JP5336943B2 - Encoding method, decoding method, encoder, decoder, program - Google Patents

Description

  The present invention relates to an encoding method and a decoding method for an acoustic signal, and particularly to an encoding method and a decoding method for vector quantization in multiple stages.

  Vector quantization is an effective means for quantizing a signal sequence with 1 bit or less per sample. This method selects an input vector and a vector with the least distortion from the codebook and transmits the number (hereinafter referred to as “index”). A codebook suitable for the quantization target is created. If this is done, quantization with small quantization distortion is possible. However, since the amount of codebook memory and the amount of computation for calculating distortion increase with an exponential function of the number of quantization bits, it is difficult to realize quantization when the number of quantization bits is large. One technique for solving this is multistage vector quantization (Patent Document 1). Further, in recent years, as an efficient vector quantization method with little quantization noise, for example, the frequency component is within a predetermined number of quantization bits such as the Spherical Vector Quantization (SVQ) method (Non-patent Document 1). A vector quantization method that stands as a pulse is widely used.

  FIG. 1 is a diagram illustrating a functional configuration example of an encoder and a decoder using a multistage vector quantization method. FIG. 2 is a diagram showing a processing flow of the encoder and the decoder. FIG. 2A shows a processing flow of the encoder, and FIG. 2B shows a processing flow of the decoder. In multi-stage vector quantization, quantizers are connected in multiple stages, the first stage first quantizer operates on the input signal, and the subsequent stages operate on errors between the input signal and the quantized output up to the previous stage. In this way, it is possible to perform vector quantization while keeping the memory amount and calculation amount within a practical range. The number of stages of multistage vector quantization in the figure is 2 for the sake of clarity, but a configuration using two or more stages may be used. As a quantization method, a method (Non-Patent Document 1) for quantizing a component converted into a frequency domain is used as an example.

  The encoder 100 includes a frequency domain transform unit 101, a normalized value calculation unit 102, a normalized value quantization unit 103, a first vector quantization unit 104, an error calculation unit 105, and a second vector quantization unit 106. The decoder 100 ′ includes a normalized value decoding unit 107, a first vector decoding unit 108, a second vector decoding unit 109, an error correction unit 110, and a time domain conversion unit 111.

  The frequency domain transform unit 101 receives an input signal x (n) as an input, and converts it into a frequency domain signal X (k) that is a signal indicating a frequency component at L points for every L input signals x (n) (for each frame). Conversion is performed (S101). Here, n represents a signal number (discrete time number) in the time domain, and k represents a signal number (discrete frequency number) in the frequency domain. The predetermined number L is, for example, 64 or 80. As a frequency domain transform method, for example, there is MDCT (Modified Discrete Cosine Transform).

The normalized value calculation unit 102 receives the frequency domain signal X (k) as an input and responds to the L point or each subband (a frequency band obtained by further dividing the L point, for example, one subband is formed by 8 points). The square root of the average power of the frequency domain signal X (k) is output as the normalized value X ^- ₀ of the frequency domain signal X (k). Assuming that the normalized value of point L is X ^- ₀ , for example, X ^- ₀ is calculated as in the following equation (S102).

In the following description, the case of using the normalization value X ^- ₀ of the L point instead of the subband normalization will be described.
Normalization value quantizer 103, the normalized value X calculated by the normalization value calculation section 102 ^- as input _0, normalized value X ^- ₀ quantizes the normalization value quantization index C _S, normal quantized normalized values corresponding to reduction value quantization index _{C S} X ^- outputting a (S103).

The first vector quantization unit 104 receives the frequency domain signal X (k) and the quantized normalized value X ⁻ and divides the frequency domain signal X (k) by the quantized normalized value X ⁻ or an inverse number. Is normalized by multiplying to obtain a normalized frequency domain signal. Then, vector quantization of the normalized frequency domain signal more samples collectively, quantized first vector quantization index C ₁ is the index of the quantization representative vectors, the signal corresponding to the first vector quantization index C ₁ normalized value X ^- seeking first quantized signal X ^ (k) by multiplying the first vector quantization index _{C 1} and the first quantized signal X ^ (k) and outputs (S104).

The error calculation unit 105 receives the frequency domain signal X (k) and the first quantized signal X ^ (k) as input, and the error between the frequency domain signal X (k) and the first quantized signal X ^ (k). For example,
E (k) = X (k) -X ^ (k) (2)
And an error signal E (k) is output (S105).

Second vector quantization section 106 receives error signal E (k) and quantized normalized value X ⁻ as input, and divides error signal E (k) by quantized normalized value X ⁻ or multiplies the inverse. To obtain a normalized error signal. Then, the normalized error signal a plurality of samples collectively by vector quantization, and outputs the index of the quantization representative vector as a second vector quantization index C ₂ (S106).

Encoder 100, a code included first vector quantization index C ₁ and the second vector quantization index C ₂ and the normalized value quantization index C _S, and sends to the decoder 100 '. The decoder 100 ′ performs the following processing.
The normalized value decoding unit 107 receives the normalized value quantization index C _S as input, obtains a decoded normalized value corresponding to the normalized value quantization index C _S, and outputs it as a decoded quantized normalized value X ⁻ ( S107).

The first vector decoding unit 108 receives the first vector quantization index C ₁ and the decoded quantization normalized value X ⁻ as input, decodes the first vector quantization index C ₁ , and normalizes the normalized signal in units of frames. determined as the first decoded quantized signal of normalized first decoded quantized signal to the decoding quantized normalization value X ^- inversely normalized by multiplying the outputs as a first decoded quantized signal X ^ (k) (S108).

Second vector decoding section 109 receives second vector quantization index C ₂ and decoded quantization normalized value X ⁻ as inputs, decodes second vector quantization index C ₂ , and normalizes the normalized signal as a normalization error. calculated as the signal, a normalized error signal to the decoding quantized normalization value X ^- inversely normalized by multiplying the outputs as a decoded quantization error signal E ^ (k) (S109) .

The error correction unit 110 receives the first decoded quantized signal X ^ (k) and the decoded quantized error signal E ^ (k) as input, and converts the decoded signal Z (k) into, for example,
Z (k) = X ^ (k) + E ^ (k) (3)
Thus, the decoded signal Z (k) is output (S110).
The time domain conversion unit 111 receives the decoded signal Z (k) as an input, performs time domain conversion using, for example, inverse MDCT for the number of frame points L, and outputs an output signal z (n) for the number of frame points L (S111). ).

Japanese Patent No. 2767474

ITU-T Recommendation G.729.1, SERIES G: TRANSMISSION SYSTEMS AND MEDIA, DIGITAL SYSTEMS AND NETWORKS, Digital terminal equipments-Coding of analogue signals by methods other than PCM, G.729-based embedded variable bit-rate coder: An 8- 32 kbit / s scalable wideband coder bitstream interoperable with G.729, 05/2006.

  In the above-described encoder 100 and decoder 100 ′, if the number of bits necessary to quantize the frequency domain signal is insufficient, the frequency component that should exist in the input signal does not exist in the output signal (frequency Component deficiency) may occur.

  However, the encoder 100 and the decoder 100 ′ have the first quantized signal regardless of whether the value of the first quantized signal (the value of the quantized signal obtained by the preceding vector quantization) is 0 or other than 0. A signal indicating an error between the quantized signal and the signal to be encoded is a target of vector quantization in the subsequent stage. In other words, the conventional technology has a special device for eliminating noise that humans feel sensitive, such as musical noise (noise that humans often feel when the presence or absence of a certain frequency component changes over time). There is no. Therefore, even if the power of error between the original signal and the decoded signal is small in numerical value, it is not an encoding that focuses on errors that are sensitive to human senses, so there is a problem that musical noise tends to remain. is there.

  The present invention has been made in view of such a problem, and even when the number of bits necessary for multistage quantization of a signal to be encoded is insufficient and a signal loss occurs after decoding. An object of the present invention is to provide a multistage vector quantization method capable of reducing musical noise by second-stage and subsequent vector quantization.

  The encoding method of the present invention includes at least a normalized value calculation step, a normalized value quantization step, a first vector quantization step, an error calculation step, and a second vector quantization step. In the normalization value calculation step, a normalization value of a signal sample to be encoded for each predetermined number of samples (hereinafter referred to as “first encoding target signal”) is obtained. As the “first encoding target signal”, an acoustic signal converted into the frequency domain, an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz, etc.) in the frequency domain acoustic signal, or divided into frame units There may be time-domain acoustic signals, but it is not necessary to limit them. The predetermined number of samples includes numbers such as 80 and 64 in the case of the number of samples of a signal constituting one frame. In the case of units called subframes, subbands, and subvectors obtained by further dividing a signal constituting one frame, there are numbers such as eight. In the normalized value quantization step, the normalized value is quantized to obtain a normalized value quantization index and a quantized normalized value. The first vector quantization step normalizes the first encoding target signal using the quantized normalized value to obtain a normalized first encoding target signal, and collects the normalized first encoding target signal by a plurality of samples. Vector quantization is performed to obtain a first vector quantization index and a first quantized signal. In the error calculation step, when the value of the first quantized signal is 0, an error signal is obtained from the absolute value of the first encoding target signal and the quantized normalized value, and the value of the first quantized signal is other than 0. In this case, an error signal is obtained from the absolute value of the first encoding target signal and the absolute value of the first quantized signal. In the second vector quantization step, a plurality of samples of the error signal or a signal corresponding to the error signal are vector-quantized to obtain a second vector quantization index.

