JP3144284B2 - Audio coding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress deterioration of sound quality by using a comparatively small amount of computation even when the bit rate is low by providing a code book which quantizes one of the parameters of amplitude and position of non-zero pulses in a lump and determining the other parameter while searching this. SOLUTION: Using an amplitude code book 13 which is used to quantize one of the amplitude parameters of pulses constituting a sound source in a lump, a sound source quantizing circuit 12 determines and quantizes the other parameter of plural non-zero pulses. Or, the circuit 12 calculates plural sets of the positions, searches the code book for the plural sets of positions of non- zero pulses and selects the combination of the set, which has the position making the value of predetermined formula maximum or minimum, and the code vector to quantize the sound source.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech coding apparatus, and more particularly to a speech coding apparatus for coding a speech signal at a low bit rate with high quality.

[0002]

2. Description of the Related Art A speech encoding device is used opposite to a speech decoding device, and a speech decoding device decodes speech encoded by the speech encoding device. Here, as a method of encoding a speech signal with high efficiency, for example, M. Schroeder and B. Atal
Etc. are IEEE Proceedings (IEEE Pro
c.) ICASSP-85, Code Excited Linear Prediction: High Quality Speech at Very Low Bit Rates, pp. 937-940, 1985.
Peech at very lowbit rates) (Reference 1), and Claizin (Kleijn) et al., IEEE Proc. IACSSP-88,
Improve Speech Quality and Efficient Vector Quantization in SLP, 1988, pp. 155-158.
d speech quality and efficient vector quantization
in CELP (Code Excited Linear Prediction Co.)
ding) is known. In this method, the transmitting side performs linear prediction (LP) from a speech signal every frame (for example, 20 ms).
C) Using analysis, extract a spectral parameter representing a spectral characteristic of the audio signal, further divide the frame into a plurality of subframes (for example, 5 ms), and apply an adaptive codebook to each subframe based on a past sound source signal. (A delay parameter and a gain parameter corresponding to a pitch period) are extracted, a speech signal of a corresponding subframe is pitch-predicted by an adaptive codebook, and a residual signal obtained by pitch prediction is determined in advance. The excitation signal is quantized by selecting an optimal excitation code vector from an excitation codebook (vector quantization codebook) composed of various types of noise signals and calculating an optimal gain.
The excitation code vector is selected in such a manner as to minimize the error power between the signal synthesized from the selected noise signal and the above-mentioned residual signal. Then, the index and gain indicating the type of the selected code vector, the spectrum parameter and the parameter of the adaptive codebook are combined and transmitted by the multiplexer unit. Description on the receiving side is omitted.

[0003]

The above-described conventional speech coding apparatus requires a large amount of calculation to select an optimal excitation code vector from an excitation codebook.
According to the methods of Documents 1 and 2, in order to select a sound source code vector, filtering or convolution operation is once performed on each code vector, and this operation is repeated by the number of code vectors stored in the code book. It is due to For example, when the number of bits in the codebook is B and the number of dimensions is N, the amount of operation is N per second, assuming that the filter or impulse response length in the filtering or convolution operation is K.
× K × 2 ^B × 8000 / N is required. As an example, if B = 10, N = 40 and K = 10, 81,920,000 operations are required per second, which is extremely large and economically infeasible.

Here, various methods have been proposed so far when considering a method of reducing the amount of calculation required for the sound source codebook search.

For example, ACELP (Argebraic Code Exci
ted Linear Prediction) has been proposed. As a concrete method of this, C. Laflamme (C. Laflamme) etc.
ICASSP-91, pp. 13-16, 16K BB Wideband Speech Coding Technique Based on Algibreak CLP (16 kbps wideband speech codingtechnique base)
d on algebraic CELP) (Reference 3). According to the method of Document 3, the sound source signal is represented by a plurality of pulses, and the position of each pulse is represented by a predetermined number of bits and transmitted. Here, since the amplitude of each pulse is limited to +1.0 or -1.0, it is possible to greatly reduce the calculation amount of the pulse search.

[0006] The method disclosed in Reference 3 can greatly reduce the amount of computation, but has the problem that the sound quality is not sufficient. The reason is that each pulse has only positive and negative polarities and the absolute value amplitude is always 1.0 regardless of the position of the pulse. Had deteriorated.

An object of the present invention is to provide a speech coding apparatus capable of suppressing deterioration of sound quality even with a low bit rate even with a relatively small amount of calculation.

[0008]

According to the present invention, there is provided a speech coding apparatus comprising: a spectrum parameter calculating section for obtaining a spectrum parameter from an input speech signal and quantizing the spectrum parameter; and using the spectrum parameter to generate a sound source signal of the speech signal. In a speech coding apparatus having a sound source quantization unit for quantizing and outputting, a sound source is constituted by a plurality of non-zero pulses, and one of parameters of the amplitude or position of the non-zero pulse is quantized collectively. has a codebook, while the sound source quantization section searches the codebook, the code
Any position or amplitude not represented in the book
It has a function of obtaining parameters and quantizing the non-zero pulse.

