JPH09269798A - Voice coding method and voice decoding method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a voice coding method capable of obtaining the voice of high quality by expressing the pitch cycle of a voice signal correctly. SOLUTION: The pitch cycle information of a voice signal to be inputted from an input terminal 101 is analyzed in a pitch information analyzing part 103 and the information of this pitch cycle are applied to an auditory sensation weighting filter 104 for making a quantization noise to be hardly heard by utilizing a masking effect. Then, at the time of coding the information of the pitch cycle of the input voice signal by searching an adaptive code book 108 by a minimum distortion searching part 116 based on the voice signal passing through the auditor sensation weighting filter 104, the analysis range of pitch cycles in the pitch information analyzing part 103 to be applied to a pitch filter is set boarder with respect to the range of pitch cycles which can be expressed by the coded data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech coding method for compressing and coding a speech signal and a speech decoding method for decoding an original speech signal from encoded data.

[0002]

2. Description of the Related Art A technique for compressing and encoding a voice signal at a low bit rate with high efficiency is an important technique for effective use of radio waves and reduction of communication cost in mobile communications such as car telephones and in-house communications. Is. A CELP (Code Excited Linear Prediction) method is known as a voice encoding method capable of performing high-quality voice synthesis at a bit rate of 8 kbps or less. This CELP method was introduced by Mr. MR Schroeder and Mr. BSAtal of AT & T Bell Labs in "Code-Excit.
ed Linear Prediction (CELP) High-Quality Speech at
Very low Bit Rates ”, Proc.ICASSP; 1985, pp.937-939”
Since it was announced in (Reference 1), it has attracted attention as a method capable of synthesizing high-quality speech, and various studies have been made on improvement of quality and reduction of calculation amount.

An adaptive codebook is one of the constituent elements of a speech coding apparatus based on the CELP method. The adaptive codebook is
The pitch prediction analysis of the input voice is performed by a closed loop operation or an analysis by synthesis. In general, the pitch prediction analysis by the adaptive codebook is 20
In many cases, a pitch period is searched for in a search range of 147 samples (128 candidates) to find a pitch period that minimizes distortion with respect to a target signal, and information on this pitch period is transmitted as 7-bit encoded data. .

However, when the input speech signal includes a pitch period outside the search range, the pitch period cannot be expressed by the adaptive codebook, so that a pitch period different from the actual one is selected, and the decoded speech has a high quality. There was a problem of deterioration. Further, if the pitch cycle search range of the adaptive codebook is widened in order to prevent this, it is necessary to increase the number of bits of encoded data representing the pitch cycle, resulting in a problem that the transmission rate becomes high. there were.

[0005]

As described above, in the conventional speech coding method, the pitch period of the predetermined search range is coded into the coded data of the predetermined number of bits. When a voice having a pitch period outside the range is input, quality deterioration occurs. In general, the range of the pitch period to be encoded is experimentally verified and a valid one is selected, but there is no guarantee that it will always be within this range, and the characteristics of the speaker and variations in the pitch period with the same speaker are not always guaranteed. Depending on the situation, there is always the danger of going outside the search range of the pitch period.

The present invention has been made to solve the above-mentioned problems of the prior art, and a speech coding method and a speech decoding method capable of accurately expressing the pitch period of a speech signal to obtain a high quality speech. The purpose is to provide a method of conversion.

[0007]

In order to solve the above problems, a speech coding method according to the present invention analyzes a pitch period of an input speech signal and suppresses a pitch period component of the input speech signal. When the pitch period of the input voice signal is searched for and encoded based on the voice signal passed through the perceptual weighting filter including the pitch filter, the pitch filter is applied to the pitch period range that can be represented by the encoded data. The feature is that the analysis range of the given pitch period is set wide. More specifically, the range of the pitch period (TL) that can be represented by the encoded data of the pitch period is TLL ≦ TL ≦ TLH, and the analysis range of the pitch period (TW) given to the pitch filter is TWL ≦ TW ≦ TW.
When H is set, at least one of TLL> TWL and TLH <TWH is satisfied.

The perceptual weighting filter improves the subjective quality by making the quantization noise difficult to hear by utilizing the masking effect. The masking effect referred to here is a phenomenon in which the noise is masked by the spectrum of the input voice and becomes hard to hear in the frequency region where the power spectrum of the input voice is large, even if the quantization noise is large. On the contrary, the masking effect does not work in the frequency region where the power spectrum of the input signal is small, and the quantization noise is easily heard. The perceptual weighting filter has a function of shaping the spectrum of the quantization noise so as to be close to the spectrum of the input voice, corresponds to the LPC synthesis filter corresponding to the spectrum envelope of the voice, and corresponds to the fine structure of the spectrum of the voice. And a pitch filter having a function of suppressing the pitch period component of.