  The decoding method of the present invention includes a normalized value decoding step, a first vector decoding step, a second vector decoding step, and an error correction step. In the decoding method, a decoded signal is obtained from a code including a first vector quantization index, a second vector quantization index, and a normalized value quantization index. The normalized value decoding step obtains a decoded normalized value corresponding to the normalized value quantization index and outputs it as a decoded quantized normalized value. The first vector decoding step obtains a normalized first decoded quantized signal corresponding to the first vector quantization index, and uses the normalized first decoded quantized signal and the decoded quantized normalized value to perform the first decoding quantization Find the signal. The second vector decoding step obtains a normalized second decoded quantized signal corresponding to the second vector quantization index, and uses the normalized second decoded quantized signal and the decoded quantized normalized value to determine the second decoded quantum Obtain the quantified signal.

  In the error correction step, when the value of the first decoded quantized signal is 0, randomly generated 1 or −1 is calculated as a calculation result using weighted addition of the decoded quantized normalized value and the second decoded quantized signal. Multiplication is performed, and the multiplication result is used as a decoded signal. Alternatively, when the value of the first decoded quantized signal is 0 and the decoded quantized normalized value of one frame past is 0, the weighted addition of the decoded quantized normalized value of the current frame and the second decoded quantized signal When the result of the above is multiplied by 1 or −1 randomly generated, the multiplication result is a decoded signal, and the value of the first decoded quantized signal is 0 and the decoded quantized normalized value in the past of one frame is other than 0, The result of weighted addition of the decoded quantization normalized value of the current frame, the second decoded quantized signal, and the decoded quantized normalized value of one frame past is multiplied by 1 or −1 generated at random, and the multiplication result is It may be a decoded signal. In addition, when the value of the first decoded quantized signal is other than 0, a calculation result using the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal is set as a decoded signal. For example, when the value of the first decoded quantized signal is negative, the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal is multiplied by −1, and the multiplication result is used as a decoded signal. When the value of the first decoded quantized signal is positive, the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal is used as the decoded signal.

  According to the encoding method and the decoding method of the present invention, in the vector quantization unit after the second stage, the first encoding target signal is not the error between the first encoding target signal and the first quantization signal. The absolute value region is replaced, and the difference between the absolute value of the first encoding target signal and the absolute value of the first quantized signal is quantized. Therefore, compared to the error between the first encoding target signal and the first quantized signal, the dynamic range of the error to be quantized in the second and subsequent stages can be reduced to half or less.

  Further, since the value of the first quantized signal for which no pulse is generated in the vector quantizing unit up to the previous stage is 0, an error between the absolute value of the first encoding target signal is large. In the present invention, for the first encoding target signal whose quantization output value up to the previous stage is 0, the difference between the absolute value of the first encoding target signal and the quantization normalized value is quantized. The error of vector quantization can be reduced.

  As described above, the encoding method and the decoding method of the present invention can greatly reduce the magnitude of the error to be quantized, and can increase the efficiency of vector quantization.

The figure which shows the function structural example of the encoder and decoder using the conventional multistage vector quantization method. The figure which shows the processing flow of the conventional encoder and decoder. The figure which shows the image of the conventional multistage vector quantization. The figure which shows the image of the multistage vector quantization of this invention. The figure which represented the 1st encoding object signal and the 1st quantization signal by the absolute value after the first stage vector quantization. The figure which showed the quantization normalization value calculated | required for every subvector with the signal shown in FIG. The figure for comparing the example of the 1st encoding object signal, and the signal used as the standard for generating the error signal of the present invention. The figure which shows the structural example of the encoder of Example 1, Example 2, and Example 3, and a decoder. The figure which shows the example of a processing flow of the encoder of Example 1, Example 2, and Example 3 and a decoder. The figure which shows the example of a specific processing flow of the error calculation part 205. FIG. The figure which shows the example of a specific process flow of the error correction part 210. FIG. The figure which shows the example of the specific processing flow of the error correction part 275. FIG. The figure for demonstrating the principle added in Examples 4-6. The figure which shows the structural example of the encoder of Example 4, Example 4 modification 1, Example 4 modification 2, and a decoder. The figure which shows the example of a processing flow of the encoder of Example 4, Example 4 modification 1, and Example 4 modification 2 of a decoder. The figure which shows the example of the processing flow for calculating | requiring the error signal E (m) and the number M. The figure which shows the example of the processing flow which calculates | requires the number M of the 1st decoding quantization signal X ^ (k) whose value is 0. The figure which shows the example of the specific processing flow of the error correction part 310,375. The figure which shows the example of the processing flow of step S3751. The figure which shows the structural example of the encoder of Example 5, Example 5 modification 1, Example 5 modification 2, and a decoder. The figure which shows the processing flow example of the encoder of Example 5, Example 5 modification 1, Example 5 modification 2, and a decoder. FIG. 20 is a diagram illustrating an example of a processing flow for obtaining a recalculated normalization value E ^{− according} to the fifth embodiment. FIG. 16 is a diagram illustrating an example of another processing flow for obtaining a recalculated normalized value E ^{− according} to the fifth embodiment. The figure which shows the structural example of the encoder of Example 6, Example 6 modification 1, and Example 6 modification 2 and a decoder. The figure which shows the example of a processing flow of the encoder of Example 6, Example 6 modification 1, and Example 6 modification 2 and a decoder. FIG. 20 is a diagram illustrating an example of a processing flow for obtaining a recalculated normalization value E ^{− according} to the sixth embodiment. It illustrates another example of processing flow for determining the ^- recalculation normalized value E of Example 6. The figure which shows the experimental result when targeting the orchestra only of a musical instrument. The figure which shows the experimental result when targeting the orchestra containing a song. The figure which shows the experimental result when targeting only the string of musical instruments. The figure which shows the experimental result when targeting the pops containing a song. The figure which shows the experimental result when targeting wind instrument (no song). The figure which shows the experimental result when only a song is made into object.

  Hereinafter, embodiments of the present invention will be described in detail. In addition, the same number is attached | subjected to the structure part which has the same function, and duplication description is abbreviate | omitted.

Principle FIG. 3 is a diagram showing an image of conventional multistage vector quantization. The horizontal axis indicates the frequency component number, and the vertical axis indicates the value of each frequency component. In this figure, the signal (first encoding target signal) to be subjected to vector quantization in the first stage is a frequency domain signal (MDCT coefficient) and is indicated by a thin line. The signal after the first stage vector quantization (first quantized signal) is indicated by a thick line. When vector quantization is performed with a small number of bits (for example, 37 bits), a large number of first quantized signals having a value of 0 (locations having a value of 0 in the frequency component) are generated. Further, in the conventional vector quantization in the latter stage, a plurality of samples of the difference between the first encoding target signal and the first quantized signal in the figure are vector-quantized.

  FIG. 4 is a diagram showing an image of multistage vector quantization according to the present invention. In this figure, the horizontal axis indicates the frequency component number, and the vertical axis indicates the value of each frequency component. The thin line indicates the absolute value of the first encoding target signal, and the thick line indicates the absolute value of the first quantized signal. In the present invention, when the value of the first quantized signal is 0, the difference between the absolute value of the first encoding target signal and the quantized normalized value is set to vector quantization in the subsequent stage. In addition, the quantization normalization value of FIG. 4 is an example calculated | required for every 8 samples.

  Next, the procedure will be described with reference to FIGS. As shown in FIG. 5, after the first-stage vector quantization, the first encoding target signal and the first quantized signal are represented by absolute values. Then, as shown in FIG. 6, a quantized normalized value is obtained for each subvector. This figure also shows that the difference between the quantized normalized value and the first encoding target signal is smaller than the error between the first quantized signal having a value of 0 and the first encoding target signal. Therefore, when the value of the first quantized signal is 0, it is more efficient to vector quantize the difference between the absolute value of the first encoding target signal and the quantized normalized value in the subsequent stage. Further, since the quantized normalized value is information that can be known by a conventional decoder, it is not necessary to transmit new information to the decoder even if the quantized normalized value is used for encoding. In FIG. 6, the quantized normalized value is obtained every 8 samples, but one quantized normalized value may be obtained from 64 samples, or the quantized normalized value may be obtained every other number. . However, if the quantization normalization value is obtained for each appropriate number of samples, the difference between the quantization normalization value and the first encoding target signal can be reduced. FIG. 7 is a diagram for comparing an example of the first encoding target signal with a signal serving as a reference for generating the error signal of the present invention. The thin line is an example of the absolute value of the first encoding target signal. A thick line is a line when the first quantized signal is an absolute value of the first quantized signal when the value of the first quantized signal is other than 0, and a quantized normalized value when the value of the first quantized signal is 0. From this figure, it can be seen that the difference can be reduced if the normalized value is used as a reference when obtaining the difference.

Specific Example FIG. 8 shows a configuration example of the encoder and decoder of the first embodiment. FIG. 9 shows an example of the processing flow of the encoder and decoder of the first embodiment. FIG. 9A shows the processing flow of the encoder, and FIG. 9B shows the processing flow of the decoder. The encoder 200 includes a frequency domain transform unit 101, a normalized value calculation unit 102, a normalized value quantization unit 103, a first vector quantization unit 104, an error calculation unit 205, and a second vector quantization unit 106. The decoder 200 ′ includes a normalized value decoding unit 107, a first vector decoding unit 108, a second vector decoding unit 109, an error correction unit 210, and a time domain conversion unit 111. The encoder 200 is different from the encoder 100 in the error calculation unit 205. The decoder 200 ′ is different from the decoder 100 ′ in the error correction unit 210. Other components are the same as those of the encoder 100 and the decoder 100 ′. Note that the first encoding target signal in this embodiment is a frequency domain signal.