A speech encoding apparatus according to the present invention comprises a spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter from an input speech signal, and a sound source for quantizing and outputting a speech signal of the speech signal using the spectrum parameter. In a speech encoding device having a quantization unit, a sound source is configured by a plurality of non-zero pulses, and a codebook is provided for quantizing the amplitude of the non-zero pulses collectively. The position of the non-zero pulse is calculated for a plurality of sets, the code book is searched for the positions of the plurality of sets, and a set having a position where the value of a given equation is one of a maximum and a minimum is set. And the pulse amplitude
This is a configuration having a function of selecting a combination with a code vector of a code book to be quantized and quantizing a sound source signal.

A speech coding apparatus according to the present invention comprises a spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter at predetermined time intervals from an input speech signal, and quantizing a sound source signal of the speech signal using the spectrum parameter. A sound source comprising a plurality of non-zero pulses, a code book for quantizing the amplitudes of the non-zero pulses together, and A mode discrimination circuit that discriminates a mode by extracting a feature amount, and calculates a position of the non-zero pulse for a plurality of sets when a discrimination result of the mode discrimination circuit is a predetermined mode. the codebook searches for each of the positions of the set, the set having a position to either one maximum and minimum values of a predetermined equation, Pas Code to be quantized together the amplitude of the scan
In this configuration, the sound source signal is quantized by selecting a combination with the code vector of the book .

A speech coding apparatus according to the present invention includes a spectrum parameter calculating section for obtaining and quantizing a spectrum parameter at predetermined time intervals from an input speech signal, and quantizing a sound source signal of the speech signal using the spectrum parameter. And a sound source quantization unit having a sound source quantization unit for outputting the sound source.
A code book that collectively quantizes the amplitude of the pulse, and a mode discriminating circuit that discriminates a mode by extracting a feature amount from the audio signal, wherein the mode discriminating result of the mode discriminating circuit is a predetermined mode In the case of, the position of the pulse is calculated for at least one set, the code book is searched for a position of a set having a position where the value of a given equation is one of a maximum and a minimum,
The source signal is quantized by selecting a combination of a good position set and a code vector, and in another predetermined mode, the source is a linear combination of a plurality of pulses and a source code vector selected from a source code book. In this configuration, the pulse and the excitation code vector are searched and quantized.

A speech encoding apparatus according to the present invention comprises a frame dividing circuit for dividing an inputted speech signal into frames of a predetermined time length, and a sub-segment for dividing the speech signal of the frame into sub-frames having a shorter time length than the frame. A frame dividing circuit, receiving a sound signal of a series of frames output from the frame dividing circuit, cutting out the sound signal by applying a window longer than the time length of the subframe to the sound signal of at least one subframe; A spectral parameter calculating circuit for calculating spectral parameters up to a predetermined order, and vector quantization of a line spectrum pair parameter quantized by a predetermined subframe calculated by the spectrum parameter calculating circuit using a line spectrum pair parameter codebook. A spectral parameter quantization circuit, A perceptual weighting circuit that receives the linear prediction coefficients of the plurality of subframes calculated by the spectral parameter calculation circuit, performs perceptual weighting on the audio signal of each subframe, and outputs a perceptual weighting signal; A response signal calculation circuit for inputting the linear prediction coefficient of the subframe and the linear prediction coefficient restored by the spectrum parameter quantization circuit for each subframe, calculating a response signal for one subframe, and outputting the response signal to a subtractor; An impulse response calculation circuit that receives the linear prediction coefficient restored by the parameter quantization circuit and calculates a predetermined score for the impulse response of the auditory weighting filter; a past sound source signal that is fed back from the output side; Input the impulse response of the weighting filter and The other of a plurality of non-zero pulses using an adaptive codebook circuit that calculates the delay and outputs an index representing the delay, and a codebook that collectively quantizes one of the parameters of the amplitude and position of the pulse constituting the sound source Search and select a sound source quantization circuit and gain codebook for
Given the selected amplitude code vector and location,
Gain code vector that maximizes or minimizes the
Select vector, and the gain quantization circuit to output an index representing the selected gain code vector to the multiplexer, the input excitation signal read a code vector corresponding to the index output of the gain quantization circuit And a weighting signal calculation circuit to be obtained.

[0013] In the speech coding apparatus according to the present invention, the sound source quantizing section may preliminarily limit a position where at least one pulse can be taken.

In the speech coding apparatus according to the present invention, the excitation quantization section preliminarily selects a plurality of code vectors from the codebook, and then searches for the preselected code vector to obtain the other parameter and quantizes the other parameter. You may.

[Operation] In the first aspect of the present invention, the sound source quantization unit has a code book for quantizing one of the amplitude parameter and the position parameter collectively for a plurality of non-zero pulses. In the following, a description will be given assuming that the code book includes a code book for quantizing a plurality of pulse amplitude parameters collectively.

At regular intervals, M pulses are generated as a sound source. The time length is N samples. The amplitude of the i th pulse, position, respectively, g _i, and m _i. At this time, the sound source signal can be expressed by the following equation.