Conventionally, the analysis range (TWL ≦ T) of the pitch period (TW) given to the pitch filter in the perceptual weighting filter.
W ≦ TWH) is the pitch period (TL) according to the adaptive codebook
Was the same as the search range (TLL ≦ TL ≦ TLH).
That is, TWL = TLL and TWH = TLH. This is based on the assumption that a reasonable pitch period search range by the adaptive codebook has already been selected, and the pitch period analysis range given to the pitch filter must match the pitch period search range of the adaptive codebook. There is nothing but because there is.

By the way, since the perceptual weighting filter is used as a distortion measure for the codebook search in the speech coding apparatus, it is not necessary to send information indicating the structure of the perceptual weighting filter to the speech decoding apparatus. Therefore, unlike the pitch period search range of the adaptive codebook which is limited by the number of bits of encoded data, the pitch period analysis range given to the pitch filter in the perceptual weighting filter can be originally set freely. Focusing on this point, in the present invention, the analysis range of the pitch period given to the pitch filter in the perceptual weighting filter is set sufficiently wider than the pitch period search range of the adaptive codebook.

By doing so, even if an input speech signal having a pitch period that cannot be represented by the pitch period search range of the adaptive codebook is given, the pitch period given to the pitch filter can be accurately obtained. Therefore, based on this pitch period, a pitch filter suppresses the pitch period component of the input voice signal, and a perceptual weighting filter including this pitch filter performs spectrum shaping of the quantization noise to improve the voice quality by the masking effect. Can be planned. Further, this process does not change the connectivity between the speech coding apparatus and the speech decoding apparatus, so that the quality improvement can be realized while maintaining the compatibility.

On the other hand, the speech decoding method according to the present invention analyzes the pitch period of the decoded speech signal obtained by decoding the coded data, and analyzes the pitch period of the decoded speech signal so as to enhance the pitch period component. In the speech decoding method of outputting through a filter, the analysis range of the pitch period given to the pitch filter is set wider than the range of the pitch period that can be represented by the encoded data. More specifically, the range of the pitch period (TL) that can be represented by the encoded data is TLL ≦ TL ≦ TLH,
When the analysis range of the pitch period (TP) given to the pitch filter is TPL ≦ TP ≦ TPH, at least one of TLL> TPL and TLH <TPH is satisfied.

The post filter is a filter which enhances the subjective quality by emphasizing the peak portion and attenuating the valley portion of the spectrum of the decoded speech signal obtained by the speech decoding apparatus. As a component of this post filter,
There is a pitch filter that emphasizes the pitch period component of the decoded speech signal.

Conventionally, as the information of the pitch period (TP) used for the pitch filter in the post filter, the pitch period obtained by decoding the encoded data is used as it is, or otherwise, the pitch period search range of the adaptive codebook is used. Pitch period (TP) analysis range {TPL≤TP≤
TPH} is the same as the search range {TLL ≦ TL ≦ TLH} of the pitch period (TL) of the adaptive codebook. That is, TPL = TLL and TPH = TLH. This is based on the assumption that a valid range has already been selected as the search range for the pitch period by the adaptive codebook.
This is because the analysis range of the pitch period given to the pitch filter matches the pitch period search range of the adaptive codebook.

By the way, since the post filter processes the decoded speech signal, the analysis range of the pitch period given to the pitch filter in the post filter is limited by the number of bits of the encoded data. Unlike the pitch period search range of, it can be originally set freely. Focusing on this point, in the present invention, the analysis range of the pitch period given to the pitch filter in the post filter is set sufficiently wider than the range of pitch frequencies that can be represented by the encoded data, that is, the pitch period search range of the adaptive codebook.

In this way, even if a decoded voice signal having a pitch period that cannot be represented by the pitch period search range of the adaptive codebook is given to the post filter, if the pitch period of the decoded voice signal is obtained, that is obtained. It is possible to improve the quality by emphasizing the pitch period component based on the pitch period and restoring the pitch period component that could not be transmitted.

[0017]

BEST MODE FOR CARRYING OUT THE INVENTION First, an embodiment of a speech coding method according to the present invention will be described. FIG. 1 is a block diagram for explaining the basic operation of a perceptual weighting filter used in a speech coding method according to an embodiment of the present invention. In the figure, a digitized audio signal (input audio signal) is sequentially input from the input terminal 101 frame by frame with a plurality of samples as one frame. In this embodiment, one frame has 80 samples. This input audio signal is L
It is given to the PC coefficient analysis unit 102, the pitch information analysis unit 103 and the perceptual weighting filter 104.

