate vector (2 ^Mb), total 2 ^Mb number of gain A total of 2 ^Mb codes Δ ₃₄ (1), Δ ₃₄ (2),..., Idx ₃₄ (2 ^Mb ) respectively corresponding to the correction amount candidate vectors are stored.

The third VQ gain correction amount codebook includes a gain correction amount candidate vector composed of Δ ₁₂ (1) and Δ ₃₄ (1), and a gain correction amount composed of Δ ₁₂ (2) and Δ ₃₄ (2). Candidate vectors, ..., 2 ^Mc gain correction amount candidate vectors composed of Δ ₁₂ (2 ^Mb ) and Δ ₃₄ (2 ^Mb ), and a total of 2 ^Mc gain correction amount vectors A total of 2 ^Mc codes Δ ₁₂₃₄ (1), Δ ₁₂₃₄ (2),..., Idx ₁₂₃₄ (2 ^Mc ) corresponding to the candidate vectors are stored.

Thus, a plurality of gain correction amount candidates are associated with each of the divided ranges and the ranges obtained by collecting the divided ranges by 2 ^k pieces (k is an integer from 1 to D-1). Has been. In this example, Δ ₁ (1), ..., Δ ₁ (2 ^Ma ) is associated with the range R1, and Δ ₂ (1), ..., Δ ₂ (2 ^Ma ) is associated with the range R2. It is, in the range _{R3 Δ 3 (1), ...} , Δ 3 (2 Mb) have been associated, in the range _{R4 Δ 4 (1), ...} , Δ 4 (2 Mb) is correlated Δ ₁₂ (1),..., Δ ₁₂ (2 ^Mc ) is associated with the range R12, and Δ ₃₄ (1),..., Δ ₃₄ (2 ^Mc ) is associated with the range R34. Has been.

  The gain correction amount candidates have a relationship that the absolute value of the gain correction amount candidate corresponding to the large range is larger than the absolute value of the gain correction amount candidate corresponding to the smaller range. May be. That is, the absolute value of the gain correction amount candidate corresponding to a range having a large number of divided ranges included in the range is included in the range than the number of the divided ranges included in the range. Alternatively, there may be a relationship that the absolute value of gain correction amount candidates corresponding to a range with a small number of ranges is larger.

  In this example, the range R12 and the range R34 are larger than the range R1, the range R2, the range R3, and the range R4.

_{Thus, Δ 12 (1), ...} , the absolute value of Δ ₁₂ (2 ^Mc) _{is, Δ 1 (1), ...} , the absolute value of ^{_{Δ 1 (2 Ma), Δ}} 2 (1), ..., Δ 2 ( 2 ^Ma ) absolute value, Δ ₃ (1), ..., Δ ₃ (2 ^Mb ) absolute value and Δ ₄ (1), ..., Δ ₄ (2 ^Mb ) absolute value .

_{Similarly, Δ 34 (1), ...} , the absolute value of Δ ₃₄ (2 ^Mc) _{is, Δ 1 (1), ...} , the absolute value of ^{_{Δ 1 (2 Ma), Δ}} 2 (1), ..., Δ 2 The absolute value of (2 ^Ma ), the absolute value of Δ ₃ (1), ..., Δ ₃ (2 ^Mb ) and the absolute value of Δ ₄ (1), ..., Δ ₄ (2 ^Mb ) Good.

  For example, a gain correction amount candidate vector can be generated as follows.

First, 2 ^Md number of normalized gain correction amount candidate vectors composed of a values are stored in the storage unit 141. For example, Md = Ma = Mb = Mc. When a value constituting the normalized gain correction amount candidate vector is expressed as Δ ¹ (m),..., Δ ^a (m), the normalized gain correction amount candidate vector is (Δ ¹ (m),. ^a (m)). The storage unit 141, 2 ^Md pieces of normalized gain correction amount candidate vectors, i.e. ^{(Δ 1 (1), ...} , Δ a (1)), ..., (Δ 1 (2 Md), ..., Δ a ( 2 ^Md )) is stored.

  In addition, it is assumed that a predetermined coefficient is determined for each size of the range. This coefficient is larger as the corresponding range is larger. In other words, this coefficient is larger as the number of divided ranges included in the range is larger.

In the above example, the ranges R12, R34 are larger than the ranges R1, R2, R3, R4. For this reason, the coefficient step ₁₂₃₄ corresponding to the ranges R12 and R34 is larger than the coefficient step ₁₂ corresponding to the ranges R1 and R2. Similarly, the coefficient step ₁₂₃₄ corresponding to the ranges R12 and R34 is larger than the coefficient step ₃₄ corresponding to the ranges R3 and R4.

At this time, a vector obtained by multiplying the normalized gain correction amount candidate vector by a coefficient corresponding to the size of the range is set as the gain correction amount candidate vector of the range. In other words, the normalized gain correction amount candidate vector ^{(Δ 1 (m), ...} , Δ a (m)) a number of values delta ¹ constituting the (m), ..., to each of the delta ^a (m), A vector (stepΔ ¹ (m),..., StepΔ ^a composed of ^a values stepΔ ¹ (m),..., StepΔ ^a (m) obtained by multiplying the coefficient step corresponding to the size of the range. (m)) is a gain correction amount candidate vector in that range. This multiplication is performed by the decoded signal sequence generation unit 250, for example. Since there are 2 ^Md normalized gain correction amount candidate vectors (Δ ¹ (m),..., Δ ^a (m)), by performing this multiplication for each of m = 1,..., 2 ^Md , 2 ^Md Gain correction amount candidate vectors (step Δ ¹ (m),..., Step Δ ^a (m)) are obtained.