[0017]

Assuming that the k-th amplitude code vector stored in the codebook is _g'ik and the amplitude is quantized, the sound source is

[0019]

It can be expressed as Here, B is the number of bits of the codebook for quantizing the amplitude. At this time, the distortion between the signal reproduced using equation (2) and the input audio signal is

[0021]

Here, x _w (n) and h _w (n) are a perceptually weighted speech signal and a perceptually weighted impulse response, respectively, which will be described in the following embodiments.

[0023] To minimize equation (3) may be determined to k-th code vector and a combination of the position m _i that maximizes the following equation.

[0024]

Here, s _wk (m _i ) can be expressed by the following equation.

[0026]

Therefore, for each amplitude code vector,
Calculate the position and find the combination that maximizes equation (4).

According to a second aspect, in the first aspect, a position where at least one pulse can be set is set in advance. Although various limiting methods are conceivable, for example, the method in ACELP of the above-mentioned Document 3 can be used. As an example, if N = 40 and M = 5, the position of each pulse can be limited as shown in Table 1.

[0029]

[Table 1]

According to this, since the positions of each pulse are limited to eight types, the number of combinations of positions is greatly reduced, and the amount of calculation for calculating the equation (4) is smaller than that of the first invention. Can be reduced.

In the third invention, instead of calculating Equation (4) for all 2B types of code vectors included in the code book, a plurality of code vectors are preliminarily selected and the preselected code vectors are preliminarily selected. By calculating equation (4) only for the code vector, the amount of calculation is reduced.

In the fourth invention, M is used as a code book.
It has a codebook that quantizes the amplitudes of the pulses collectively. Further, the positions of the M pulses are calculated for a plurality of sets, and for each of the plurality of sets of positions, Equation (4) is calculated for the code vector in the codebook, and the position and code for maximizing Equation (4) are calculated. Select a vector combination.

According to a fifth aspect, in the fourth aspect, the positions where at least one pulse can be taken are limited in advance, as in the second aspect.

In the sixth aspect, the mode is discriminated by extracting the characteristic amount from the audio signal, and in the case of a predetermined mode, the same processing as in the fourth aspect is performed.

According to a seventh aspect, in the sixth aspect, at least one possible pulse position is limited in advance, as in the second aspect.

In the eighth invention, the sound source signal is switched according to the mode. That is, in the predetermined mode, the sound source is represented by a plurality of pulses as in the sixth invention, and in the other predetermined modes, the sound source signal is expressed by the plurality of pulses and the sound source selected from the sound source codebook. Expressed as a linear combination with the code vector. For example, it can be expressed as the following equation.

[0037]

Here, c _j (n) is the j-th sound source code vector stored in the sound source code book. G ₁ and G ₂ are the respective gains. R is the number of bits in the sound source codebook.

Also, in the predetermined mode,
The same processing as in the sixth invention is performed.

According to a ninth aspect, in the eighth aspect, the position where at least one pulse can be taken is limited in advance, as in the second aspect.

[0041]

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing a first embodiment of the present invention.

The speech coding apparatus 1 according to the first embodiment of the present invention comprises a frame dividing circuit 2 for dividing an inputted speech signal into frames of a predetermined time length, and a speech signal having a shorter time than the frame. A sub-frame division circuit 3 for dividing into a long sub-frame, and receiving at least one audio signal of a series of frames output from the frame division circuit 2
A spectrum parameter calculation circuit 4 that cuts out the audio signal by applying a window longer than the subframe time length to the audio signal of one subframe and calculates spectral parameters to a predetermined order, and a line spectrum versus parameter codebook (Hereinafter referred to as LSP codebook) 6
A parameter quantization circuit 5 for vector-quantizing an LSP parameter quantized in a predetermined subframe calculated by the spectrum parameter calculation circuit 4 using
A perceptual weighting circuit 7 that receives the linear prediction coefficients of a plurality of subframes calculated by the spectrum parameter calculation circuit 4 and weights the perceptual weight of the audio signal of each subframe and outputs a perceptual weighting signal; Are input for each sub-frame, the linear prediction coefficients restored by the spectral parameter quantization circuit 5 are calculated for each sub-frame, the response signal is calculated for one sub-frame, and output to the subtractor 8 A signal calculation circuit 9, an impulse response calculation circuit 10 that receives the linear prediction coefficient restored by the spectrum parameter quantization circuit 5 and calculates a predetermined score of an impulse response of the perceptual weighting filter; The output signal of the subtractor 8 and the impulse response of the perceptual weighting filter are input and An adaptive codebook circuit 11 for outputting a Intex representing a delay determined delay corresponding to the switch,
A sound source quantization circuit 12 for obtaining and quantizing the other parameters of a plurality of non-zero pulses using an amplitude codebook 13 for quantizing one of the amplitude parameters of the pulses constituting the sound source collectively; A gain quantization circuit 14 that reads a gain code vector from the code book 15 and selects one gain code vector from the amplitude code vector and the position, and outputs an index representing the selected gain code vector to the multiplexer 16;
The output of the gain quantization circuit 14 is inputted, a code vector corresponding to the index is read out from the index, and a weighting signal calculation circuit 17 for obtaining a drive excitation signal.