The LPC coefficient analysis unit 102 analyzes the input speech signal on a frame-by-frame basis using an existing technique such as an autocorrelation method, and LPC coefficients {α (i); i = 1 to N
P}. In this LPC analysis, it is necessary to use data having a length sufficient to obtain a stable analysis result centering on the frame to be analyzed of the input audio signal. NP
Represents the analysis order, and here NP = 10. LPC coefficient thus obtained {α (i); i = 1 to NP}
Is the pitch information analysis unit 103 and the perceptual weighting filter 1.
04.

The pitch information analysis unit 103 analyzes the input voice signal frame by frame, for example, as described later, to obtain the pitch period TW and the pitch filter coefficient g. Details of the pitch information analysis unit 103 will be described later with reference to FIG.

The perceptual weighting filter 104 is a filter for shaping the spectrum of the quantization noise so as to be close to the spectrum of the input speech signal, and corresponds to the LPC synthesis filter corresponding to the speech spectral envelope and the spectral fine structure of speech. And a pitch filter having a function of suppressing the pitch period component of the input audio signal. Specifically, the perceptual weighting filter 104 includes the LPC coefficient analysis unit 1
02 LPC coefficient {α (i); i = 1 to N
P} and the pitch period TW and the pitch filter coefficient g obtained from the pitch information analysis unit 103, (1)
A filter having a transfer function W (z) defined by the formula is configured to filter the input audio signal input in frame units, and output the weighted input audio signal to the output terminal 105.

[0021]

[Equation 1]

A (z / β) / A (z / γ) corresponds to a perceptual weighting filter corresponding to the spectral envelope of the voice, and Q
(Z) corresponds to the perceptual weighting filter corresponding to the fine spectral structure of the voice. Here, as specific numerical values of each parameter, the present inventors have β = 0.9, γ = 0.4, λ
= 0.4 is recommended. However, since the values of these parameters depend on subjective preferences, these numbers are not always optimal. The weighted input voice signal obtained by passing the input voice signal through the perceptual weighting filter 104 of the transfer function W (z) defined by the equation (1) is output terminal 105.
Output.

Next, referring to FIG. 2, the pitch information analysis unit 1
03 will be described. In FIG. 2, the input terminal 301
Input audio signal from the input terminal 302 and LP from the input terminal 302.
C coefficients {α (i); i = 1 to NP} are respectively input and given to the prediction residual signal calculation unit 303. Similar to the LPC coefficient analysis unit 102, the prediction residual signal calculation unit 303 performs analysis using data having a length sufficient to obtain a stable analysis result centering on the analysis target frame of the input speech signal. In the prediction residual signal calculation unit 303, the data of the input speech signal used for analysis is {u (n); n = 0 to NU−
1}, the LPC coefficient {α (i); i = 1 to NP}
Using the prediction residual signal {e
(N); n = 0 to NU-1} is calculated according to the following equation.

[0024]

[Equation 2]

The prediction residual signal {e (n); n = 0 to NU-1} obtained in this way is provided to the pitch period analysis unit 304. In the pitch period analysis unit 304, based on the signal {ew (n); n = 0 to NU-1} obtained by multiplying the prediction residual signal {e (n); n = 0 to NU-1} by a Hamming window, The autocorrelation value φ (t) defined by the equation (6) is calculated in the pitch period analysis range {TWL ≦ t ≦ TWH}.

[0026]

(Equation 3)

In this embodiment, the lower limit TWL and the upper limit TWH of the pitch period analysis range are set to TWL = 10, for example.
And TWH = 200. On the other hand, a pitch period search range {TLL ≦ TL ≦ TL of a pitch period coding means (for example, an adaptive codebook described later) not shown in FIG. 1 is used.
The lower limit TLL and the upper limit TLH of H} are, for example, TLL =
20 and TLH = 147. That is, TLL> TW
L, TLH <TWH, and is characterized in that the pitch period analysis range is wider than the pitch period search range.

The autocorrelation value φ (t) thus obtained
Is provided to the pitch filter coefficient analysis unit 305 as the pitch cycle TW. In the pitch filter coefficient analysis unit 305, the prediction residual signal {e (n); n = 0 to NU-1} obtained by the prediction residual signal calculation unit 303.
And the pitch period TW obtained by the pitch period analysis unit 304
Using, the pitch filter coefficient g is calculated according to the following equation.