Error calculator 205, if the value of the first quantized signal X ^ (k) is 0, the absolute value and the quantized normalized values X of the frequency domain signals X (k) ^- seek error signal from the first When the value of the quantized signal X ^ (k) is other than 0, the error signal E (k) is obtained from the absolute value of the frequency domain signal X (k) and the absolute value of the first quantized signal X ^ (k). (S205). FIG. 10 shows a specific processing flow example of the error calculation unit 205. The process is started with k = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S2051). When step S2051 is Yes, it is confirmed whether X ^ (k) is 0 (S2052). When Step S2052 is Yes (when the value of the first quantized signal X ^ (k) is 0), the difference between the absolute value of the frequency domain signal X (k) and the quantized normalized value X ⁻ | X (k) | -A ・ X ⁻
Is an error signal E (k) (S2053). Here, A is a positive number (for example, 1.0) for adjusting the normalized value. When Step S2052 is No (when the value of the first quantized signal X ^ (k) is other than 0), the absolute value of the frequency domain signal X (k) and the absolute value of the first quantized signal X ^ (k) | X (k) |-| X ^ (k) |
Is an error signal E (k) (S2054). k is increased by 1 (S2055), and the process returns to step S2051. If step S2051 is No, the process ends.

Error correction unit 210 of the decoder 200 ', when the value of first decoded quantized signal X ^ (k) is 0, the decoding quantized normalization value X ^- and the second decoded quantized signal E ^ (k) When the calculation result using the weighted addition is multiplied by 1 or −1 randomly generated, the multiplication result is the decoded signal Z (k), and the value of the first decoded quantized signal X ^ (k) is other than 0 Uses the sum of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k) as a decoded signal Z (k) (S210). FIG. 11 shows a specific processing flow example of the error correction unit 210. The process is started with k = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S2101). When step S2101 is Yes, it is confirmed whether X ^ (k) is 0 (S2102). Step (If the value of the first decoded quantized signal X ^ (k) of 0) S2102 If Yes, the decoded quantized normalization value X ^- and weighted addition A second decoded quantized signal E ^ (k) X ⁻ + E ^ (k) is multiplied by 1 or −1 randomly generated, and the multiplication result is set as a decoded signal Z (k) (S2103). Here, A is a positive number (for example, 1.0) for adjusting the normalized value. Further, rand (k) in the figure indicates a function that randomly generates 1 or −1, and may be calculated using a random number or the like. Thus, the reason why 1 or −1 is randomly multiplied is that if the decoded signal becomes a positive value at all frequencies, it becomes a distorted sound. By multiplying, a natural sound can be created. When step S2102 is No, it is confirmed whether X ^ (k) is negative (S2104). When Step S2104 is Yes (when the value of the first decoded quantized signal is negative), the sum of the absolute value of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k). Is multiplied by −1, and the multiplication result is set as a decoded signal Z (k) (S2105). When Step S2104 is No (when the value of the first decoded quantized signal is positive), the sum of the absolute value of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k). Is the decoded signal Z (k) (S2106). k is increased by 1 (S2107), and the process returns to step S2101. If step S2101 is No, the process ends.

  In the encoding method and the decoding method according to the first embodiment, since the absolute value is used, code information (± information) may be lost. However, in the second stage and subsequent vector quantization units, the first encoding target signal is not an absolute value but an error between the first encoding target signal (frequency domain signal in this embodiment) and the first quantization signal. The difference between the absolute value of the first encoding target signal and the absolute value of the first quantized signal is quantized by replacing with the region. Therefore, compared to the error between the first encoding target signal and the first quantized signal, the dynamic range of the error to be quantized in the second and subsequent stages can be reduced to half or less. Further, since the value of the first quantized signal for which no pulse is generated in the vector quantizing unit up to the previous stage is 0, an error between the absolute value of the first encoding target signal is large. In the present invention, for the first encoding target signal whose quantization output value up to the previous stage is 0, the difference between the absolute value of the first encoding target signal and the quantization normalized value is quantized. The error of vector quantization can be reduced.

  As described above, since the encoding method and the decoding method of the present embodiment can greatly reduce the magnitude of the error to be quantized, even if the code information is lost, the vector quantization is comprehensively performed. Efficiency can be increased.

  In the present embodiment, the signal (first encoding target signal) to be encoded by the first vector quantization unit 104 is assumed to be an acoustic signal converted into the frequency domain. However, it is not necessary to limit to this signal, and an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz) among the acoustic signals in the frequency domain may be used as the first encoding target signal or divided into frames. The time domain acoustic signal thus made may be used as the first encoding target signal.

  The second embodiment is different from the first embodiment in that the quantization normalization value of the previous frame is also considered when determining the quantization normalization value on the decoder side. The diagram showing the configuration and the processing flow is the same as that of the first embodiment, FIG. 8 shows a configuration example of the encoder and decoder of the second embodiment, and FIG. 9 shows the processing flow of the encoder and decoder of the second embodiment. An example is shown. FIG. 9A shows the processing flow of the encoder, and FIG. 9B shows the processing flow of the decoder. The encoder 200 of the second embodiment is the same as that of the first embodiment. The decoder 250 ′ according to the second embodiment includes a normalized value decoding unit 157, a first vector decoding unit 108, a second vector decoding unit 109, an error correction unit 210, and a time domain conversion unit 111. The decoder 250 ′ is different from the decoder 200 ′ in the normalized value decoding unit 157. Other components are the same as those in the first embodiment.

The normalized value decoding unit 157 obtains a decoded normalized value X ^- _C corresponding to the normalized value quantization index. The decoding normalization value X ^- _C is the decoding normalization value of the current frame being processed. Then, the decoded normalization value X before one frame (one frame previous) ^- _P is the case of 0, the decoded normalization value X of the current frame ^- the _C, decoded quantized normalization value X ^- is output as. Decoded normalization value of the previous frame and X ^- If _P is not 0, the decoded normalization value X of the current frame ^- _C and the previous frame of the decoded normalization value X ^- weighted addition of _P X ^- = .alpha.X ^- _C + ΒX ⁻ _P
And decoding quantized normalization value X ^- is output as (S157). α and β represent adjustment coefficients, for example, each may be set to 0.5.

Since the encoding method and decoding method of the second embodiment are such processes, the same effects as the encoding method and decoding method of the first embodiment can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity. In the present embodiment, a method of calculating a decoded quantization normalized value based on a decoded normalized value of all frequency components has been shown. However, for example, a quantization method for performing vector quantization for each sub-vector of 8 points is used. When used, the decoded quantized normalized value of each subvector may be calculated based on the decoded quantized normalized value obtained from the decoded normalized value of 8 points.

  In the present embodiment, the signal (first encoding target signal) to be encoded by the first vector quantization unit 104 is assumed to be an acoustic signal converted into the frequency domain. However, it is not necessary to limit to this signal, and an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz) among the acoustic signals in the frequency domain may be used as the first encoding target signal or divided into frames. The time domain acoustic signal thus made may be used as the first encoding target signal.

  In the third embodiment, the method of generating a decoded signal in the error correction unit is different from the first embodiment. The diagram showing the configuration and the processing flow is the same as that of the first embodiment, FIG. 8 is a configuration example of the encoder and decoder of the third embodiment, and FIG. 9 is the processing flow of the encoder and decoder of the third embodiment. An example is shown. FIG. 9A shows the processing flow of the encoder, and FIG. 9B shows the processing flow of the decoder. The encoder 200 of the third embodiment is the same as that of the first embodiment. The decoder 270 ′ according to the third embodiment includes a normalized value decoding unit 107, a first vector decoding unit 108, a second vector decoding unit 109, an error correction unit 275, and a time domain conversion unit 111. The decoder 270 'is different from the decoder 200' in the error correction unit 275. Other components are the same as those in the first embodiment.

If the value of the first decoded quantized signal X ^ (k) is 0 and the decoded quantized normalization value X ⁻ ′ of one frame in the past is 0, the error correcting unit 275 normalizes the decoded quantization of the current frame. the value X ^- and the result of the weighted addition of the second decoded quantized signal E ^ (k) multiplied by 1 or -1 randomly generated, the multiplication result as the decoded signal Z (k). When the value of the first decoded quantized signal X ^ (k) is 0 and the decoded quantization normalized value X ⁻ ′ of one frame past is other than 0, the decoded quantized normalized value X ⁻ of the current frame and the first 2 dequantized signal E ^ (k) and a frame previously decoded quantized normalization value X ^- multiplied by 1 or -1 randomly generated on the result of weighted addition of a ', a decoded signal a multiplication result . When the value of the first decoded quantized signal is other than 0, the calculation result using the sum of the absolute value of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k) Is a decoded signal (S275).