In the above example of a = 2, when Md = Ma, the gain correction amount candidate vectors (Δ ₁ (m), Δ ₂ (m)) corresponding to the ranges R1, R2 constituting the group G12 are (Δ ₁ (m), Δ ₂ (m)) = (step ₁₂ Δ ¹ (m), step ₁₂ Δ ² (m)) [m = 1,..., 2 ^Ma ]. When Md = Mb, the gain correction amount candidate vectors (Δ ₃ (m), Δ ₄ (m)) corresponding to the ranges R3, R4 constituting the group G34 are (Δ ₃ (m), Δ ₄ (m )) = (step ₃₄ Δ ¹ (m), step ₃₄ Δ ² (m)) [m = 1,..., 2 ^Mb ]. When Md = Mb, the gain correction amount candidate vectors (Δ ₁₂ (m), Δ ₃₄ (m)) corresponding to the ranges R 12, R 34 constituting the group G 1234 are (Δ ₁₂ (m), Δ ₃₄ (m )) = (step ₁₂₃₄ Δ ¹ (m), step ₁₂₃₄ Δ ² (m)) [m = 1,..., 2 ^Mc ].

[[Example of Decoded Signal Sequence Generation Processing 1: Example Using Different Addition Formulas in Three Cases]]
Specific example 1 is an example in which the number of bits U _g of the input gain correction amount code idx is any one of Mc, Mc + Ma, and Mc + Ma + Mb.

(a) When the number of bits U _g of the input gain correction amount code idx is Mc When the number of bits U _{g of the} input gain correction amount code idx is Mc, the gain correction amount code idx includes the third VQ Only the code idx ₁₂₃₄ is included. Accordingly, the decoded signal sequence generation unit 250 first performs the third VQ decoding on the third VQ code idx ₁₂₃₄ , and obtains the decoding gain correction amount Δ ₁₂ corresponding to the range R12 and the decoding gain correction amount Δ ₃₄ corresponding to the range R34. obtain. Specifically, a set of Δ ₁₂ (1), Δ ₃₄ (1) and idx ₁₂₃₄ (1) stored in the storage unit 251, Δ ₁₂ (2), Δ ₃₄ (2) and idx ₁₂₃₄ (2) ,..., Δ ₁₂ (2 ^Mc ), Δ ₃₄ (2 ^Mc ) and idx ₁₂₃₄ (2 ^Mc ), and the same idx ₁₂₃₄ (mc) as the third VQ code idx ₁₂₃₄ The corresponding Δ ₁₂ (mc) is obtained as the decoding gain correction amount Δ ₁₂ corresponding to the range R12, and Δ ₃₄ (mc) corresponding to the same idx ₁₂₃₄ (mc) as the third VQ code idx ₁₂₃₄ corresponds to the range R34. Obtained as a decoding gain correction amount Δ ₃₄ . This is a well-known vector quantization decoding method.

Next, the range R1 includes a decoding gain correction amount delta ₁₂ corresponding to the range R12 is decoded gain correction amount for the range R1, the correction gain g ^ + delta ₁₂ obtained by adding the decoded global gain g ^ , The decoded decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L ₍₁₎ −1] and the signal sequence obtained by multiplying by the output signal Obtained as a sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1 is obtained by the equation (D7).
X ^ (ω) = (g ^ + Δ ₁₂ ) X ^ ' _Q (ω) (D7)

As for the range R2, the decoded gain correction amount delta ₁₂ corresponding to the range R12 is decoded gain correction amount for the range R2, the correction gain g ^ + delta ₁₂ obtained by adding the decoded global gain g ^, Post-correction decoded normalized signal sequence X ^ ' _Q (ω) Outputs signal sequence obtained by multiplying each sample value of [ω∈ {L ₍₁₎ , ..., L ₍₂₎ -1]] The signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1] is obtained. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1] in the range R2 is also obtained by the equation (D7).

As for the range R3, and the decoded gain correction amount delta ₃₄ corresponding to the range R34 is decoded gain correction amount for the range R3, the correction gain g ^ + delta ₃₄ obtained by adding the decoded global gain g ^, Post-correction decoded normalized signal sequence X ^ ' _Q (ω) Outputs signal sequence obtained by multiplying each sample value of [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1]] The signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] is obtained. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] in the range R3 is obtained by the equation (D8).
X ^ (ω) = (g ^ + Δ ₃₄ ) X ^ ' _Q (ω) (D8)

Also, the range R4, the decoded gain correction amount delta ₃₄ corresponding to the range R34 is decoded gain correction amount for range R4, the correction gain g ^ + delta ₃₄ obtained by adding the decoded global gain g ^, The signal sequence obtained by multiplying the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max ] ₎ with the value of each sample is output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ] in the range R4 is also obtained by the equation (D8).

(b) When the number of bits U _g of the input gain correction amount code idx is Mc + Ma When the number of bits U _{g of the} input gain correction amount code idx is Mc + Ma, the gain correction amount code idx Includes a third VQ code idx ₁₂₃₄ and a first VQ code idx ₁₂ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain a decoding gain correction amount Δ ₁ corresponding to the range R1 and a decoding gain correction amount Δ ₂ corresponding to the range R2. Each VQ decoding is performed by a well-known vector quantization decoding method.