[0029]

(Equation 4)

The pitch period TW and the pitch filter coefficient g thus obtained are output from the output terminal 306. Although the case of using the first-order pitch filter has been described here, it may be realized by using a higher-order pitch filter. In that case, more accurate pitch information can be obtained although the amount of calculation increases somewhat. Further, the pitch period analysis method and the pitch filter coefficient analysis method are not limited to the methods described above, and other techniques may be used.

The above processing is summarized as a flow chart shown in FIG. First of all, step S
11, the LPC coefficient {α (i); i = 1 to NP} is obtained,
Next, in step S12, the prediction residual signal {e (n); n = 0
~ NU-1} is obtained. In step S13, pitch cycle T
W is analyzed, and then the pitch filter coefficient g in the pitch period TW is obtained in step S14. In step S15, steps S11, S13 and S
14, LPC coefficient {α (i); i = 1 to NP},
A pitch weight TW and a pitch filter coefficient g are used to form a perceptual weighting filter defined by the equation (1), and in step S16, the input voice signal is passed through the perceptual weighting filter to generate and output a weighted input voice signal. .

Next, a CELP-based speech encoding apparatus using the above-described perceptual weighting filter will be described with reference to FIG. 4, the same components as those of FIG. 1 are designated by the same reference numerals, and detailed description thereof will be omitted.

First, the LPC coefficients {α (i); i = 1 to NP} output from the LPC coefficient analysis unit 102 are given to the LPC coefficient quantization unit 106 and quantized. In the weighted synthesis filter 107, the LPC coefficient analysis unit 102 outputs information on the LPC coefficients {α (i); i = 1 to NP}, the pitch information analysis unit 103 outputs information on the pitch period TW and the pitch filter coefficient g, and the LPC coefficient quantum. By receiving the quantized information of the LPC coefficients {αq (i); i = 1 to NP} from the conversion unit 106, the filter of the transfer function Hw (z) is configured. This weighted synthesis filter 107
The transfer function Hw (z) of is expressed by the following equation.

Hw (z) = W (z) · H (z) (8) Here, the transfer function W (z) of the perceptual weighting filter 104.
Is the same as the equation (1) shown above. The synthesis filter H (z) is expressed by the following equation.

[0035]

(Equation 5)

The drive signal given to the weighted synthesis filter 107 is expressed by a combination of candidates of adaptive codebook 108, adaptive vector gain codebook 113, noise codebook 109, and noise vector gain codebook 111.

The adaptive codebook 108 always holds the immediately preceding drive signal sequence, repeats this drive signal sequence at a desired pitch period to generate an adaptive vector, and efficiently expresses the periodicity. However, since this pitch period must be transmitted through the multiplexer, the pitch period is searched only within the range of the number of candidates that can be expressed by the predetermined number of bits. In this embodiment, the adaptive codebook 108
The pitch period search range {TLL ≦ TL ≦ TLH} by TLL = 20 and TLH = 147 will be described.

The noise codebook 109 is a codebook having a noise sequence as a candidate vector. In general, the random codebook 10
Structure 9 is structured for the purpose of reducing the amount of calculation and improving quality, but since it is not particularly related to the gist of the present invention, its explanation is omitted.

Adaptive vector and adaptive vector gain are selected from adaptive codebook 108 and adaptive vector gain codebook 113, and these are multiplied by multiplier 110. Similarly, the adaptive vector and the noise vector gain are selected from the noise codebook 109 and the noise vector gain codebook 114, respectively, and are multiplied by the multiplier 111. Then, in the adder 112, the multipliers 110, 1
A drive signal is generated by adding the output vectors of 11 to each other, and is supplied to the weighted synthesis filter 107 as an input.

With the output signal of the perceptual weighting filter 104 as the target signal, the error of the output signal of the weighted synthesis filter 107 is taken by the subtractor 115, and the minimum distortion searching unit 11 is further used.
The square distortion is calculated at 6. The minimum distortion search unit 116, for the adaptive codebook 108, the adaptive vector gain codebook 113, the noise codebook 109, and the noise vector gain codebook 111, the adaptive vector, the adaptive vector gain, and the noise vector that minimize the square distortion. And a vector of noise vector gains are efficiently searched for, and index information of each candidate of an adaptive vector, an adaptive vector gain, a noise vector, and a noise vector gain when the square distortion is minimized is given to the multiplexer 117.

On the other hand, the LPC coefficient quantization unit 106
Index information obtained as a result of quantizing the coefficient is given to the multiplexer 117. Multiplexer 117
Is an LPC coefficient quantization unit 106 and a minimum distortion search unit 11
The index information input from 6 is converted into a bit stream as encoded data and output to the output terminal 118. Finally, the drive signal when the squared distortion calculated by the minimum distortion search unit 116 is minimized is given to the adaptive codebook 108 to update the internal state, and prepares for the input speech signal of the next frame.