FIG. 12 shows an example of a specific processing flow of the error correction unit 275. The process is started with k = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S2801). When step S2801 is Yes, it is confirmed whether X ^ (k) is 0 (S2802). Step S2802 (if the value of the first decoded quantized signal X ^ (k) is 0) If Yes, 1 frame previously decoded quantized normalization value X ^- 'To check if 0 (S2803). When Step S2803 is Yes, the weighted addition A · X ⁻ + E ^ (k) of the decoded quantized normalized value X ⁻ and the second decoded quantized signal E ^ (k) is multiplied by 1 or −1 randomly generated. The multiplication result is set as a decoded signal Z (k) (S2804). When Step S2803 is No, the weighted addition of the decoded quantization normalized value X ⁻ of the current frame, the second decoded quantized signal E ^ (k), and the decoded quantized normalized value X ⁻ ′ of one frame past [ α (A · X ⁻ + E ^ (k)) + β · A · X ⁻ ′]
Is multiplied by randomly generated 1 or −1, and the multiplication result is set as a decoded signal Z (k) (S2805). Rand (k) in the figure indicates a function that randomly generates 1 or −1, and may be calculated using a random number or the like. Thus, the reason why 1 or −1 is randomly multiplied is that if the decoded signal becomes a positive value at all frequencies, it becomes a distorted sound. By multiplying, a natural sound can be created. When step S2802 is No, it is confirmed whether X ^ (k) is negative (S2806). When step S2806 is Yes (when the value of the first decoded quantized signal is negative), the sum of the absolute value of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k). Is multiplied by −1, and the multiplication result is set as a decoded signal Z (k) (S2807). When step S2806 is No (when the value of the first decoded quantized signal is positive), the sum of the absolute value of the first decoded quantized signal X ^ (k) and the second decoded quantized signal E ^ (k). Is the decoded signal Z (k) (S2808). k is incremented by 1 (S2809), and the process returns to step S2801. If step S2801 is No, the process ends.

Since the encoding method and decoding method of the third embodiment are such processes, the same effects as the encoding method and decoding method of the first embodiment can be obtained. Furthermore, decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity. In the present embodiment, a method of calculating a decoded quantization normalized value based on a decoded normalized value of all frequency components has been shown. However, for example, a quantization method for performing vector quantization for each sub-vector of 8 points is used. When used, the decoded quantized normalized value of each subvector may be calculated based on the decoded quantized normalized value obtained from the decoded normalized value of 8 points.

  In the present embodiment, the signal (first encoding target signal) to be encoded by the first vector quantization unit 104 is assumed to be an acoustic signal converted into the frequency domain. However, it is not necessary to limit to this signal, and an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz) among the acoustic signals in the frequency domain may be used as the first encoding target signal or divided into frames. The time domain acoustic signal thus made may be used as the first encoding target signal.

Principle FIG. 13 is a diagram for explaining the principle added in the fourth to sixth embodiments. The horizontal axis represents 64 samples indicating the frequency, and the vertical axis represents the value of the signal at each frequency. A thin line indicates a signal (frequency domain signal) indicating 64 frequency components obtained by converting one frame (64 time domain signals) into the frequency domain. A thick line indicates an example of the first quantized signal obtained when the first vector quantizing unit 104 vector-quantizes the frequency domain signal indicated by the thin line.

In the conventional encoder 100 and the encoder 200 according to the first to third embodiments, an error signal between the frequency domain signal and the first quantized signal is obtained. That is, when the error between the first quantized signal and the frequency domain signal having a value other than 0 is large, the bits of the second vector quantization index are used to correct this error. Therefore, the number of bits of the second vector quantization index used for correcting the error of the first quantized signal having a value of 0 is reduced, and the error between X ⁻ (k) and | X (k) | is not corrected. There is a case. Therefore, if the number of bits necessary to quantize the frequency domain signal is insufficient, the frequency component that should be present in the input signal does not exist in the decoded signal (missing frequency component), resulting in musical noise. It was one of the causes. In addition, a frequency component that should not be present in the input signal is present in the decoded signal, resulting in generation of quantization noise.

Therefore, in Embodiments 4 to 6, a signal (error signal) to be subjected to the next vector quantization is generated using only the frequency domain signal corresponding to the first quantized signal having a value of 0. In the subsequent vector quantization, the error signal is vector-quantized by collecting a plurality of samples. That is, since only the frequency domain signal corresponding to the first quantized signal having a value of 0 is vector quantized, the error between X ⁻ (k) and | X (k) | is easily corrected in the decoded signal. . Therefore, loss of frequency components and quantization noise are less likely to occur, and musical noise can be reduced.

Specific Example FIG. 14 shows a configuration example of an encoder and a decoder according to the fourth embodiment. FIG. 15 shows an example of the processing flow of the encoder and decoder of the fourth embodiment. FIG. 15A shows the processing flow of the encoder, and FIG. 15B shows the processing flow of the decoder. The encoder 300 includes a frequency domain transform unit 101, a normalized value calculation unit 102, a normalized value quantization unit 103, a first vector quantization unit 104, an error calculation unit 312, and a second vector quantization unit 306. The decoder 300 ′ includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a second vector decoding unit 309, an error correction unit 310, and a time domain conversion unit 111. The encoder 300 is different from the encoder 200 in an error calculator 312 and a second vector quantizer 306. The decoder 300 ′ is different from the decoder 200 ′ in the M-value calculating unit 313, the second vector decoding unit 309, and the error correcting unit 310. Other components are the same as those in the first embodiment. Note that the first encoding target signal in this embodiment is a frequency domain signal.

  The error calculator 312 receives the frequency domain signal X (k) and the first quantized signal X ^ (k) as inputs, and the frequency domain signal X (k) ( The error signal E (m) obtained by extracting only the frequency domain signal corresponding to the first quantized signal having a value of 0) and the number M (value of the frequency domain signal for which no pulse was generated in the first vector quantizing unit) The number of first quantized signals of 0) is obtained, and error signal E (m) and number M are output (S312). Here, m is an integer value representing the sequence number.

FIG. 16 shows an example of a processing flow for obtaining the error signal E (m) and the number M. First, the process is started with k = 0 and m = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S3121). When step S3121 is Yes, it is confirmed whether X ^ (k) is 0 (S3122). When Step S3122 is Yes, it is an error signal between the absolute value of the frequency domain signal X (k) and the quantized normalized value X ⁻ | X (k) | −A · X ⁻
Is set to E (m), and the value of m is increased by 1 (S3123). If step S3122 is No, the process proceeds to step S3124. In step S3124, the value of k is incremented by 1, and the process returns to step S3121. When S3121 is No, the value of m is set to M (S3125), and the process ends.

The second vector quantization unit 306, the error signal E (m) and quantization normalized value wherein X ^- and the value as an input the number M of first quantized signal 0. The second vector quantization unit 306 normalizes the error signal E (m) by dividing the error signal E (m) by the quantized normalized value X ⁻ or by multiplying by an inverse number to obtain a normalized error signal. Then, the normalized error signal, and vector quantization in the range of order of multiple Th of M points or quantized vector, and outputs the index of the quantization representative vector as a second vector quantization index C ₂ (S306). The order of the quantization vector is, for example, 8. In this case, the multiple Th has 8, 16,..., 64, etc., and the number closest to M, the number closest to M, or the number closest to M may be selected as Th. For example, when the closest number less than or equal to M is selected as Th, vector quantization may be performed for Th out of M normalized error signals.

Next, the decoder 300 ′ will be described. The M-value calculation unit 313 receives the first decoded quantized signal X ^ (k) obtained by the first vector decoding unit 108 as an input, and the first decoded quantized signal X ^ (k) has a value of zero. The number M of one decoded quantized signal X ^ (k) is obtained. FIG. 17 is an example of a processing flow for obtaining the number M of first decoded quantized signals X ^ (k) whose value is 0. The process is started with k = 0 and m = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S3131). When step S3131 is Yes, it is confirmed whether X ^ (k) is 0 (S3132). When step S3132 is Yes, the value of m is increased by 1 (S3133). If step S3132 is No, the process proceeds to step S3134. In step S3134, the value of k is incremented by one, and the process returns to step S3131. When S3131 is No, the value of m is set to M (S3135), and the process ends. In the normalization of the encoder side dequantized normalized value X ^- and that is used, by the decoder side is provided with M value calculation unit 313, a decoder information indicating the number M from the encoder (No need to use bits to convey the number M).

The second vector decoding unit 309, the second vector quantization index C ₂ and the decoded quantized normalization value X ^- an input of the number M of the value is 0 the first decoded quantized signal X ^ (k) . The second vector decoding unit 209, a second vector by decoding the quantization index C ₂ in order of multiple Th of M points or quantized vector to obtain the normalized signal as the normalized second decoded quantized signal, normal The normalized second decoded quantized signal is denormalized by multiplying by the decoded quantized normalized value X ⁻ and output as the second decoded quantized signal E ^ (m) (S309).

  The error correction unit 310 uses the first decoded quantized signal X ^ (k), the second decoded quantized signal E ^ (m), and a multiple of the order of the quantization vector used when the second vector decoding unit 309 decodes. The number M of the first decoded quantized signals X ^ (k) whose values are Th and 0 is input. When the value of the first decoded quantized signal X ^ (k) is other than 0, the error correcting unit 310 sets the first decoded quantized signal X ^ (k) as the decoded signal Z (k), and the first decoded quantized signal When the value of the signal X ^ (k) is 0, the second decoded quantized signal E ^ (m) is sequentially set as the decoded signal Z (k), thereby obtaining and outputting the decoded signal Z (k) (S314). ). In other words, in the case of the present embodiment, since there is no second decoded quantized signal when the value of the first decoded quantized signal is other than 0, the first decoded quantized signal is simply used as the decoded signal. This is the same when 0 is substituted for E ^ (m) in steps S2105 and S2106 in FIG. 11 and steps S2807 and S2808 in FIG.