Next, for the range R1, a decoding gain correction amount delta ₁ corresponding to the decoding gain correction amount delta ₁₂ and range R1 corresponds to the range R12 is decoded gain correction amount for the range R1, and a decoding global gain g ^ The correction gain g ^ + Δ ₁₂ + Δ ₁ obtained by the addition and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₁₎ -1] A signal sequence obtained by multiplying the value by is obtained as an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1 is obtained by the equation (D9).
X ^ (ω) = (g ^ + Δ ₁₂ + Δ ₁ ) X ^ ' _Q (ω) (D9)

As for the range R2, the decoded gain correction amount delta ₂ of decoded gain correction amount in a range R12 decoding gain correction amount delta ₁₂ and the range R2 for the range R1, is obtained by adding the decoded global gain g ^ The correction gain g ^ + Δ ₁₂ + Δ ₂ and the value of each sample of the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1]; Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] in the range R2 is obtained by the equation (D10).
X ^ (ω) = (g ^ + Δ ₁₂ + Δ ₂ ) X ^ ' _Q (ω) (D10)

As for the range R3, and the decoded gain correction amount delta ₃₄ corresponding to the range R34 is decoded gain correction amount for the range R3, the correction gain g ^ + delta ₃₄ obtained by adding the decoded global gain g ^, Post-correction decoded normalized signal sequence X ^ ' _Q (ω) Outputs signal sequence obtained by multiplying each sample value of [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1]] The signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] is obtained. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] in the range R3 is obtained by the equation (D8).

Also, the range R4, the decoded gain correction amount delta ₃₄ corresponding to the range R34 is decoded gain correction amount for range R4, the correction gain g ^ + delta ₃₄ obtained by adding the decoded global gain g ^, The signal sequence obtained by multiplying the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max ] ₎ with the value of each sample is output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ] in the fourth range is also obtained by Expression (D8).

(c) When the number of bits U _g of the input gain correction amount code idx is Mc + Ma + Mb When the number of bits U _{g of the} input gain correction amount code idx is Mc + Ma + Mb, the gain The correction amount code idx includes a third VQ code idx ₁₂₃₄ , a first VQ code idx _12, and a second VQ code idx ₃₄ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain the decoding gain correction amount Δ ₁ corresponding to the range R1 and the decoding gain correction amount Δ ₂ corresponding to the range R2, and further, the second VQ code idx It performs the 2VQ decoding to ₃₄ to obtain decoded gain correction amount delta ₄ corresponding to the decoding gain correction amount delta ₃ and scope R4 corresponding to the range R3. Each VQ decoding is performed by a well-known vector quantization decoding method.

Next, for the range R1, a decoding gain correction amount delta ₁ corresponding to the decoding gain correction amount delta ₁₂ and range R1 corresponds to the range R12 is decoded gain correction amount for the range R1, and a decoding global gain g ^ The correction gain g ^ + Δ ₁₂ + Δ ₁ obtained by the addition and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₁₎ -1] A signal sequence obtained by multiplying the value by is obtained as an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the _{first range} is obtained by the equation (D9).

As for the range R2, added to the decoded gain correction amount delta ₂ which corresponds to the decoded gain correction amount delta ₁₂ and range R2 corresponding to the range R12 is decoded gain correction amount for the range R2, and a decoding global gain g ^ And the corrected gain g ^ + Δ ₁₂ + Δ ₂ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₁₎ , ..., L ₍₂₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1}] in the range R2 is obtained by the equation (D10).

As for the range R3, added to the decoded gain correction amount delta ₃ corresponding to the decoding gain correction amount delta ₃₄ and range R3 corresponding to a range R34 is decoded gain correction amount for the range R3, and a decoding global gain g ^ Correction gain g ^ + Δ ₃₄ + Δ ₃ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1}] in the _{third range} is obtained by the equation (D11).
X ^ (ω) = (g ^ + Δ ₃₄ + Δ ₃ ) X ^ ' _Q (ω) (D11)

Also, the range R4, the decoded gain correction amount delta ₄ decoding gain correction amount delta ₃₄ and scope R4 ranging R34 is decoded gain correction amount for range R4, obtained by adding the decoded global gain g ^ Multiplying the correction gain g ^ + Δ ₃₄ + Δ _{4 by} the value of each sample of the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max ] The obtained signal sequence is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max }]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max }] in the range R4 is obtained by the equation (D12).
X ^ (ω) = (g ^ + Δ ₃₄ + Δ ₄ ) X ^ ' _Q (ω) (D12)

[[Example of decoded signal sequence generation process 2: Example using the same addition formula in three cases]]
The specific example 2 is also an example in which the number of bits U _g of the input gain correction amount code idx is any one of Mc, Mc + Ma, and Mc + Ma + Mb, as in the specific example 1.

(a) When the number of bits U _g of the input gain correction amount code idx is Mc When the number of bits U _{g of the} input gain correction amount code idx is Mc, the gain correction amount code idx includes the third VQ Only the code idx ₁₂₃₄ is included. Accordingly, the decoded signal sequence generation unit 250 first performs the third VQ decoding on the third VQ code idx ₁₂₃₄ , and obtains the decoding gain correction amount Δ ₁₂ corresponding to the range R12 and the decoding gain correction amount Δ ₃₄ corresponding to the range R34. obtain. The method for obtaining the decoding gain correction amount is the same as in the first specific example.

Also, each of the decoding gain correction amounts Δ ₁ , Δ ₂ , Δ ₃ , Δ ₄ is set to 0.

Next, for the range R1, a decoding gain correction amount delta ₁ corresponding to the decoding gain correction amount delta ₁₂ and range R1 corresponds to the range R12 is decoded gain correction amount for the range R1, and a decoding global gain g ^ The correction gain g ^ + Δ ₁₂ + Δ ₁ obtained by the addition and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₁₎ -1] A signal sequence obtained by multiplying the value by is obtained as an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1 is obtained by the equation (D9).

As for the range R2, added to the decoded gain correction amount delta ₂ which corresponds to the decoded gain correction amount delta ₁₂ and range R2 corresponding to the range R12 is decoded gain correction amount for the range R2, and a decoding global gain g ^ And the corrected gain g ^ + Δ ₁₂ + Δ ₂ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₁₎ , ..., L ₍₂₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1] in the range R2 is obtained by the equation (D10).