Next, the operation of the present apparatus will be described.

First, an audio signal is input from an input terminal, and the frame dividing circuit 2 converts the audio signal into a frame (for example, 10 m).
s), and the subframe division circuit 3 divides the audio signal of the frame into subframes (for example, 2.5 ms) shorter than the frame. In the spectrum parameter calculation circuit 4, a window (for example, 24 ms) longer than the subframe length is provided for at least one subframe audio signal.
, The speech is cut out, and the spectral parameters are calculated in a predetermined order (for example, P = 10th order). Here, a well-known LPC analysis, a Burg analysis, or the like can be used for calculating the spectrum parameters. Here, Burg analysis is used. Berg (B
urg) For details of the analysis, see the book “Corona Analysis and System Identification” written by Nakamizo (Corona Publishing Co., 1988).
The description is omitted since it is described on pages 82 to 87 (Document 5) and the like.

Further, in the spectrum parameter calculation circuit 4, the linear prediction coefficient α calculated by the Burg method
_i (i = 1,..., 10) is converted to LSP parameters suitable for quantization and interpolation. Here, the conversion from the linear prediction coefficient to the LSP is performed by Sugamura et al. In a paper entitled "Speech Information Compression by Line Spectrum Pair (LSP) Speech Analysis / Synthesis Method" (Transactions of the Institute of Electronics, Information and Communication Engineers, J64-A, pp.599). -606, 1981) (Reference 5)
Can be referred to. For example, the linear prediction coefficient obtained by the Burg method in the second and fourth subframes is represented by LS
Convert to P parameter, LSP of 1st and 3rd subframe
Is obtained by linear interpolation, and the LS of the first and third sub-frames is calculated.
P is inversely transformed back to a linear prediction coefficient, and the linear prediction coefficients α _il (i = 1,..., 10, l = 1,..., 5) of the first to fourth subframes are output to the audibility weighting circuit 7. In addition, L of the fourth sub-frame
The SP is output to the spectrum parameter quantization circuit 5.

In the speckle parameter quantization circuit 5, L
The LSP parameter of a predetermined subframe is efficiently quantized using the SP record book 6, and a quantized value that minimizes the following equation is output.

[0048]

Here, LSP (i), QLSP (i) _j and W (i) are the i-th LTP and LSP codebooks 6 before quantization, respectively.
The result of the j-th code vector is the weight coefficient.

In the following, it is assumed that the LSP parameter of the fourth subframe is quantized. A well-known method can be used for the method of vector quantization of LSP parameters. Specific methods are described in, for example, JP-A-4-171500 (Reference 6), JP-A-4-363000 (Reference 7), and JP-A-5-6199 (Reference 8).
And IEE Proceedings by T. Nomura and others. Mobile multimedia
Communications (IEEE Proc. Mobile Multimedia
Communications.) 1993; On page 2.5, LSP Coding Using Viku-SV-Q-with with Interpolation in 4.0
75KPS M-LSP Coding Using VQ-SVQ Wit
h Interpolation in 4.075 kbps M-LCELP Speech Code
r), the description of which is omitted here.

The spectrum parameter quantization circuit 5
Then, the LSP parameters of the first to fourth subframes are restored based on the LSP parameters quantized in the fourth subframe. Here, the quantization LSP parameter of the fourth sub-frame of the current frame and the fourth
The LSPs of the first to third subframes are restored by linearly interpolating the quantized LSPs of the subframe. Here, after selecting one type of code vector that minimizes the error power between the LSP before quantization and the LSP after quantization, the LSPs of the first to fourth subframes can be restored by linear interpolation. In order to further improve performance, after selecting a plurality of candidates for the code vector that minimizes the error power, evaluate the cumulative distortion for each candidate, and select a combination of the candidate and the interpolation LSP that minimizes the cumulative distortion. You can make it. For details, see, for example, Japanese Patent Application No. 5-8737 (Reference 10).
Can be referred to.

The LSPs of the first to third sub-frames and the quantized LSPs of the fourth sub-frame, which have been reconstructed as described above, are converted into linear prediction coefficients α ′ _il (i = 1,..., 10, l =,. ) And outputs it to the impulse response calculation circuit 10. Further, an index representing the code vector of the quantized LSP of the fourth subframe is output to the multiplexer 16. The audibility weighting circuit 7 includes a spectrum parameter calculation circuit 4
From the linear prediction coefficient α before quantization for each subframe.
_il (i = 1,..., 10, l =,..., 5) is input, and the perceptual weighting is performed on the audio signal of the subframe based on Document 1 to output a perceptual weighting signal.

The response signal calculation circuit 9 receives the linear prediction coefficient α _il for each subframe from the spectrum parameter calculation circuit 4, and quantizes, interpolates and restores the linear prediction coefficient α _il from the spectrum parameter quantization circuit 5. α ′ _il is input for each sub-frame, a response signal with the input signal set to zero d (n) = 0 is calculated for one sub-frame using the stored value of the filter memory, and output to the subtractor 8 I do. Here, the response signal x _z (n) is represented by the following equation.