As described above, in this embodiment, the pitch period analysis range {T of the pitch information analysis unit 103 used in the perceptual weighting filter 104 and the weighted synthesis filter 107 is used.
WL ≦ TW ≦ TWH} and the weighted synthesis filter 107
Represents the periodicity of the drive signal given to the multiplexer 11
Of the pitch cycle search range {TLL ≦ TL ≦ TLH} of the adaptive codebook 108 represented by the coded data of the pitch cycle (encoded data of the adaptive vector index) output from the output terminal 118. Relationship is TWL
<TLL and TWH> The condition of TLH is satisfied.
That is, the pitch period analysis range {TWL ≦ TW ≦ TW
H} is set wider than the pitch period search range {TLL ≦ TL ≦ TLH}.

By satisfying such a condition, the pitch period search range {TLL≤TL≤T of the adaptive codebook 108 which must be expressed by a predetermined number of bits.
Even if an input speech signal having a pitch period deviating from LH} is given, the pitch filter analysis range {TW in the perceptual weighting filter 104 and the weighting synthesis filter 107 is given.
Since L ≤ TW ≤ TWH} is wider than the pitch period search range of the adaptive codebook 108, it is possible to perform spectrum shaping of the quantization noise in the pitch period of the input speech signal and reduce the noise feeling by the masking effect. As a result, the subjective quality can be effectively improved.

In the present embodiment, the pitch cycle analysis range {TWL≤TW≤TWH} is set to the pitch cycle search range {TL.
L ≦ TL ≦ TLH} has a relationship of TLL> TWL and T
The condition of LH <TWH is satisfied, but TLL> TWL
Alternatively, only one of the conditions of TLH <TWH and TLH <TWH may be satisfied.

Next, an embodiment of the speech decoding method according to the present invention will be described. FIG. 5 is a block diagram for explaining the basic operation of the post filter used in the speech decoding method according to the embodiment of the present invention. In the figure, a digitized audio signal (for example, a decoded audio signal) is sequentially input from the input terminal 201 frame by frame with a plurality of samples as one frame. In this embodiment, one frame is described as 80 samples.

On the other hand, from the input terminal 202, the input terminal 2
The LPC prediction residual signal of the audio signal from 01 or a signal similar thereto, for example, a drive signal for driving a synthesis filter in a CELP-type audio decoding device described later is input. The pitch information analysis unit 203 obtains the pitch period using the LPC prediction residual signal or the drive signal of the synthesis filter. Details of the pitch information analysis unit 203 will be described later.

The post filter 205 has an input terminal 20.
For example, the decoded speech signal is given from 1, the information of the pitch period TP and the pitch filter coefficient g is given from the pitch information analysis unit 203, and the information of the LPC coefficient {α (i); i = 1 to NP} from the input terminal 204. Is given. LP
The C coefficient represents the spectrum envelope of the audio signal from the input terminal 201. The post filter 205 uses the pitch period TP and the LPC coefficient {α (i); i = 1 to N.
The filter represented by the transfer function R (z) of the following equation is configured using the information of P}, and the audio signal from the input terminal 201 is filtered. The filtered output signal is output from the output terminal 206. R (z) = F (z) * P (z) * U (z) (10) F (z), P (z), and B (z) are represented as follows.

[0048]

(Equation 6)

Here, as concrete numerical values of each parameter, the present inventors have ν = 0.5, ξ = 0.8, η = 0.
7, μ = 0.4 is recommended. The values of these parameters are
These numbers are not always optimal as they depend on subjective preferences.

Next, the pitch information analysis unit 203 in this embodiment will be described with reference to FIG. 6, the same elements as those of FIG. 2 are designated by the same reference numerals, and the description thereof will be omitted.

The pitch information analysis unit 203 shown in FIG.
The difference between the pitch information analysis unit 103 and the pitch information analysis unit 103 shown in FIG. 2 in the above embodiment is the input signal. That is, the pitch information analysis unit 203 in FIG. 6 receives the prediction residual signal or a signal corresponding thereto, for example, a drive signal generated by a speech decoding device (not shown). Therefore, it is not necessary to input the input voice signal and the LPC coefficient to the pitch information analysis unit 203 unlike the pitch information analysis unit 103 in FIG.
3 is not necessary either. Pitch information analysis unit 203 in FIG.
Is the pitch period T obtained by the pitch period analysis unit 304.
Information on the pitch filter coefficient g obtained by P and the pitch filter coefficient analysis unit 305 is output from the output terminal 308.