As for the range R3, added to the decoded gain correction amount delta ₃ corresponding to the decoding gain correction amount delta ₃₄ and range R3 corresponding to a range R34 is decoded gain correction amount for the range R3, and a decoding global gain g ^ Correction gain g ^ + Δ ₃₄ + Δ ₃ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] in the range R3 is obtained by the equation (D11).

Also, the range R4, added to the decoded gain correction amount delta ₄ corresponding to the decoding gain correction amount delta ₃₄ and scope R4 corresponding to a range R34 is decoded gain correction amount for range R4, and a decoding global gain g ^ And the corrected gain g ^ + Δ ₃₄ + Δ ₄ obtained as described above and the value of each sample of the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max ], Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ] in the range R4 is obtained by the equation (D12).

(b) When the number of bits U _g of the input gain correction amount code idx is Mc + Ma When the number of bits U _{g of the} input gain correction amount code idx is Mc + Ma, the gain correction amount code idx Includes a third VQ code idx ₁₂₃₄ and a first VQ code idx ₁₂ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain a decoding gain correction amount Δ ₁ corresponding to the range R1 and a decoding gain correction amount Δ ₂ corresponding to the range R2. The method for obtaining the decoding gain correction amount is the same as in the first specific example.

Each of the decoding gain correction amounts Δ ₃ and Δ ₄ is set to 0.

After obtaining the decoding gain correction amounts Δ ₁₂ , Δ ₃₄ , Δ _1, Δ _2, Δ ₃ , Δ ₄ , the processing for obtaining the output signal series in each range from the range R 1 to the range R 4 is performed as described in (a ).

In this way, performing the correction of the decoded global gain g ^ with the gain correction amount to which no bit is assigned as 0 uses the gain correction amount corresponding to not performing correction for a range where there is no corresponding gain correction amount. It is synonymous with that. For example, in “(a) When the number of bits U _{g of the} input gain correction amount code idx is Mc”, the ranges R1, R2, R3, and R4 are ranges in which there is no corresponding gain correction amount.

(c) When the number of bits U _g of the input gain correction amount code idx is Mc + Ma + Mb When the number of bits U _{g of the} input gain correction amount code idx is Mc + Ma + Mb, the gain The correction amount code idx includes a third VQ code idx ₁₂₃₄ , a first VQ code idx _12, and a second VQ code idx ₃₄ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain the decoding gain correction amount Δ ₁ corresponding to the range R1 and the decoding gain correction amount Δ ₂ corresponding to the range R2, and further, the second VQ code idx It performs the 2VQ decoding to ₃₄ to obtain decoded gain correction amount delta ₄ corresponding to the decoding gain correction amount delta ₃ and scope R4 corresponding to the range R3. The method for obtaining the decoding gain correction amount is the same as in the first specific example.

After obtaining the decoding gain correction amounts Δ ₁₂ , Δ ₃₄ , Δ ₁ , Δ ₂ , Δ ₃ , Δ ₄ , the process of obtaining the output signal series in each range from the range R 1 to the range R 4 is performed as described in (a ).

[[Specific Example 3 of Decoded Signal Sequence Generation Processing: An Example Including the Case where the Number of Bits of the Gain Correction Amount Code is Halfway]]
In the third specific example, the number of bits U _g of the input gain correction amount code idx includes other than Mc, Mc + Ma, and Mc + Ma + Mb, that is, the number of bits U _{g of the} input gain correction amount code idx is This is an example when the value is one of 1 or more.

(a) When the number of bits U _{g of the} input gain correction amount code idx is greater than 0 and less than or equal to Mc (0 <U ≦ Mc)
When the number of bits U _{g of the} input gain correction amount code idx is greater than 0 and equal to or less than Mc, the gain correction amount code idx includes only the third VQ code idx ₁₂₃₄ . Accordingly, the decoded signal sequence generation unit 250 first performs the third VQ decoding on the third VQ code idx ₁₂₃₄ , and obtains the decoding gain correction amount Δ ₁₂ corresponding to the range R12 and the decoding gain correction amount Δ ₃₄ corresponding to the range R34. obtain. In this case, the portion of the U _g bit that can distinguish all mc from 1 to 2 ^Ug is set as the gain correction amount code idx. For example, in the case of U _g = 1 and Mc = 2, {0} which is the first bit of the {0,0} bits of idx ₁₂₃₄ (1) or {x of idx ₁₂₃₄ (2) {1}, which is the first bit of the two bits of 1,0}, is the third VQ code idx ₁₂₃₄ . The method for obtaining the decoding gain correction amount is the same as in the first specific example.

Each of the decoding gain correction amounts Δ ₁ , Δ ₂ , Δ ₃ , Δ ₄ is set to 0.

Next, for the range R1, a decoding gain correction amount delta ₁ of decoding gain correction amount delta ₁₂ and range R1 corresponds to the range R12 is decoded gain correction amount for the range R1, and adds the decoded global gain g ^ And the corrected gain g ^ + Δ ₁₂ + Δ ₁ and the value of each sample of the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L _min , ..., L ₍₁₎ -1] , Are obtained as an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L _min ,..., L ₍₁₎ −1}] in the range R1 is obtained by the equation (D9).

As for the range R2, added to the decoded gain correction amount delta ₂ which corresponds to the decoded gain correction amount delta ₁₂ and range R2 corresponding to the range R12 is decoded gain correction amount for the range R2, and a decoding global gain g ^ And the corrected gain g ^ + Δ ₁₂ + Δ ₂ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₁₎ , ..., L ₍₂₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₁₎ ,..., L ₍₂₎ −1] in the range R2 is obtained by the equation (D10).