[0054]

However, when ni ≦ 0, y (ni) = p (N + (ni)) (9) x _z (ni) = s _w (N + (ni)) (10) where N is the subframe length Is shown. γ is a weight coefficient for controlling the hearing weighting amount, and is the same value as the following equation (12). s _w (n) and p (n) respectively represent the output signal of the weighting signal calculation circuit 17 and the output signal of the denominator term of the filter of the first term on the right side in Expression (12) described later.

The subtractor 8 subtracts the response signal by one subframe from the perceptual weighting signal according to the following equation, and outputs x ′ _w (n) to the adaptive codebook circuit 11. x ′ _w (n) = x _w (n) −x _z (n) (11) The impulse response calculation circuit 10 calculates in advance the impulse response h _w (n) of the auditory weighting filter whose z-transformation is expressed by the following equation. Calculation is performed for a predetermined number L, and the adaptive codebook circuit 11, the sound source quantization circuit 12, and the gain quantization circuit 14
And output to

[0057]

In the adaptive codebook circuit 11, the past sound source signal v (n) is output from the gain quantization circuit 14, the output signal x ′ _w (n) is output from the subtracter 8, and the audibility is output from the impulse response calculation circuit 10. Enter the weighted impulse response h _w (n). The delay T corresponding to the pitch is determined so as to minimize the distortion of the following expression, and an index representing the delay is output to the multiplexer 16.

[0059]

Here, y _w (n−T) = v (n−T) * h _w (n) (14), and the symbol * represents a convolution operation. The gain β is obtained according to the following equation.

[0061]

Here, in order to improve the accuracy of extracting delays for female sounds and children's voices, the delays may be determined not by integer samples but by decimal sample values. A concrete method is, for example, IACSSP by P. Kroon et al. (IEEE Proc.)
-90, 1990, Pitch Predictors with High Temporal Solution (Pit
For example, a paper (Reference 11) published under the title of "Ch predictors with high temporal resolution" can be referred to.

Further, the adaptive codebook circuit 11 performs pitch prediction according to the following equation, and outputs a prediction residual signal e _w (n) to the excitation quantization circuit 12. e _w (n) = x ′ _w (n) −βv (nT) * h _w (n) (16) In the sound source quantization circuit 12, as described in the operation, it is assumed that M pulses are emitted. . In the following, a description will be given assuming that a B-bit amplitude codebook 13 is provided in order to collectively quantize the pulse amplitude for M pulses.

The sound source quantization circuit 12 reads out the amplitude code vectors from the amplitude code book 13, applies all positions to each code vector, and selects a combination of a code vector and a position that minimizes the following expression.

[0065]

Here, h _w (n) is an auditory weighting impulse response.

[0067] To minimize equation (16) may be determined a combination of an amplitude code vector k and a position m _i that maximizes the following equation.

[0068]

Here, s _wk (m _i ) is calculated by equation (5).
Alternatively, a choice may be made to maximize: In this case, the amount of calculation required for the calculation of the numerator is reduced.

[0070]

Here,

[0072]

Is as follows.

Then, the index representing the code vector is output to the multiplexer 16. Further, the position of the pulse is quantized by a predetermined number of bits, and an index representing the position is output to the multiplexer 16.

A method of searching for a position in a pulse is described in the above-mentioned reference 3 or, for example, K. Ozaw (K. Ozaw).
a) IEE Journal of
Selected Area On Communications (I
EEE Journal of Selected Areas on Communications.),
1986, pages 133-141, A Study on Pulse Search Algorithms for Multipulse Excite Speech Coder Realization (A
study on pulse searchalgorithms for multipulse ex
cited speech coder realization.) (Reference 12).

Further, a code book for quantizing the amplitude of a plurality of pulses can be learned and stored in advance using an audio signal. Codebook learning methods include, for example, Linde et al.'S IEEE Trans.
mun.), January, 1980, pp. 84-95, An algorithm for vector quantization desig.
n,) can be referred to.

The information on the amplitude and the position is obtained by the gain quantization circuit 14.
Is output to The gain quantization circuit 14 reads the gain code vector from the gain code book 15 and selects a gain code vector for the selected amplitude code vector and position so as to minimize the following expression. Here, an example will be described in which both the gain of the adaptive codebook and the gain of the sound source expressed in pulses are simultaneously vector-quantized.

[0078]

Here, β′t, G′t are the k-th code vector in the two-dimensional gain codebook stored in the gain codebook 15. An index representing the selected gain code vector is output to the multiplexer 16.

The weighting signal calculation circuit 17 receives the respective indexes, reads out the corresponding code vectors from the indexes, and obtains the driving sound source signal v (n) based on the following equation.

[0081]

V (n) is output to the adaptive codebook circuit 11.

Next, the response signal s _w (n) is calculated for each subframe by the following equation using the output parameter of the spectrum parameter calculation circuit 4 and the output parameter of the spectrum parameter quantization circuit 5, and the response signal calculation circuit 9 is output.