Here, the lower limit TPL and the upper limit TP of the analysis range {TPL≤TP≤TPH} of the pitch period TP in the pitch period analysis unit 304 in the pitch information analysis unit 203.
H is set to TPL = 10 and TPH = 200, for example. On the other hand, the lower limit TLL and the upper limit TLH of the pitch period search range {TLL ≦ TL ≦ TLH} of the pitch period encoding means (for example, adaptive codebook) are TLL = 20 and TLH.
= 147. That is, TLL> TPL, TPH>
TLH is characterized in that the pitch period analysis range is wider than the pitch period search range.

The above processing is summarized as a flow chart shown in FIG. First of all, step S
In step 21, the pitch period TP is analyzed, and in step S22, the pitch filter coefficient g in the pitch period TP is obtained. In step S23, steps S21 and S
A pitch filter TP and pitch filter coefficient g obtained in step 22 and the LPC coefficient input from the input terminal 204 are used to form a post filter defined by equation (10),
In step S24, the audio signal input from the input terminal 201 is passed through the post filter and output.

Next, the CELP speech decoding apparatus using the above-described post filter will be described with reference to FIG. 8, the same elements as those of FIG. 5 are designated by the same reference numerals, and detailed description thereof will be omitted.

In FIG. 8, a bit stream, which is encoded data output from a CELP audio encoding apparatus (not shown), is input to an input terminal 401 via a transmission line or a storage medium (not shown). The voice encoding device is configured, for example, as shown in FIG. The demultiplexer 402 decodes the parameters required to generate an audio signal from the input bitstream.
The type and number of these parameters vary depending on the configuration of the speech coder, but here, LPC coefficient index, adaptive vector index, adaptive vector gain index, noise vector index and noise vector gain index are used as parameters. It shall be decrypted.

From the adaptive codebook 403 and the adaptive vector gain codebook 404, the adaptive vector and the adaptive vector gain designated by the adaptive vector index and the adaptive vector gain index are selected, and these are multiplied by the multiplier 405. Similarly, the noise vector and the noise vector gain designated by the noise vector index and the noise vector gain index are selected from the noise codebook 406 and the noise vector gain codebook 407, respectively.
It is multiplied by 8.

Then, the adder 409 causes the multiplier 40 to
5, 408, the drive signals are generated by summing the vectors output, and the drive signals are provided to the synthesis filter 411 and the pitch information analysis unit 203. This drive signal is also given to the adaptive codebook 403 to update the internal state and prepare for the next input.

On the other hand, when the LPC coefficient index is given to the LPC coefficient decoding section 410, the LPC coefficient {αq (i); i = 1 to NP} is decoded, and this LPC coefficient is decoded.
The coefficients are the synthesis filter 411 and the post filter 205.
Given to. The transfer function of the synthesis filter 411 is (9)
It is expressed in the same manner as the equation, and the driving signal from the adder 409 is input to the synthesis filter 411 to perform filtering and obtain a decoded speech signal. This decoded audio signal is input to the post filter 205.

The post filter 205 and the pitch information analysis unit 203 are as described with reference to FIGS. 5 to 7, so detailed description will be omitted here. In the present embodiment, the decoded audio signal output from the synthesis filter 411 is input to the post filter 205, and the drive signal output from the adder 409 is input to the pitch information analysis unit 203. In the speech decoding apparatus of this embodiment, the post filter 205
The decoded voice signal that has passed through is finally output from the output terminal 206.

As described above, in this embodiment, the pitch period analysis range {TPL≤T in the pitch information analysis unit 203 that analyzes the pitch information used in the post filter 205.
P ≦ TPH} and the periodicity of the drive signal given to the synthesis filter 411, which can be the pitch period (TL) specified by the adaptive vector index decoded by the demultiplexer 402 and used in the adaptive codebook 403. The relation of the range {TLL ≦ TL ≦ TLH} is TPL <T
The conditions of LL and TPH> TLH are satisfied. That is, the pitch cycle analysis range {TPL≤TP≤TPH} is set wider than the pitch cycle range {TLL≤TL≤TLH} that can be expressed by pitch cycle encoded data (adaptive vector index encoded data). .

By satisfying such a condition, the pitch period search range {TLL≤TL≤T of the adaptive codebook 403 which has to be expressed by a predetermined number of bits.
Even if a decoded speech signal having a pitch period deviating from LH} is input to the post filter 205, the pitch analysis range {TPL ≦ TP ≦ TPH} of the pitch information analysis unit 203 used in the post filter 205 is determined by the adaptive codebook 40.
Since it is wider than the pitch period search range of 3, the pitch period that could not be transmitted as the encoded data of the adaptive vector index can be restored. As a result, the subjective quality can be improved.