FIG. 18 is a diagram illustrating an example of a specific processing flow of the error correction unit 310. The process is started with k = 0 and m = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S3101). If step S3101 is Yes, it is confirmed that m is smaller than M (S3102). If step S3102 is Yes, it is confirmed that m is smaller than Th (S3103). When step S3103 is Yes, it is confirmed whether X ^ (k) is 0 (S3104). When Step S3104 is Yes, the weighted addition A · X ⁻ + E ^ (k) of the decoded quantized normalized value X ⁻ and the second decoded quantized signal E ^ (k) is multiplied by 1 or −1 randomly generated. Then, the multiplication result is the decoded signal Z (k), and the value of m is incremented by 1 (S3105). Here, A is a positive number (for example, 1.0) for adjusting the normalized value. Further, rand (k) in the figure indicates a function that randomly generates 1 or −1, and may be calculated using a random number or the like. Thus, the reason why 1 or −1 is randomly multiplied is that if the decoded signal becomes a positive value at all frequencies, it becomes a distorted sound. By multiplying, a natural sound can be created. When Steps S3102, S3103, and S3104 are No, X ^ (k) is set to Z (k) (S3106). Then, the value of k is incremented by 1 (S3107), and the process returns to step S3101. If S3101 is No, the process ends.

  Because of such processing, the same effects as the encoding method and decoding method of the first embodiment can be obtained. Further, in the encoding method and the decoding method of the fourth embodiment, the error signal does not include the first encoding target signal corresponding to the first quantized signal having a value other than 0, and the 0th value is 0. Only the first encoding target signal corresponding to one quantized signal is included. That is, the second vector quantization step is a process of adding a value other than 0 to the first encoding target signal whose value has become 0 in the first vector quantization. Therefore, the second code is a code for reducing the error between the first encoding target signal and the first quantized signal while avoiding the presence or absence of the frequency component for each frame that is one of the causes of musical noise. A vector quantization index can be determined. Therefore, efficient vector quantization with a limited number of bits can be realized.

  Further, according to the encoding method and the decoding method of the first embodiment, for example, the number of frequency components in one frame is set to 80 points (80 frequency domain signals), and pulses are generated in the first vector quantization unit. If the number of frequency components obtained is 40 points (40 first quantized signals having a value other than 0), the number of points to be vector quantized by the first vector quantization unit 104 is 80 points. Since the number of points to be vector quantized by the vector quantization unit 206 can be halved to 40 points, it can be expected to improve the efficiency of vector quantization.

  In the present embodiment, the signal (first encoding target signal) to be encoded by the first vector quantization unit 104 is assumed to be an acoustic signal converted into the frequency domain. However, it is not necessary to limit to this signal, and an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz) among the acoustic signals in the frequency domain may be used as the first encoding target signal or divided into frames. The time domain acoustic signal thus made may be used as the first encoding target signal.

[Modification 1]
Embodiment 4 Modification 1 is different from Embodiment 4 in that the decoding quantization normalization value of the previous frame is also taken into consideration when determining the decoding quantization normalization value on the decoder side. The diagram showing the configuration and the processing flow is the same as that of the fourth embodiment. FIG. 14 shows a configuration example of the encoder and decoder of the first modification of the fourth embodiment, and FIG. 15 shows the encoder of the fourth modification of the first embodiment. An example of the processing flow of the decoder is shown. FIG. 15A shows the processing flow of the encoder, and FIG. 15B shows the processing flow of the decoder. Fourth Embodiment An encoder 300 according to the first modification is the same as that according to the fourth embodiment. Fourth Embodiment A decoder 350 ′ of the first modification includes a normalized value decoding unit 157, a first vector decoding unit 108, an M value calculation unit 313, a second vector decoding unit 309, an error correction unit 310, and a time domain conversion unit 111. Is provided. The decoder 350 ′ is different from the decoder 300 ′ in the normalized value decoding unit 157. Other components are the same as those in the fourth embodiment. Further, the normalized value decoding unit 157 performs the same processing as that described in the second embodiment.

Since the encoding method and decoding method of the fourth modification example are such processes, the same effects as the encoding method and decoding method of the fourth example can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

[Modification 2]
Fourth Embodiment A second modification is different from the fourth embodiment in the method of generating a decoded signal in the error correction unit. The diagram showing the configuration and the processing flow is the same as that of the fourth embodiment. FIG. 14 shows a configuration example of the encoder and decoder of the second modification of the fourth embodiment, and FIG. 15 shows the encoder of the fourth modification of the fourth embodiment. An example of the processing flow of the decoder is shown. FIG. 15A shows the processing flow of the encoder, and FIG. 15B shows the processing flow of the decoder. Fourth Embodiment An encoder 300 according to the second modification is the same as that according to the fourth embodiment. Fourth Embodiment A decoder 370 ′ of the second modification includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a second vector decoding unit 309, an error correction unit 375, and a time domain conversion unit 111. Is provided. Decoder 370 ′ differs from decoder 300 ′ in error correction section 375. Other components are the same as those in the fourth embodiment.

Error correction unit 375, decoding quantized normalization value value of 0 and 1 frame past the first decoded quantized signal X ^ (k) X ^- If 'is 0, decoding quantized normalization of the current frame the value X ^- and the result of the weighted addition of the second decoded quantized signal E ^ (k) multiplied by 1 or -1 randomly generated, the multiplication result as the decoded signal Z (k). When the value of the first decoded quantized signal X ^ (k) is 0 and the decoded quantization normalized value X ⁻ ′ of one frame past is other than 0, the decoded quantized normalized value X ⁻ of the current frame and the first 2 dequantized signal E ^ (k) and a frame previously decoded quantized normalization value X ^- multiplied by 1 or -1 randomly generated on the result of weighted addition of a ', a decoded signal a multiplication result . When the value of the first decoded quantized signal is other than 0, the first decoded quantized signal X ^ (k) is set as a decoded signal (S375).

FIG. 18 shows an example of a specific processing flow of the error correction unit 375. FIG. 19 shows an example of the processing flow of step S3751. The process is started with k = 0 and m = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S3101). If step S3101 is Yes, it is confirmed that m is smaller than M (S3102). If step S3102 is Yes, it is confirmed that m is smaller than Th (S3103). When step S3103 is Yes, it is confirmed whether X ^ (k) is 0 (S3104). When step S3104 is Yes, the decoded signal Z (k) is calculated, and the value of m is increased by 1 (S3751). In step S3751, the following processing is performed (see FIG. 19). It is confirmed whether the decoded quantization normalized value X ⁻ ′ of one frame in the past is 0 (S3752). When Step S3752 is Yes, the weighted addition A · X ⁻ + E ^ (k) of the decoded quantized normalized value X ⁻ and the second decoded quantized signal E ^ (k) is multiplied by 1 or −1 randomly generated. The multiplication result is set as a decoded signal Z (k) (S3753). When Step S3752 is No, the weighted addition of the decoded quantization normalized value X ⁻ of the current frame, the second decoded quantized signal E ^ (k), and the decoded quantized normalized value X ⁻ ′ of the previous frame [ α (A · X ⁻ + E ^ (k)) + β · A · X ⁻ ′]
Is multiplied by randomly generated 1 or −1, and the multiplication result is set as a decoded signal Z (k) (S3754). Here, A is a positive number (for example, 1.0) for adjusting the normalized value. Further, rand (k) in the figure indicates a function that randomly generates 1 or −1, and may be calculated using a random number or the like. Thus, the reason why 1 or −1 is randomly multiplied is that if the decoded signal becomes a positive value at all frequencies, it becomes a distorted sound. By multiplying, a natural sound can be created. When Steps S3102, S3103, and S3104 are No, X ^ (k) is set to Z (k) (S3106). Then, the value of k is incremented by 1 (S3107), and the process returns to step S3101. If S3101 is No, the process ends.

Since the encoding method and decoding method of the fourth modification example are such processes, the same effects as the encoding method and decoding method of the fourth example can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

In the fourth embodiment, the second vector quantization unit 306 performs normalization using the quantization normalized value X ⁻ . Further, in the second vector decoding unit 309, decoded quantized normalization value X ^- inversely normalized using. However, the second vector quantization unit vector quantizes the error signal E (m) composed only of the frequency domain signal corresponding to the first quantized signal X ^ (k) whose value is 0. Therefore, using the normalized value for the error signal E (m) improves the accuracy of normalization and enables efficient vector quantization. Therefore, in this embodiment, normalization is performed using a normalized value obtained from the error signal E (m) (hereinafter referred to as “recalculated normalized value E ⁻ ”).

  FIG. 20 shows a configuration example of the encoder and decoder of the fifth embodiment. FIG. 21 shows an example of the processing flow of the encoder and decoder of the fifth embodiment. FIG. 21A shows the processing flow of the encoder, and FIG. 21B shows the processing flow of the decoder. The encoder 400 includes a frequency domain transform unit 101, a normalized value calculation unit 102, a normalized value quantization unit 103, a first vector quantization unit 104, an error calculation unit 312, a normalized value recalculation unit 415, a second A vector quantization unit 406 is provided. The decoder 400 ′ includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 416, a second vector decoding unit 409, an error correction unit 310, and a time domain conversion unit. 111. The encoder 400 is different from the encoder 300 in the normalized value recalculation unit 415 and the second vector quantization unit 406. The decoder 400 ′ is different from the decoder 300 ′ in a normalized value recalculation unit 416 and a second vector decoding unit 409. Other components are the same as those in the fourth embodiment.