[0084]

The description of the first embodiment of the present invention has been completed.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

Speech Encoding Apparatus 1 of Second Embodiment
8 is different from the first embodiment in that the operation of the sound source quantization circuit 19 reads the position of each pulse from the position storage circuit 20 storing the position of the pulse shown in the table in the section of the operation. However, only in the combination of these positions, the combination of the position and the amplitude code vector that maximizes the expression (18) or (19) is selected. This concludes the description of the second invention.

FIG. 3 is a block diagram showing a third embodiment of the present invention.

Speech coding apparatus 2 according to a third embodiment
1 is different from the first embodiment in that the preselection circuit 2
2 is newly provided, and a plurality of code vectors are preliminarily selected from the code vectors stored in the code book 13. Here, the following method is used as a preselection method. The residual signal z (n) is calculated by the following equation using the adaptive codebook output signal e _w (n) and the spectrum parameter α _i .

[0090]

Subsequently, a plurality of amplitude code vectors are preliminarily selected in the order of maximizing the equation (25) or the equation (26), and output to the sound source quantization circuit 23.

[0092]

The sound source quantization circuit 23 calculates the equation (18) or (19) for the combination of positions only for the amplitude code vector preliminarily selected, and calculates the combination of the position and the code vector that maximizes this. Is output.

FIG. 4 is a block diagram showing a fourth embodiment of the present invention.

Speech coding apparatus 2 according to a fourth embodiment
4 is different from the first embodiment in that the sound source quantization circuit 25 calculates the positions of a predetermined number M of pulses for a plurality of sets by the method of Reference 12 or Reference 3. Here, for simplicity, it is assumed that M positions are calculated for two sets.

First, the amplitude code vector is read from the amplitude code book 25 for the first set of positions, and the expression (1)
8) or (19) to select the amplitude code vectors which maximize the following equation, it calculates a first distortion D _1. Next, the second
The amplitude code vector is read from the amplitude codebook 25 for the position of the set, and the same processing as above is repeated to calculate the _second distortion D2. Next, the first distortion and the second distortion are compared, and a combination of a position giving the smaller distortion and an amplitude code vector is selected. Here, the definition of distortion follows the following formula.

[0097]

Then, an index representing the position and the amplitude code vector is output to the multiplexer 16.

FIG. 5 is a block diagram showing a fifth embodiment of the present invention.

Speech coding apparatus 2 according to a fifth embodiment
The fourth embodiment differs from the fourth embodiment in that the sound source quantization circuit 28 and the sound source quantization circuit 25 shown in FIG. The sound source quantization circuit 28 reads the limited positions from the position storage circuit 20, selects two sets of M positions from the combination of these positions, and performs the same processing as the sound source quantization circuit 25.
A combination of the position and the amplitude code vector that maximizes the expression (18) or (19) is selected. Next, the position of the pulse is obtained by the same method as in the first embodiment, and is quantized and output to the multiplexer 16 and the gain quantization circuit 14.

FIG. 6 is a block diagram showing a sixth embodiment of the present invention.

Speech coding apparatus 2 according to a sixth embodiment
9 is different from the fourth embodiment in that a mode discriminating circuit 31 is newly provided. The mode determination circuit 31
It receives a perceptual weighting signal from the perceptual weighting circuit 7 in frame units and outputs mode discrimination information to the sound source quantization circuit 30. Here, the feature amount of the current frame is used for mode determination. As the characteristic amount, for example, a pitch prediction gain averaged in a frame is used. The calculation of the pitch prediction gain uses, for example, the following equation.

[0103]

Here, L is the number of subframes included in the frame. Pi and Ei respectively represent the speech power and the pitch prediction error power in the i-th subframe.

[0105]

Here, T is an optimal delay for maximizing the prediction gain.

The frame average pitch prediction gain G is compared with a plurality of predetermined thresholds, and classified into a plurality of types of modes. As the number of modes, for example, 4 can be used. The mode discrimination circuit 31 outputs the mode discrimination information to the sound source quantization circuit 30 and the multiplexer 16. The sound source quantization circuit 30 receives the mode discrimination information, and performs the same processing as the sound source quantization circuit of FIG. 4 when the mode discrimination information indicates a predetermined mode.

FIG. 7 is a block diagram showing a seventh embodiment of the present invention.

Speech coding apparatus 2 according to a seventh embodiment
9 is different from the sixth embodiment in that the possible positions of the pulse of the sound source quantization circuit 33 and the sound source quantization circuit 30 of the sixth embodiment are limited.
The limited positions are read from the position storage circuit 20, and two sets of M positions are selected from a combination of these positions,
The same processing as the sound source quantization circuit 30 is performed, and the equation (18) or
Select the combination of the position and the amplitude code vector that maximizes (19).

FIG. 8 is a block diagram showing an eighth embodiment of the present invention.