In this embodiment, the relationship between the pitch period analysis range {TPL≤TP≤TPH} and the pitch period range {TLL≤TL≤TLH} that can be expressed by the encoded data is TPL <TLL and TPH> TLH. Although the condition is satisfied, only one of TLL> TPL and TLH <TPH may be satisfied.

Next, another embodiment of the present invention will be described.
FIG. 9 is a block diagram for explaining the basic operation of the post filter used in the speech coding method according to another embodiment of the present invention. In the figure, the same components as those of FIG. 5 are designated by the same reference numerals, and detailed description thereof will be omitted.

The difference between this embodiment and the embodiment shown in FIG. 5 is that a speech decoding apparatus (not shown) has an adaptive codebook and a fixed codebook made up of fixed candidate vectors prepared in advance, and The difference is that the pitch period TP is obtained differently when the adaptive codebook is selected and when the fixed codebook is selected.

When the adaptive codebook is selected, the pitch period TL of the transmitted and decoded adaptive codebook is regarded as the pitch period TP given to the pitch filter in the post filter,
A pitch filter coefficient g is obtained using this pitch period TP and is given to the post filter 205. On the contrary, when the fixed codebook is selected, the pitch information analysis unit 203 newly obtains the pitch period TP, calculates the pitch filter coefficient g using this pitch period TP, and gives it to the post filter 205.

Next, the pitch information analysis unit 203 in this embodiment will be described with reference to FIG. 10, the same elements as those of FIG. 6 are designated by the same reference numerals and the description thereof will be omitted.

In FIG. 10, the input terminal 211 supplies selection information indicating which of the adaptive codebook and the fixed codebook is used in the speech decoding apparatus (not shown). When this selection information represents the adaptive codebook, switch 3
Reference numeral 10 regards the information of the pitch cycle TL of the adaptive codebook input from the input terminal 207 as the information of the pitch cycle TP used in the post filter, and the pitch filter coefficient analysis unit 3
Give to 05. On the other hand, when the selection information from the input terminal 211 represents the fixed codebook, the switch 310 switches the input terminal 20.
It works so that the input from 2 becomes effective. That is, the driving signal sequence which is a prediction residual signal or a signal corresponding thereto is input from the input terminal 202, the pitch period analysis unit 304 determines the pitch period TP based on this signal, and the pitch period TP is given to the pitch filter coefficient analysis unit 305. . At this time, it can be considered that the fixed codebook is selected because a pitch that cannot be represented by the pitch period search range {TLL ≦ TL ≦ TLH} of the adaptive codebook has occurred, so that the pitch period analysis unit 305 selects the fixed codebook. The analysis range excludes the pitch period search range of the adaptive codebook {TPL ≦ TP <TLL, TLH <TP ≦ T
PH}, which can reduce the amount of calculation required for the pitch period analysis.

The pitch filter coefficient analysis unit 305 calculates the pitch filter coefficient g by using the prediction residual signal or a drive signal sequence based on the information of the given pitch cycle TP, and calculates the pitch cycle TP and the pitch filter coefficient. The information of g is output from the output terminal 308.

The above processing is summarized as a flow chart shown in FIG. Step S3 of FIG.
The processes of 3, S34, S35, and S36 are the same as steps S21, S22, S23, and S24 of FIG. 7, and thus the description thereof is omitted here. However, step S33
The difference between the pitch cycle analysis range in step S21 and the pitch cycle analysis range in step S21 is as described above.

First, in step S31, it is determined whether the selection information represents the adaptive codebook or the fixed codebook. When the selection information represents the adaptive codebook, the process proceeds to step S32, and when the selection information represents the fixed codebook, the process proceeds to step S33. When the selection information represents the adaptive codebook, the pitch period TL obtained by the adaptive codebook search in step S32
Is set as a pitch period TP used in the pitch filter in the post filter, and the process proceeds to step S34. When the selection information represents the fixed codebook, the pitch period TP is newly calculated in step S33, and the process proceeds to step S34.

Next, the CELP speech decoding apparatus using the above-described post filter will be described with reference to FIG. 12, the same components as those of FIG. 8 are designated by the same reference numerals, and detailed description thereof will be omitted.

The present embodiment is different from the embodiment of FIG. 8 in that a fixed codebook 412 is provided in the same column as the adaptive codebook 403. Hereinafter, the description will be given with an emphasis on the difference from the embodiment of FIG.