As for the range R3, added to the decoded gain correction amount delta ₃ corresponding to the decoding gain correction amount delta ₃₄ and range R3 corresponding to a range R34 is decoded gain correction amount for the range R3, and a decoding global gain g ^ Correction gain g ^ + Δ ₃₄ + Δ ₃ and the corrected decoded normalized signal sequence X ^ ' _Q (ω) [ω∈ {L ₍₂₎ , ..., L ₍₃₎ -1] Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₂₎ ,..., L ₍₃₎ −1] in the _{third range} is obtained by the equation (D11).

Also, the range R4, added to the decoded gain correction amount delta ₄ corresponding to the decoding gain correction amount delta ₃₄ and scope R4 corresponding to a range R34 is decoded gain correction amount for range R4, and a decoding global gain g ^ And the corrected gain g ^ + Δ ₃₄ + Δ ₄ obtained as described above, and the value of each sample of the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L ₍₃₎ ,..., L _max ], Is obtained as an output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ]. That is, each sample of the output signal sequence X ^ (ω) [ω∈ {L ₍₃₎ ,..., L _max ] in the fourth range is obtained by Expression (D11).

(b) When the number of bits U _{g of the} input gain correction amount code idx is greater than Mc and less than or equal to Mc + Ma (Mc <U _g ≦ Mc + Ma)
When the number of bits U _{g of the} input gain correction amount code idx is greater than Mc and equal to or less than Mc + Ma, the gain correction amount code idx includes the third VQ code idx ₁₂₃₄ and the first VQ code idx ₁₂ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain a decoding gain correction amount Δ ₁ corresponding to the range R1 and a decoding gain correction amount Δ ₂ corresponding to the range R2. In this case, the portion of U _g -Mc bits that can distinguish all ma in the range of 2 ^Ug-Mc +1 to 2 ^Ma is the first VQ code idx ₁₂ . The method for obtaining the decoding gain correction amount is the same as in the first specific example.

Each of the decoding gain correction amounts Δ ₃ and Δ ₄ is set to 0.

After obtaining the decoding gain correction amounts Δ ₁₂ , Δ ₃₄ , Δ ₁ , Δ ₂ , Δ ₃ , Δ ₄ , the process of obtaining the output signal series in each range from the range R 1 to the range R 4 is performed as described in (a ).

(c) When the number of bits U _{g of the} input gain correction amount code idx is greater than Mc + Ma (Mc + Ma <U _g )
When the number of bits U _{g of the} input gain correction amount code idx is larger than Mc + Ma, the gain correction amount code idx includes the third VQ code idx ₁₂₃₄ , the first VQ code idx _12, and the second VQ code idx ₃₄ . Therefore, the decoded signal sequence generation unit 250 performs third VQ decoding on the third VQ code idx ₁₂₃₄ to obtain a decoding gain correction amount Δ ₁₂ corresponding to the range R12 and a decoding gain correction amount Δ ₃₄ corresponding to the range R34. Further, the first VQ decoding is performed on the first VQ code idx ₁₂ to obtain the decoding gain correction amount Δ ₁ corresponding to the range R1 and the decoding gain correction amount Δ ₂ corresponding to the range R2, and further, the second VQ code idx It performs the 2VQ decoding to ₃₄ to obtain decoded gain correction amount delta ₄ corresponding to the decoding gain correction amount delta ₃ and scope R4 corresponding to the range R3. In this case, the portion of U _g -Mc-Ma bits that can distinguish all mbs in the range of 2 ^Ug-Mc-Ma +1 to 2 ^Mb is set as the second VQ code idx ₃₄ .

After obtaining the decoding gain correction amounts Δ ₁₂ , Δ ₃₄ , Δ ₁ , Δ ₂ , Δ ₃ , Δ ₄ , the process of obtaining the output signal series in each range from the range R 1 to the range R 4 is performed as described in (a ).

  Note that the gain correction amount candidates used in each vector quantization may be stored in one gain correction amount codebook, and the gain correction amount may be generated by one-time vector quantization decoding.

It is assumed that the number of divided ranges is ^2D . 2 ^D number of segmented 2 ^k pieces by summarizing the number of ranges range is ^{2 D / 2 k = 2 Dk} . Therefore, the number of the divided ranges and the ranges obtained by collecting the divided ranges by 2 ^k pieces (k is an integer from 1 to D-1) is 2 ^D + Σ _{d = 1} ^D-1 2 ^Dd , In total, Σ _{d = 1} ^D 2 ^d = 2 ^D + Σ _{d = 1} ^D−1 2 ^Dd . Hereinafter, A = Σ _{d = 1} ^D 2 ^d .

In this case, it is assumed that the gain correction amount candidate vector includes A gain correction amount candidates. 2 ^D number of segmented range and by collectively range of 2 ^D number of segmented range of 2 ^k pieces (each integer from k 1 to D-1), respectively constituting the gain correction amount candidate vectors Assume that these are associated with A gain correction amount candidates.

D = 2 and k = 1, and [[c) Example of decoded signal sequence generation process: Example using the same addition formula in three cases] “(c) Input gain correction amount code idx” In the example of “when the number of bits U _g is Mc + Ma + Mb”, A = Σ _{d = 1} ² 2 ^d = 2 + 4 = 6, and the gain correction amount candidate vector (Δ ₁₂ (m) of index idx (m) , Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m)) are six gain correction amount candidates Δ ₁₂ (m), Δ ₃₄ ( m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m). Candidate gain correction amounts Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m) are ranges R12, R34, R1, respectively. It corresponds to R2, R3, R4.