Speech coding apparatus 3 according to the eighth embodiment
4 differs from the sixth embodiment in that two sets of gain codebooks 35 and 36 are provided and a sound source codebook 37 is newly provided. The sound source quantization circuit 38 switches the sound source according to the mode. That is, in the predetermined mode, the same operation as the sound source quantization circuit 30 of the sixth embodiment is performed, the sound source is composed of a plurality of pulses, and the combination of the pulse position and the amplitude code vector is obtained. Further, in another predetermined mode, as described in the operation, the sound source is configured by a linear combination of a plurality of pulses and a sound source code vector selected from the sound source codebook 37, and is represented by Expression (5). . Then, after searching for the pulse amplitude and position, an optimum sound source code vector is searched for. The gain quantization circuit 39 switches between the gain codebook 1 and the gain codebook 2 according to the sound source according to the mode.

FIG. 9 is a block diagram showing a ninth embodiment of the present invention.

Speech coding apparatus 4 according to the ninth embodiment
0 is different from the eighth embodiment in that the possible positions of the pulse of the sound source quantization circuit 41 and the sound source quantization circuit 38 of the eighth embodiment are limited.
A limited position is read from the position storage circuit 20, and a combination of a pulse position and an amplitude code vector is selected from a combination of these positions.

The description of the embodiment of the present invention has been completed.

The present invention is not limited to the above-described embodiment, and various modifications are possible. It is also possible to adopt a configuration in which the adaptive codebook circuit and the gain codebook are switched using the mode determination information.

The gain quantization circuit outputs a plurality of amplitude code vectors from the amplitude code book when searching for a gain code vector so as to minimize equation (21).
For each code vector, a combination of an amplitude code vector and a gain code vector that minimizes Equation (21) can be selected. In addition, when searching for an amplitude code vector using Equations (18) and (19), performance can be further improved by performing a search while orthogonalizing with an adaptive code vector.

The orthogonalization is performed as follows. q _k (n) = s _wk (n)-[Ψ _k / Υ] b _w (n) (31) where

[0118]

Here, b _w (n) is a signal weighted and reproduced by the adaptive code vector, and b _w (n) = βv (nT) * h _w (n) (34).

As a result of the orthogonalization, the term of the adaptive codebook disappears, and the amplitude code vector maximizing the following equation or equation (36) may be selected.

[0121]

Here,

[0123]

Is as follows.

[0125]

As described above, according to the present invention, the sound source in the sound source quantization unit is composed of a plurality of pulses, and has a code book for quantizing one of the amplitude and position of the pulses collectively. , The other parameter is calculated while searching the codebook, so that it is compared with the conventional method,
Even at the same bit rate, higher sound quality can be obtained with a relatively small amount of calculation. According to the present invention, there is provided a code book for quantizing pulse amplitudes collectively,
After calculating a plurality of sets of pulse positions, the set of each position and the amplitude codebook are searched and the best combination is selected, so that there is an effect that higher sound quality can be obtained as compared with the conventional method. . Furthermore, according to the present invention, the sound source is represented by a plurality of pulses or a linear combination of the plurality of pulses and a sound source code vector selected from the sound source codebook according to the mode. Therefore, there is also an effect that a better sound quality than before can be obtained.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is a block diagram showing a second embodiment of the present invention.

FIG. 3 is a block diagram showing a third embodiment of the present invention.

FIG. 4 is a block diagram showing a fourth embodiment of the present invention.

FIG. 5 is a block diagram showing a fifth embodiment of the present invention.

FIG. 6 is a block diagram showing a sixth embodiment of the present invention.

FIG. 7 is a block diagram showing a seventh embodiment of the present invention.

FIG. 8 is a block diagram showing an eighth embodiment of the present invention.

FIG. 9 is a block diagram showing a ninth embodiment of the present invention.

[Explanation of symbols]

1,18,21,24,27,29,32,34,40
Speech coding apparatus 2 Frame division circuit 3 Subframe division circuit 4 Spectrum parameter calculation circuit 5 Spectrum parameter quantization circuit 6 Line spectrum pair parameter codebook (LSP)
Codebook) 7 Perception weighting circuit 8 Subtractor 9 Response signal calculation circuit 10 Impulse response calculation circuit 11 Adaptive codebook circuit 12, 19, 23, 25, 28, 30, 33, 38, 4
Reference Signs List 1 sound source quantization circuit 13, 26 amplitude code book 14, 39 gain quantization circuit 15, 35, 36 gain code book 16 multiplexer 17 weighting signal calculation circuit 20 position storage circuit 22 preliminary selection circuit 31 mode discrimination circuit 37 sound source code book

Claims (10)