Normalization value recalculation unit 415, a first quantized signal X ^ (k), the quantization normalized value X ^- and, the number M of first quantized signal value is 0, the second vector quantization unit A multiple Th of the order of the quantization vector close to the number M used in 406 is input. The normalized value recalculation unit 415 recalculates the normalized value for the error signal E (m) and outputs it as a recalculated normalized value E ⁻ (S415). Figure 22 is recalculated normalized value E ^- is an example of a processing flow for determining the. The process is started with k = 0 and tmp = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S4151). If step S4151 is Yes, | X ^ (k) | ² is added to tmp (S4156). The value of k is incremented by 1 (S4157), and the process returns to step S4151. If S4151 is No, the recalculation normalized value E ^- the

(S4158), and the process ends.
Further, FIG. 23 is recalculated normalized value E ^- is an example of another processing flow for determining the. In this flow, the number of first quantized signals having a value of 0 is calculated again. The process starts with k = 0, m = 0, and tmp = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S4151). If step S4151 is Yes, it is confirmed that m is smaller than M (S4152). If step S4152 is Yes, it is confirmed that m is smaller than Th (S4153). When step S4153 is Yes, it is confirmed whether X ^ (k) is 0 (S4154). When step S4154 is Yes, the value of m is increased by 1 (S4155). When Steps S4152, S4153, and S4154 are No, | X ^ (k) | ² is added to tmp (S4156). Then, the value of k is increased by 1 (S4157), and the process returns to step S4151. If S4151 is No, the recalculation normalized value E ^- the

(S4158 ′) and the process ends. In both the processing flow of FIG. 22 and the processing flow of FIG. 23, the recalculated normalized value E ⁻ multiplies the square of the quantized normalized value X ⁻ by the number L of the frequency domain signals X (k), The power of the first quantized signal X ^ (k) | X ^ (k) | ² is subtracted, the subtraction result is divided by the number M of the first quantized signals having a value of 0, and the square root of the division result is obtained. The recalculated normalization value E ⁻ is used.

The second vector quantization unit 406 is different from the second vector quantization unit 306 of the fourth embodiment in that the recalculated normalization value E ⁻ is used for normalization, but the others are the same. That is, the second vector quantization unit 406, the error signal E (m) and recalculates normalized value E ^- the value is to enter the number M of first quantized signal 0. The second vector quantization unit 406 normalizes the error signal E (m) by dividing the error signal E (m) by the recalculated normalization value E ⁻ or multiplying by the reciprocal to obtain a normalized error signal. Then, the normalized error signal, and vector quantization in the order of a multiple Th of M points or quantized vector, and outputs the index of the quantization representative vector as a second vector quantization index C ₂ (S406).

Next, differences between the decoder 400 ′ will be described. The normalized value recalculator 416 obtains a recalculated normalized value E ⁻ for the decoder 400 ′ by the same process as the normalized value recalculator 415. If the recalculated normalization value E ⁻ is obtained according to the processing flow of FIGS. 22 and 23, it is not necessary to transmit the information of the recalculated normalization value E ⁻ from the encoder to the decoder, and avoid using bits. be able to.

The second vector decoding unit 409 is the same as the second vector decoding unit 309 of the fourth embodiment except that the recalculated normalization value E ⁻ is used for normalization. That is, the second vector decoding unit 409 receives the second vector quantization index C ₂ , the recalculated normalized value E ^−, and the number M of the first decoded quantized signals X ^ (k) whose values are 0. . The second vector decoding unit 409, a second vector by decoding the quantization index C ₂ in order of multiple Th of M points or quantized vector to obtain the normalized second decoded quantized signal is normalized signal, The normalized second decoded quantized signal is denormalized by multiplying it by the recalculated normalization value E ⁻ and is output as the second decoded quantized signal E ^ (m) (S409).

  Since it is such a process, according to the encoding method and decoding method of Example 5, the same effect as the encoding method and decoding method of Example 4 is acquired. Since the normalized value for the error signal E (m) is used, the accuracy of normalization is improved and vector quantization can be performed more efficiently.

  In the present embodiment, the signal (first encoding target signal) to be encoded by the first vector quantization unit 104 is assumed to be an acoustic signal converted into the frequency domain. However, it is not necessary to limit to this signal, and an acoustic signal in a specific frequency band (for example, 8 kHz to 14 kHz) among the acoustic signals in the frequency domain may be used as the first encoding target signal or divided into frames. The time domain acoustic signal thus made may be used as the first encoding target signal.

[Modification 1]
Fifth Embodiment Modification 1 is different from the fifth embodiment in that the decoding quantization normalization value of the previous frame is also taken into consideration when determining the decoding quantization normalization value on the decoder side. The diagram showing the configuration and the processing flow is the same as that of the fifth embodiment, FIG. 20 shows an example of the configuration of the encoder and decoder of the first modification of the fifth embodiment, and FIG. 21 shows the encoder of the fifth modification of the first embodiment. An example of the processing flow of the decoder is shown. FIG. 21A shows the processing flow of the encoder, and FIG. 21B shows the processing flow of the decoder. Fifth Embodiment An encoder 400 according to the first modification is the same as that of the fifth embodiment. Fifth Embodiment A decoder 450 ′ of the first modification includes a normalized value decoding unit 157, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 416, a second vector decoding unit 409, an error correction. Unit 310 and time domain conversion unit 111. The decoder 450 ′ differs from the decoder 400 ′ in the normalized value decoding unit 157. Other components are the same as those in the fifth embodiment. Further, the normalized value decoding unit 157 performs the same processing as that described in the second embodiment.

Since the encoding method and decoding method of the fifth modification example are such processes, the same effects as the encoding method and decoding method of the fifth embodiment can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

[Modification 2]
Fifth Embodiment A second modification is different from the fifth embodiment in the method of generating a decoded signal in the error correction unit. The diagram showing the configuration and the processing flow is the same as that of the fifth embodiment, FIG. 20 shows an example of the configuration of the encoder and decoder of the second modification of the fifth embodiment, and FIG. 21 shows the encoder of the fifth modification of the fifth embodiment. An example of the processing flow of the decoder is shown. FIG. 21A shows the processing flow of the encoder, and FIG. 21B shows the processing flow of the decoder. Fifth Embodiment An encoder 400 according to the second modification is the same as the fifth embodiment. Fifth Embodiment A decoder 470 ′ of the second modification includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 416, a second vector decoding unit 409, an error correction. Unit 375 and time domain conversion unit 111. Decoder 470 ′ differs from decoder 400 ′ in error correction section 375. Other components are the same as those in the fifth embodiment. The error correcting unit 375 performs the same processing as that described in the fourth modification of the fourth embodiment (see FIGS. 18 and 19).

Since the encoding method and decoding method of the fifth modification example are such processes, the same effects as the encoding method and decoding method of the fifth example can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

Quantized normalized values X ^- and the first quantized signal X ^ (k) and the target value ^- if the error between (normalized value of the frequency domain signals X ₀ and the frequency domain signal X (k)) is large, Example the calculation method, recalculated normalized value E ^{- -} 5 recalculated normalized value E of the Sometimes not be accurately calculated. In this case, for example, if the recalculated normalization value E ⁻ is estimated too small, the frequency component is lost, and if the recalculated normalization value E ⁻ is estimated too large, the quantization noise is generated. Therefore, in this embodiment, the normalized value recalculation units 515 and 516 calculate the recalculated normalized value E ⁻ by the following method.

  FIG. 24 shows a configuration example of the encoder and decoder of the sixth embodiment. FIG. 25 shows an example of the processing flow of the encoder and decoder of the sixth embodiment. FIG. 26A shows the processing flow of the encoder, and FIG. 26B shows the processing flow of the decoder. The encoder 500 includes a frequency domain transform unit 101, a normalized value calculator 102, a normalized value quantizer 103, a first vector quantizer 104, an error calculator 312, a normalized value recalculator 515, a second A vector quantization unit 406 is provided. The decoder 500 ′ includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 516, a second vector decoding unit 409, an error correction unit 310, and a time domain conversion unit. 111. The encoder 500 is different from the encoder 400 in the normalized value recalculation unit 515. The decoder 500 'is different from the decoder 400' in the normalized value recalculation unit 516. Other components are the same as those in the fifth embodiment.

The normalized value recalculation units 515 and 516 obtain a recalculated normalized value E ⁻ as follows. Figure 26 is recalculated normalized value E ^- is an example of a processing flow for determining the. The process is started with k = 0, tmp = 0, and tmp2 = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S5151). If step S5151 is Yes, | X ^ (k) | ² is added to tmp, and | X ^ (k) | is added to tmp2 (S5156). The value of k is incremented by 1 (S5157), and the process returns to step S5151. When S5151 is No, the first recalculated value E ^- ₁ and the second recalculated value E ^- ₂ are

(S5158). E ^- ₁ is E ^- large or confirms than ₁ times _{2 A} (S5159). A ₁ is an adjustment constant serving as a reference for selecting which candidate, and for example, A ₁ = 0.5. If step S5159 is Yes, the first recalculated E ^- ₁ recalculates normalized value E ^- and then (S5160), if the step S5159 is No, the second recalculated E ^- ₂ recalculation normalized value E ^- (S5161). It is confirmed whether E ⁻ is smaller than A ₂ times X ⁻ (S5162). A ₂ is a constant for confirming whether E ⁻ has a contradictory value, and for example, A ₂ = 1. Step S5162 is finished processing cases Yes, the case step S5162 is No, the recalculation normalized value E ^- quantized normalized value X ^- replaced (S5163), the process ends.