The gain correction amount code book stores a plurality of gain correction amount candidate vectors. In the above example, for example, 2 ^Me gain correction amount candidate vectors (Δ ₁₂ (m), Δ ₃₄ (m), Δ ₁ (m), Δ ₂ (m), Δ ₃ (m), Δ ₄ (m )) [m = 1,... 2 ^Me ] is stored in the gain correction amount code book. Me is an integer of 2 or more.

In this case, the decoded signal sequence generation unit 250 selects a gain correction amount candidate vector specified by the input gain correction amount code from a plurality of gain correction amount candidate vectors stored in the gain correction amount codebook. . The decoding global gain is corrected using the gain correction amount constituting the selected gain correction amount candidate vector.
[Modification of Decoded Signal Sequence Generation Unit 250]
Decoded signal sequence generation section 250 may obtain output signal sequence X ^ (ω) based on equations (D17) and (D18) instead of equations (D7) and (D8), respectively.
X ^ (ω) = (g ^ + s ₁₂ Δ ₁₂ ) X ^ ' _Q (ω) (D17)
X ^ (ω) = (g ^ + s ₃₄ Δ ₃₄ ) X ^ ' _Q (ω) (D18)
s ₁₂ and s _34, for example, is defined as the following equation.

The decoded signal sequence generation unit 250 replaces the equations (D9), (D10), (D11), and (D12) with equations (D19), (D20), (D21), and ( Based on D22), an output signal sequence X ^ (ω) may be obtained.
X ^ (ω) = (g ^ + s ₁₂ Δ ₁₂ + s ₁ Δ ₁ ) X ^ ' _Q (ω) (D19)
X ^ (ω) = (g ^ + s ₁₂ Δ ₁₂ + s ₂ Δ ₂ ) X ^ ' _Q (ω) (D20)
X ^ (ω) = (g ^ + s ₃₄ Δ ₃₄ + s ₃ Δ ₃ ) X ^ ' _Q (ω) (D21)
X ^ (ω) = (g ^ + s ₃₄ Δ ₃₄ + s ₄ Δ ₄ ) X ^ ' _Q (ω) (D22)
For example, s ₁ and s ₂ are defined as in the following equations.

Further, s ₃ and s ₄ are defined as in the following equations, for example.

Thus, the decoded global gain g ^ is the square of each of a plurality of gain correction amounts for each divided range and all sample values of the corrected decoded normalized signal sequence X ^ ' _Q (ω). The sum may be corrected for each divided range by a value obtained by multiplying a value obtained by dividing the sum by the square sum of the values of all the samples within the range corresponding to the respective gain correction amounts.

The decoded signal sequence generation unit 250 replaces the equations (D9), (D10), (D11), and (D12) with equations (D23), (D24), (D25), and ( The output signal sequence X ^ (ω) may be obtained based on D26).
X ^ (ω) = (g ^ + s ₁ (Δ ₁₂ + Δ ₁ )) X ^ ' _Q (ω) (D23)
X ^ (ω) = (g ^ + s ₂ (Δ ₁₂ + Δ ₂ )) X ^ ' _Q (ω) (D24)
X ^ (ω) = (g ^ + s ₃ (Δ ₃₄ + Δ ₃ )) X ^ ' _Q (ω) (D25)
X ^ (ω) = (g ^ + s ₄ (Δ ₃₄ + Δ ₄ )) X ^ ' _Q (ω) (D26)

In this way, the decoded global gain g ^ is a value obtained by adding a plurality of gain correction amounts for each divided range, and all the sample values of the corrected decoded normalized signal sequence X ^ ' _Q (ω). May be corrected for each divided range by a value obtained by multiplying a value obtained by dividing the sum of squares by the sum of squares of the values of all the samples in each divided range.

Note that s ₁₂ , s ₃₄ , s ₁ , s ₂ , s ₃ , and s ₄ may be defined as in the following equations, respectively.

c _12, the energy of the samples in the range R12 is the number of sample that is larger than a predetermined value. c _34, the energy of the samples in the range R34 is the number of sample that is larger than a predetermined value. c ₁₂₃₄ is the number of samples in which the energy of the samples in the range R1234 is larger than a predetermined value. c ₁ is the number of samples in which the energy of the samples in the range R1 is larger than a predetermined value. c ₂ is the energy of the samples in the range R2 is the number of sample that is larger than a predetermined value. c ₃ is the number of samples in which the energy of the sample in the range R3 is larger than a predetermined value. c ₄ is the energy of the samples in the range R4 is the number of sample that is larger than a predetermined value.

In this case, the decoded global gain g ^ has a plurality of gain correction amounts for each divided range and the energy of the sample of the corrected decoded normalized signal sequence X ^ ' _Q (ω) from a predetermined value. Is corrected by a value obtained by multiplying the number of samples obtained by dividing the energy of the samples within the range corresponding to the respective gain correction amounts by the number of samples larger than the predetermined value. The decoded global gain g ^ includes a value obtained by adding a plurality of gain correction amounts for each divided range, and a sampled energy of the corrected decoded normalized signal sequence X ^ ' _Q (ω) having a predetermined value. The energy of the samples in each divided range is multiplied by the value obtained by dividing the number of samples larger than the predetermined number by the number of samples larger than the predetermined value, and is corrected for each divided range. .

<Time domain conversion unit 270>
The output signal sequence X ^ (ω) is input to the time domain conversion unit 270 provided as necessary. The time domain transform unit 270 applies a frequency-time transform to the output signal sequence X ^ (ω) and outputs a time domain signal sequence z _F (n) in units of frames. The frequency-time conversion method is an inverse conversion corresponding to the time-frequency conversion method used in the frequency domain conversion unit 101. In the above example, the frequency-time conversion method here is IMDCT (Inverse Modified Cosine Transform) or IDCT (Inverse Discrete Cosine Transform).