(57) [Claims] 1. A speech system comprising: a spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter from an input speech signal; and a sound source quantization unit for quantizing and outputting a speech signal of the speech signal using the spectrum parameter. In the encoding device, the sound source is composed of a plurality of non-zero pulses, and has a codebook for quantizing one of the parameters of the amplitude or position of the non-zero pulse collectively, and the sound source quantization unit is configured to perform the While searching the codebook, position or amplitude not represented in said codebook
Speech coding apparatus characterized by having a function to quantize the pulse of the non-zero determined one of the parameters. 2. The speech coding apparatus according to claim 1, wherein the sound source quantization section previously limits a position where at least one pulse can be taken. 3. The sound source quantizing unit preliminarily selects a plurality of code vectors from a codebook in a descending order of a value of a given equation or in a descending order,
For each of the preselected code vectors, a pulse
Determine either the amplitude or position parameters of
Maximize or minimize the value of a given expression
Code vector and parameter combinations
Speech encoding apparatus according to claim 1, wherein the Melco. 4. A speech having a spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter from an input speech signal, and a sound source quantization unit for quantizing and outputting a speech signal of the speech signal using the spectrum parameter. In the encoding device, the sound source is composed of a plurality of non-zero pulses, and the code book has a code book for quantizing the amplitude of the non-zero pulses collectively. Is calculated for a plurality of sets, the code book is searched for the positions of the plurality of sets, and a set having a position where the value of a given equation is one of a maximum and a minimum, and a pulse amplitude are summarized. Code to quantize
A speech coding apparatus having a function of selecting a combination with a code vector of a book and quantizing a sound source signal. 5. The speech encoding apparatus according to claim 4, wherein a position where at least one pulse can be taken in the excitation quantization unit is limited in advance. 6. A spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter at predetermined time intervals from an input speech signal, and a sound source quantization unit for quantizing and outputting a speech signal of the speech signal using the spectrum parameter. And a codebook in which the sound source is composed of a plurality of non-zero pulses, the amplitude of the non-zero pulses is collectively quantized, and a feature amount is extracted from the audio signal. And a mode discriminating circuit for discriminating the position of the non-zero pulse for a plurality of sets when the discrimination result of the mode discriminating circuit is a predetermined mode, for each position of the plurality of sets. The code book is searched, and a set having a position where the value of a predetermined expression is one of a maximum and a minimum is determined .
Code vector of a codebook that quantizes the amplitude at once
Le a speech coding apparatus characterized by quantizing a sound source signal by selecting a combination of. 7. The speech coding apparatus according to claim 6, wherein the sound source quantization section preliminarily limits a position where at least one pulse can be taken. 8. A spectrum parameter calculation unit for obtaining and quantizing a spectrum parameter at predetermined time intervals from an input speech signal, and a sound source quantization unit for quantizing and outputting a speech signal of the speech signal using the spectrum parameter. Wherein the sound source is composed of a plurality of non-zero pulses, and a codebook that collectively quantizes the amplitude of the pulses and a feature amount is extracted from the audio signal to determine a mode. A mode discriminating circuit,
A set having at least one set of the positions of the pulses when the determination result of the mode determination circuit is a predetermined mode, and setting the value of a given expression to a maximum or a minimum. Search the code book for the position, quantize the sound source signal by selecting a combination of a good set of positions and a code vector, and in another predetermined mode, the sound source is A speech coding apparatus having a sound source quantization unit, represented by a linear combination of a sound source code vector selected from a codebook, and searching and quantizing the pulse and the sound source code vector. 9. The speech coding apparatus according to claim 8, wherein the sound source quantization section previously limits a position where at least one pulse can be taken. 10. A frame division circuit that divides an input audio signal into frames of a predetermined time length, a subframe division circuit that divides an audio signal of the frame into subframes of a time length shorter than a frame, A speech signal of a series of frames output from the dividing circuit is received, and a speech signal is cut out by applying a window longer than the time length of the subframe to the speech signal of at least one subframe, and a spectrum parameter is predetermined. A spectrum parameter calculation circuit that calculates up to the order, a spectrum parameter quantization circuit that vector-quantizes a line spectrum pair parameter quantized in a predetermined subframe calculated by the spectrum parameter calculation circuit using a line spectrum pair parameter codebook, , The spectral parameter calculation A perceptual weighting circuit that receives a linear prediction coefficient of a plurality of subframes calculated by the circuit, performs perceptual weighting on an audio signal of each subframe and outputs a perceptual weighting signal, and a plurality of subframes calculated by the spectrum parameter calculating circuit. A response signal calculation circuit that inputs a linear prediction coefficient of the sub-frame and a linear prediction coefficient restored by the spectrum parameter quantization circuit for each subframe, calculates a response signal for one subframe, and outputs the response signal to a subtractor; An impulse response calculation circuit that receives the linear prediction coefficient restored by the circuit and calculates a predetermined score for the impulse response of the perceptual weighting filter; Input impulse response and calculate delay corresponding to pitch to display delay Calculate and quantize the other parameter of a plurality of non-zero pulses using an adaptive codebook circuit that outputs the intex and a codebook that collectively quantizes one of the parameters of the amplitude and position of the pulse constituting the sound source and the sound source quantization circuit, a gate
Search the in-code book and select the selected amplitude code vector
Maximize or maximize a given expression for torque and position.
A gain quantization circuit that selects a gain code vector to be reduced , and outputs an index representing the selected gain code vector to a multiplexer, and an output of the gain quantization circuit that is input and a corresponding code vector is obtained from the index. And a weighting signal calculation circuit for reading out a driving excitation signal.