FIG. 27 is an example of another processing flow for obtaining the recalculated normalization value E ⁻ . In this flow, the number of first quantized signals having a value of 0 is calculated again. The process is started with k = 0, m = 0, tmp = 0, and tmp2 = 0. It is confirmed whether k is smaller than L (number of frequency domain signals X (k)) (S5151). If step S5151 is Yes, it is confirmed that m is smaller than M (S5152). If step S5152 is Yes, it is confirmed that m is smaller than Th (S5153). When step S5153 is Yes, it is confirmed whether X ^ (k) is 0 (S5154). When step S5154 is Yes, the value of m is increased by 1 (S5155). When Steps S5152, S5153, and S5154 are No, | X ^ (k) | ² is added to tmp, and | X ^ (k) | is added to tmp2 (S5156). The value of k is incremented by 1 (S5157), and the process returns to step S5151. When S5151 is No, the first recalculated value E ^- ₁ and the second recalculated value E ^- ₂ are

(S5158 ′). The processing from step S5159 to S5163 is the same as that in FIG. In the processing flow of FIG. 26, in the processing flow of FIG. 27, the first recalculated E ^- ₁ is quantized normalized values X ^- multiplied by the number L of the frequency domain signal X (k) to the square of the multiplication result Is subtracted from the sum of the powers | X ^ (k) | ² of the first quantized signal X ^ (k), and the subtraction result is divided by the number M of first quantized signals having a value of 0, and the square root of the division result Is the first recalculated value E ^- ₁ . The second recalculated E ^- ₂ is quantized normalized values X ^- multiplied by the number L of the frequency domain signal X (k), the absolute value of the first quantized signal X ^ (k) from the multiplication result The total of | X ^ (k) | is subtracted, the subtraction result is divided by the number M of first quantized signals having a value of 0, and the division result is taken as a second recalculated value E ^- ₂ .

Since it is such a process, according to the encoding method and decoding method of Example 6, the same effect as the encoding method and decoding method of Example 5 is acquired. The recalculation normalized value E ^- since the can more accurately calculated, and improves the accuracy of the normalization, can more efficiently vector quantization.

[Modification 1]
Embodiment 6 Modification 1 is different from Embodiment 6 in that the decoding quantization normalization value of the previous frame is also taken into consideration when determining the decoding quantization normalization value on the decoder side. The diagram showing the configuration and the processing flow is the same as that of the sixth embodiment, FIG. 24 shows an example of the configuration of the encoder and decoder of the first modification of the sixth embodiment, and FIG. 25 shows the encoder of the sixth modification of the first embodiment. An example of the processing flow of the decoder is shown. FIG. 25A shows the processing flow of the encoder, and FIG. 25B shows the processing flow of the decoder. Sixth Embodiment An encoder 400 according to the first modification is the same as that according to the sixth embodiment. Sixth Embodiment A decoder 550 ′ of the first modification includes a normalized value decoding unit 157, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 516, a second vector decoding unit 409, an error correction. Unit 310 and time domain conversion unit 111. The decoder 550 ′ is different from the decoder 500 ′ in the normalized value decoding unit 157. Other components are the same as those in the sixth embodiment. Further, the normalized value decoding unit 157 performs the same processing as that described in the second embodiment.

Embodiment 6 Since the encoding method and decoding method of Modification 1 are such processes, the same effects as the encoding method and decoding method of Embodiment 6 can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

[Modification 2]
Embodiment 6 Modification 2 is different from Embodiment 6 in the method of generating a decoded signal in the error correction unit. The diagram showing the configuration and the processing flow is the same as that of the sixth embodiment, FIG. 24 shows an example of the configuration of the encoder and decoder of the second modification of the sixth embodiment, and FIG. 25 shows the encoder of the sixth modification of the sixth embodiment. An example of the processing flow of the decoder is shown. FIG. 26A shows a processing flow of the encoder, and FIG. 25B shows a processing flow of the decoder. Sixth Embodiment An encoder 500 according to the second modification is the same as the sixth embodiment. Sixth Embodiment A decoder 570 ′ of the second modification includes a normalized value decoding unit 107, a first vector decoding unit 108, an M value calculation unit 313, a normalized value recalculation unit 516, a second vector decoding unit 409, an error correction. Unit 375 and time domain conversion unit 111. Decoder 570 ′ differs from decoder 500 ′ in error correction section 375. Other components are the same as those in the sixth embodiment. The error correcting unit 375 performs the same processing as that described in the fourth modification of the fourth embodiment (see FIGS. 18 and 19).

Since the encoding method and decoding method of the sixth modification example are such processing, the same effects as the encoding method and decoding method of the sixth example can be obtained. Furthermore, he decoded quantized normalization value X decoder side ^- it is possible to reduce the musical noise generated in quantizing the output by the time-axis direction of the discontinuity.

Experimental Results An experiment was conducted to confirm the effect of the method of the sixth modification. FIG. 28 is a diagram showing experimental results when targeting an orchestra containing only a musical instrument, FIG. 29 is a diagram showing experimental results when targeting an orchestra containing a song, and FIG. 30 is intended for a string containing only an instrument. The figure which shows the experimental result when doing, FIG. 31 is the figure which shows the experimental result when targeting pops containing a song, FIG. 32 is the figure which shows the experimental result when targeting wind instrument (no song), FIG. 33 is a diagram showing experimental results when only a song is targeted. In each figure, the horizontal axis indicates time, and the vertical axis indicates the S / N ratio shown in the following equation. The larger the value, the better the quality.

A dotted line indicates a case where encoding and decoding are performed by a conventional method, and a solid line indicates a case where encoding and decoding are performed by a method according to a second modification of the sixth embodiment. The first encoding target signal is a frequency domain signal converted into the frequency domain by MDCT (Modified Discrete Cosine Transform). Any experimental result shows that the quality of the present invention is better. This also shows that the present invention can realize efficient vector quantization with a limited number of bits.

The various processes described above, such as a program, are not only executed in time series according to the description, but may also be executed in parallel or individually as required by the processing capability of the apparatus that executes the process. Needless to say, other modifications are possible without departing from the spirit of the present invention.
Further, when the above-described configuration is realized by a computer, processing contents of functions that each device should have are described by a program. The processing functions are realized on the computer by executing the program 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.
The program is distributed by selling, transferring, or lending a portable recording medium such as a DVD or CD-ROM in which the program is recorded. Furthermore, the program may be distributed by storing the program in a storage device of the server computer and transferring the program from the server computer to another computer via a network.

  A computer that executes such a program first stores, for example, a program recorded on a portable recording medium or a program transferred from a server computer in its own storage device. When executing the process, the computer reads a program stored in its own recording medium and executes a process according to the read program. As another execution form of the program, the computer may directly read the program from a portable recording medium and execute processing according to the program, and the program is transferred from the server computer to the computer. Each time, the processing according to the received program may be executed sequentially. Also, the program is not transferred from the server computer to the computer, and the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes the processing function only by the execution instruction and result acquisition. It is good. Note that the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (data that is not a direct command to the computer but has a property that defines the processing of the computer).

  In this embodiment, the present apparatus is configured by executing a predetermined program on a computer. However, at least a part of these processing contents may be realized by hardware.

100, 200, 300, 400, 500 Encoder 100 ', 200', 250 ', 270', 300 ', 350', 370 'Decoder 400', 450 ', 470', 500 ', 550', 570 'Decoder 101 Frequency domain transform unit 102 Normalized value calculation unit 103 Normalized value quantization unit 104 First vector quantization unit 105, 205, 312 Error calculation units 106, 206, 306, 406 Second vector quantization unit 107 157 Normalized value decoding unit 108 First vector decoding unit 109, 209, 309, 409 Second vector decoding unit 110, 210, 275, 310, 375 Error correction unit 111 Time domain conversion unit 313 M value calculation unit 415, 416 515, 516 Normalization value recalculation part

Claims (14)

A normalization value calculation step for obtaining a normalization value for a signal sample to be encoded for each predetermined number of samples (hereinafter referred to as “first encoding target signal”);
A normalized value quantization step of obtaining a normalized quantization value obtained by quantizing the normalized value and a normalized value quantization index corresponding to the quantized normalized value;
First normalized signal is obtained by normalizing the first coding target signal using the quantized normalized value, and a plurality of samples of the normalized first coding target signal are vector quantized. A first vector quantization step for obtaining a sequence of one quantized signal and a first vector quantization index corresponding to the sequence of the first quantized signal;
When the value of the first quantized signal is 0, a weighted error signal between the absolute value of the first encoding target signal and the quantized normalized value is obtained, and the value of the first quantized signal is other than 0. In this case, an error calculation step for obtaining an error signal between the absolute value of the first encoding target signal and the absolute value of the first quantized signal;
A second vector quantization step of obtaining a second vector quantization index by vector-quantizing a plurality of samples of the error signal or a signal corresponding to the error signal;
An encoding method comprising:   A normalization value calculation step for obtaining a normalization value for a signal sample to be encoded for each predetermined number of samples (hereinafter referred to as “first encoding target signal”);
  A normalized value quantization step of obtaining a normalized quantization value obtained by quantizing the normalized value and a normalized value quantization index corresponding to the quantized normalized value;
  First normalized signal is obtained by normalizing the first coding target signal using the quantized normalized value, and a plurality of samples of the normalized first coding target signal are vector quantized. A first vector quantization step for obtaining a sequence of one quantized signal and a first vector quantization index corresponding to the sequence of the first quantized signal;
  An error calculating step for obtaining only a weighted error signal between the absolute value of the first encoding target signal and the quantized normalized value when the value of the first quantized signal is 0;
  A second vector quantization step of obtaining a second vector quantization index by vector-quantizing a plurality of samples of the error signal or a signal corresponding to the error signal;
  An encoding method comprising: An encoding method according to claim 2, wherein
The quantization normalized value is X ⁻ , the predetermined number of samples is L, the total power of the first quantized signal is tmp, the number of the first quantized signals having a value of 0 is M, and a recalculated normalized value E ^- when the,