In the processing of the control unit 170 of the encoding device 1, when it is determined that U _XQ = 0, the processing of the error encoding unit 180 and the addition unit 190 is not performed. In this case, instead of the modified quantized normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }], the quantized normalized signal sequence X ^ _Q (ω) [ Based on ω∈ {L _min ,..., L _max }], the processing of the classification unit 150 and the gain correction amount encoding unit 140 is performed. Similarly, in the processing of the control unit 205 of the decoding device 2, when U _XQ = 0 is determined, the processing of the error decoding unit 230 and the addition unit 240 is not performed. In this case, instead of the modified decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }], the decoded normalized signal series X ^ _Q (ω) [ω∈ Based on {L _min ,..., L _max }], processing of the sorting unit 260 and the decoded signal sequence generation unit 250 is performed.

On the other hand, when U _g = 0 is determined in the processing of the control unit 170 of the encoding device 1, the processing of the adding unit 190, the sorting unit 150, and the gain correction amount encoding unit 140 is not performed. In this case, the decoded signal sequence generation unit 250 uses the corrected decoded normalized signal sequence X ^ ′ _Q (ω) [ω∈ {L _min ,..., L _max }] generated by the addition unit 240 of the decoding device 2. A sequence obtained by multiplying by the decoded global gain g ^ is an output signal sequence X ^ (ω) [ω∈ {L _min ,..., L _max }].

<< Second Embodiment >>
As illustrated in FIG. 27, the control unit 170 determines that the sample value of the quantized normalized signal sequence X ^ _Q (ω) [ω∈ {L _min ,..., L _max }] has a predetermined value. The number of samples of the normalized signal sequence X _Q (ω) [ω∈ {0,..., L-1}] is not the number of samples larger than the number of samples larger than a predetermined value. Accordingly, one of the process of generating the gain correction amount code and the process of generating the error code may be performed with priority over the other process.

  In this case, the control unit 170 further transmits priority information, which is information regarding which processing is preferentially performed, not only to the error encoding unit 180 and the gain correction amount encoding unit 140 but also to the synthesis unit 160. . The synthesizing unit 160 transmits the bit stream including the priority information to the decoding device 2. The priority information transmitted to the combining unit 160 is, for example, 1-bit information.

  In this case, as illustrated in FIG. 28, the decoding device 2 is not provided with the control unit 205. The separation unit 210 separates priority information from the bitstream and transmits the priority information to the correction code separation unit 220. The correction code separation unit 200 separates the gain correction amount code and the error code from the correction code based on the priority information.

  Since other processes are the same as those in the first embodiment, description thereof is omitted here.

<< Third Embodiment >>
The third embodiment is an example in which information corresponding to the number N of divided ranges is transmitted from the encoding device 1 to the decoding device 2.

  The sorting unit 150 of the encoding device 1 determines the number N of ranges after sorting based on some standard or information transmitted from outside the sorting unit 150, and performs sorting processing so that the number of ranges after sorting becomes N. . The division unit 150 of the encoding device 1 also outputs an auxiliary code that can specify N, which is the number of ranges after the division. The sorting unit 260 of the decoding device 2 receives the auxiliary code, and performs the sorting process so that the number of ranges after the division becomes the number N specified by the auxiliary code.

  In addition to the above embodiments, the encoding device, the encoding method, the decoding device, and the decoding method according to the present invention are not limited to the above-described embodiments, and can be appropriately changed without departing from the spirit of the present invention. Is possible. In addition, the processing described in the above embodiment may be executed not only in time series according to the order of description but also in parallel or individually as required by the processing capability of the apparatus that executes the processing. .

  When the processing functions in the encoding device / decoding device are realized by a computer, the processing contents of the functions that the encoding device / decoding device should have are described by a program. By executing this program on a computer, the processing functions of the encoding device / decoding device are realized on the computer.

  The program describing the processing contents can be recorded on a computer-readable recording medium. As the computer-readable recording medium, for example, any recording medium such as a magnetic recording device, an optical disk, a magneto-optical recording medium, and a semiconductor memory may be used.

  In this embodiment, the encoding device and the decoding device are configured by executing a predetermined program on the computer. However, at least a part of the processing contents may be realized by hardware. Good.

Claims (20)