A normalized value recalculation step for obtaining a recalculated normalized value E ⁻ which is
In the second vector quantization step, the error signal is normalized using the recalculated normalization value to obtain a normalized error signal, and the normalized error signal is vector-quantized by collecting a plurality of samples, and the second vector quantum is obtained. A coding method characterized by obtaining a coding index. An encoding method according to claim 2, wherein
The quantized normalization value X ^-, the predetermined number of samples L, and total power of the first quantized signal tmp, the first tmp2 the sum of the absolute values of the quantized signal, the value of 0 first When the number of 1 quantized signals is M, the first recalculation value is E ^- ₁ and the second recalculation value is E ^- ₂ ,

E the first recalculated as a relation ^- ₁ and the second recalculated E ^- ₂ a determined,
A normalized value recalculation step in which any one of the first recalculated value, the second recalculated value, and the quantized normalized value is a recalculated normalized value;
The second vector quantization step obtains a normalized error signal by normalizing the error signal using the recalculated normalized value, and vector-quantizes the normalized error signal by collecting a plurality of samples, and An encoding method characterized by obtaining a vector quantization index. A decoding method for obtaining a decoded signal from a code including a first vector quantization index, a second vector quantization index, and a normalized value quantization index,
A normalized value decoding step of obtaining a decoded normalized value corresponding to the normalized value quantization index and outputting as a decoded quantized normalized value;
First, a normalized first decoded quantized signal corresponding to the first vector quantization index is obtained, and a first decoded quantized signal is obtained using the normalized first decoded quantized signal and the decoded quantized normalized value. One vector decoding step;
First, a normalized second decoded quantized signal corresponding to the second vector quantization index is obtained, and a second decoded quantized signal is obtained using the normalized second decoded quantized signal and the decoded quantized normalized value. A two-vector decoding step;
When the value of the first decoded quantized signal is 0, the calculation result using the weighted addition of the decoded quantized normalized value and the second decoded quantized signal is multiplied by 1 or −1 randomly generated, When the multiplication result is a decoded signal and the value of the first decoded quantized signal is other than 0, an error using the calculation result using the sum of the first decoded quantized signal and the second decoded quantized signal as the decoded signal A decoding method comprising: a correction step. The decoding method according to claim 5, wherein
The error correction step comprises:
When the value of the first decoded quantized signal is 0, a result obtained by multiplying the weighted addition of the decoded quantized normalized value and the second decoded quantized signal by randomly generated 1 or −1 is used as a decoded signal, When the value of the first decoded quantized signal is negative, the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal is multiplied by −1 as a decoded signal, and the first decoded quantum When the value of the quantized signal is positive, the decoded signal is a sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal. The decoding method according to claim 6, wherein
The normalized value decoding step comprises:
If the decoding normalization value of the previous frame is 0, the decoding normalization value of the current frame is output as the decoding quantization normalization value,
If the decoding normalization value of the previous frame is not 0, the result of the weighted addition of the decoding normalization value of the current frame and the decoding normalization value of the previous frame is output as the decoding quantization normalization value. A characteristic decoding method. The decoding method according to claim 5, wherein
The error correction step
When the value of the first decoded quantized signal is 0 and the decoded quantized normalized value of one frame past is 0, the result of weighted addition of the decoded quantized normalized value of the current frame and the second decoded quantized signal The result of multiplying randomly generated 1 or -1 is the decoded signal,
When the value of the first decoded quantized signal is 0 and the decoded quantized normalized value of one frame past is other than 0, the decoded quantized normalized value of the current frame, the second decoded quantized signal, and the past of one frame A result obtained by multiplying a result of weighted addition with the decoded quantization normalized value of 1 or −1 randomly generated as a decoded signal,
When the value of the first decoded quantized signal is negative, the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal is multiplied by −1 as a decoded signal, and the first decoded quantum When the value of the quantized signal is positive, the decoding method includes: using the sum of the absolute value of the first decoded quantized signal and the second decoded quantized signal as a decoded signal.   A decoding method for obtaining a decoded signal from a code including a first vector quantization index, a second vector quantization index, and a normalized value quantization index,
  A normalized value decoding step of obtaining a decoded normalized value corresponding to the normalized value quantization index and outputting as a decoded quantized normalized value;
  First, a normalized first decoded quantized signal corresponding to the first vector quantization index is obtained, and a first decoded quantized signal is obtained using the normalized first decoded quantized signal and the decoded quantized normalized value. One vector decoding step;
  First, a normalized second decoded quantized signal corresponding to the second vector quantization index is obtained, and a second decoded quantized signal is obtained using the normalized second decoded quantized signal and the decoded quantized normalized value. A two-vector decoding step;
  When the value of the first decoded quantized signal is 0, the calculation result using the weighted addition of the decoded quantized normalized value and the second decoded quantized signal is multiplied by 1 or −1 randomly generated, An error correction step in which the multiplication result is a decoded signal, and when the value of the first decoded quantized signal is other than 0, the first decoded quantized signal is the decoded signal;
  A decryption method. The decoding method according to claim 9, wherein
An M value calculation step for obtaining the number M of the first decoded quantized signals having a value of 0;
The decoded quantized normalized value is X ⁻ , the number of the first decoded quantized signals is L, the total power of the first decoded quantized signals is tmp, and the number of the first decoded quantized signals is 0. Is M and the recalculated normalization value E ⁻

A normalized value recalculation step for obtaining a recalculated normalized value E ⁻ which is
The second vector decoding step obtains a normalized second decoded quantized signal corresponding to the second vector quantization index, and uses the normalized second decoded quantized signal and the recalculated normalized value. 2. A decoding method characterized by obtaining a decoded quantized signal. The decoding method according to claim 9, wherein
An M value calculation step for obtaining the number M of the first decoded quantized signals having a value of 0;
The decoded quantization normalized value is X ⁻ , the number of the first decoded quantized signals is L, the total power of the first decoded quantized signals is tmp, and the total absolute value of the first quantized signals is tmp 2 , the number of the first decoded quantized signal values are 0 M, the first recalculated E ^- _1, the second recalculated E ^- when a _2,

E the first recalculated as a relation ^- ₁ and the second recalculated E ^- ₂ a determined,
A normalized value recalculation step in which any one of the first recalculated value, the second recalculated value, and the decoded quantized normalized value is a recalculated normalized value;
The second vector decoding step obtains a normalized second decoded quantized signal corresponding to the second vector quantization index, and uses the normalized second decoded quantized signal and the recalculated normalized value. 2. A decoding method characterized by obtaining a decoded quantized signal. A normalized value calculation unit for obtaining a normalized value for a signal sample to be encoded for each predetermined number of samples (hereinafter referred to as “first encoding target signal”);
A normalized value quantizing unit that obtains a normalized quantization value that is obtained by quantizing the normalized value and a normalized value quantization index that corresponds to the quantized normalized value;
First normalized signal is obtained by normalizing the first coding target signal using the quantized normalized value, and a plurality of samples of the normalized first coding target signal are vector quantized. A first vector quantization unit for obtaining a sequence of 1 quantized signal and a first vector quantization index corresponding to the sequence of the first quantized signal;
When the value of the first quantized signal is 0, a weighted error signal between the absolute value of the first encoding target signal and the quantized normalized value is obtained, and the value of the first quantized signal is other than 0 An error calculation unit for obtaining an error signal between the absolute value of the first encoding target signal and the absolute value of the first quantized signal;
A second vector quantization unit for obtaining a second vector quantization index by vector-quantizing a plurality of samples of the error signal or a signal corresponding to the error signal;
An encoder comprising: A decoder for obtaining a decoded signal from a code including a first vector quantization index, a second vector quantization index, and a normalized value quantization index,
A normalized value decoding unit that obtains a decoded normalized value corresponding to the normalized value quantization index and outputs the decoded normalized value;
First, a normalized first decoded quantized signal corresponding to the first vector quantization index is obtained, and a first decoded quantized signal is obtained using the normalized first decoded quantized signal and the decoded quantized normalized value. A one-vector decoding unit;
First, a normalized second decoded quantized signal corresponding to the second vector quantization index is obtained, and a second decoded quantized signal is obtained using the normalized second decoded quantized signal and the decoded quantized normalized value. A two-vector decoding unit;
When the value of the first decoded quantized signal is 0, the calculation result using the weighted addition of the decoded quantized normalized value and the second decoded quantized signal is multiplied by 1 or −1 randomly generated, When the multiplication result is a decoded signal and the value of the first decoded quantized signal is other than 0, an error using the calculation result using the sum of the first decoded quantized signal and the second decoded quantized signal as the decoded signal A decoder comprising: a correction unit.   The program which performs each step of the method in any one of Claim 1 to 11 by computer.

Images