In an encoding method for encoding an input signal sequence in units of frames, which is constituted by a plurality of input signal samples,
A normal signal that obtains a normalized signal code obtained by encoding a signal sequence obtained by normalizing each input signal sample included in the input signal sequence, and a quantized normalized signal sequence corresponding to the normalized signal code. Encoded signal encoding step;
A global gain encoding step for obtaining a quantized global gain corresponding to the input signal sequence and a global gain code corresponding to the quantized global gain;
The number of samples whose sample value of the quantized normalized signal sequence is larger than a predetermined value or the sample value size of a sequence obtained by normalizing the input signal sample is larger than a predetermined value. According to the number of large samples, the gain obtained by correcting the quantized global gain with the gain correction amount is corrected for the value of each sample of the quantized normalized signal sequence or the quantized normalized signal sequence. A gain correction amount code for specifying a gain correction amount that minimizes an error between the signal sequence obtained by multiplying the value of each sample of the corrected quantized normalized signal sequence obtained by the above and the input signal sequence. A gain correction amount encoding step to be generated; and each quantized global gain or each quantized global gain of the input signal sequence is corrected by the gain One step of an error encoding step for generating an error code corresponding to an error between a signal sequence obtained by dividing by a gain obtained by correcting with a gain correction amount corresponding to a code and the quantized normalized signal sequence A control step for performing the operation in preference to the other step;
An encoding method including: The encoding method according to claim 1, comprising:
The control step causes the gain correction amount encoding step to be preferentially performed over the error encoding step as the number increases.
Encoding method. The encoding method according to claim 1, comprising:
In the control step, when the number is larger than a predetermined value, the gain correction amount encoding step is preferentially performed over the error encoding step.
Encoding method. The encoding method according to any one of claims 1 to 3,
The control step allocates a predetermined number of surplus bits to the code generated in the step performed preferentially, and the code generated in the step not performed preferentially Assign bits other than the assigned bits,
Encoding method. In a decoding method for obtaining an output signal sequence by decoding a code in frame units,
A normalized signal decoding step of decoding a normalized signal code included in the code to obtain a decoded normalized signal sequence;
A global gain decoding step of obtaining a decoded global gain by decoding a global gain code included in the code;
A decoding global gain that corrects the decoding global gain based on a gain correction amount code included in the code according to the number of samples in which the value of the sample of the decoding normalized signal sequence is larger than a predetermined value A control step for preferentially performing one step of the decoded normalized signal sequence correction step for correcting the decoded normalized signal sequence based on the error code included in the correction step and the code over the other step;
A decoding step of obtaining a signal sequence obtained by multiplying the corrected decoded global gain by each sample of the corrected decoded normalized signal sequence as an output signal sequence;
A decoding method including: The decoding method according to claim 5, wherein
The control step causes the decoding global gain correction step to be performed in preference to the decoding normalized signal sequence correction step as the number increases.
A decryption method. The decoding method according to claim 5, wherein
When the number is greater than a predetermined value, the control step causes the decoding global gain correction step to be performed in preference to the decoding normalized signal sequence correction step.
Decryption method. A decoding method according to any one of claims 5 to 7,
A predetermined number of surplus bits are assigned to the code used in the step performed preferentially, and the code used in the step not preferentially assigned is assigned the above-mentioned surplus bits. Bits other than those assigned are assigned,
Decryption method. In an encoding apparatus that encodes an input signal sequence in units of frames, which is configured by a plurality of input signal samples,
A normal signal that obtains a normalized signal code obtained by encoding a signal sequence obtained by normalizing each input signal sample included in the input signal sequence, and a quantized normalized signal sequence corresponding to the normalized signal code. An encoded signal encoding unit;
A global gain encoding unit for obtaining a quantized global gain corresponding to the input signal sequence and a global gain code corresponding to the quantized global gain;
The number of samples whose sample value of the quantized normalized signal sequence is larger than a predetermined value or the sample value size of a sequence obtained by normalizing the input signal sample is larger than a predetermined value. According to the number of large samples, the gain obtained by correcting the quantized global gain with the gain correction amount is corrected for the value of each sample of the quantized normalized signal sequence or the quantized normalized signal sequence. A gain correction amount code for specifying a gain correction amount that minimizes an error between the signal sequence obtained by multiplying the value of each sample of the corrected quantized normalized signal sequence obtained by the above and the input signal sequence. Processing to generate, and each sample value of the input signal sequence is the gain corresponding to the quantized global gain or the quantized global gain to the gain correction amount code One process of generating an error code corresponding to an error between a signal sequence obtained by dividing by a gain obtained by correcting with a positive amount and the quantized normalized signal sequence has priority over the other process. A control unit to be
An encoding device including: The encoding device according to claim 9, comprising:
The control unit causes the process of generating the gain correction amount code to be preferentially performed over the process of generating the error code as the number increases.
Encoding device. The encoding device according to claim 9, comprising:
When the number is greater than a predetermined value, the control unit causes the process of generating the gain correction amount code to be performed with priority over the process of generating the error code.
Encoding device. The encoding device according to any one of claims 9 to 11,
The control unit allocates a predetermined number of surplus bits to a code generated by a process that is performed preferentially, and the code among the surplus bits is generated by a code that is generated by a process that is not performed preferentially. Assign bits other than the assigned bits,
Encoding device. In a decoding device that obtains an output signal sequence by decoding a code in frame units,
A normalized signal decoding unit for decoding a normalized signal code included in the code to obtain a decoded normalized signal sequence;
A global gain decoding unit for decoding the global gain code included in the code to obtain a decoded global gain;
Processing for correcting the decoding global gain based on a gain correction amount code included in the code according to the number of samples whose sample value of the decoding normalized signal sequence is larger than a predetermined value; and A control unit that preferentially performs one process of the process of correcting the decoded normalized signal sequence based on an error code included in the code over the other process;
A decoding unit that obtains, as an output signal sequence, a signal sequence obtained by multiplying the corrected decoded global gain and each sample of the corrected decoded normalized signal sequence;
A decoding device. The decoding device according to claim 13,
The control unit causes the process of correcting the decoded global gain to be performed in preference to the process of correcting the decoded normalized signal sequence as the number increases.
A decoding device. The decoding device according to claim 13,
When the number is greater than a predetermined value, the control unit causes the process of correcting the decoded global gain to be performed with higher priority than the process of correcting the decoded normalized signal sequence.
Decoding device. The decoding device according to any one of claims 13 to 15,
A predetermined number of surplus bits are allocated to a code used in a process that is preferentially performed, and a code used in a process that is not performed preferentially is allocated to the code that is allocated in the surplus bits. Bits other than those assigned are assigned,
Decoding device. Program for executing the respective steps of the encoding method according to the computer in any one of claims 1 to 4. The program for making a computer perform each step of the decoding method in any one of Claim 5 to 8. Recording medium for recording a program for executing the respective steps of the encoding method according to the computer in any one of claims 1 to 4. A recording medium on which a program for causing a computer to execute each step of the decoding method according to claim 5 is recorded.

Images