In FIG. 12, the demultiplexer 402
The adaptive vector index output from
3 is applied to the adaptive codebook 4 to be decoded.
Is a vector generated from the fixed codebook 412
Is determined from the vector. The determination result is given to the switches 414, 415 and the pitch information analysis unit 203. Here, the adaptive vector indexes are adaptive codebook 403 and fixed codebook 412.
Although the case where both vectors are represented in the same manner has been described, the determination information may be directly generated from the demultiplexer, and in that case, the determination unit 413 is not necessary. In this case, the speech coding apparatus (not shown) has a configuration in which the multiplexer is newly provided with decision information as information to be transmitted, and the decision information is 1 bit to indicate whether it is an adaptive codebook or a fixed codebook. Additional information is required.

The switch 414 switches whether the adaptive vector index is given to the adaptive codebook 403 or the fixed codebook 412 based on the decision information given from the decision unit 413. Similarly, the switch 415 serves as the determination unit 41.
A vector to be given to the multiplier 405 is determined based on the judgment information given from No. 3.

In the pitch information analysis unit 203, the judgment unit 41
Based on the determination information given from No. 3, the method of calculating the pitch period TP of the pitch filter used in the post filter 205 is switched as described in FIGS. 10 and 11,
The pitch period TP and the pitch filter coefficient g calculated by the pitch information analysis unit 203 are given to the post filter 205.

Next, the effect of this embodiment will be described. Adaptive codebook 403 generates an adaptive vector that can efficiently represent the pitch period using the immediately preceding drive signal sequence, whereas fixed codebook 412 prepares a plurality of predetermined fixed vectors. The pitch cycle of the input speech signal given to the speech coding apparatus (not shown) is the pitch cycle search range of adaptive codebook 403 {TLL≤TL≤TLH}.
, The adaptive vector of the adaptive codebook 403 is selected and its index is encoded.

However, the input voice signal is the adaptive codebook 403.
If the pitch period is not included in the pitch period search range of, the fixed codebook 412 is used instead of the adaptive codebook 403. This is because the adaptive codebook 403 was used,
This means that it is possible to determine whether the pitch cycle of the input speech signal is included in the pitch cycle search range of the adaptive codebook 403 depending on whether the fixed codebook 412 is used.

Furthermore, when fixed codebook 412 is used, the pitch period analysis range in pitch information analysis section 203 is the pitch period search range of adaptive codebook 403 {TLL≤TL.
Since it can be determined that ≦ TLH} is not included, the pitch period analysis range is set to {TPL ≦ TP <TLL, TLH <TP ≦.
TPH} and the amount of calculation can be reduced. On the other hand, when the adaptive codebook 403 is selected, it is considered that the pitch period of the input speech signal is represented by the pitch period TL of the adaptive codebook 403, and the pitch filter in the post-filter 205 is based on this pitch period TL. Pitch emphasis should be performed.

Although the above-described embodiment has described the example in which the present invention is applied to the speech encoding / decoding method of the CELP system, the present invention is not limited to the APC (Adaptive Predictive Codin).
It can also be applied to voice encoding / decoding methods of other methods such as g) method.

[0080]

As described above, according to the present invention, it is possible to provide a voice coding method and a voice decoding method which can accurately express the pitch period of a voice signal to obtain high quality voice.

That is, in the speech coding method according to the present invention, the pitch period of the adaptive codebook is set by setting the analysis range of the pitch period given to the pitch filter in the perceptual weight filter wider than the pitch period search range of the adaptive codebook. Even if an input speech signal having a pitch period that cannot be represented by the search range is given, the pitch period given to the pitch filter can be accurately obtained. Therefore, by suppressing the pitch period component of the input speech signal with the pitch filter based on this pitch period, and performing the spectral shaping of the quantization noise by the perceptual weighting filter including this pitch filter,
The quality of voice can be improved by the masking effect. Further, this process does not change the connectivity between the speech coding apparatus and the speech decoding apparatus, so that the quality improvement can be realized while maintaining the compatibility.

Further, in the voice decoding method according to the present invention,
By setting the analysis range of the pitch period given to the pitch filter in the post filter wider than the range of the pitch frequency that can be represented by the encoded data, the decoded audio signal with the pitch period that cannot be represented by the encoded data is transmitted to the post filter. Even if given, if the pitch period of the decoded speech signal is obtained, the pitch period component can be emphasized based on the pitch period and the pitch period component that could not be transmitted can be restored to improve the quality. .

[Brief description of drawings]

FIG. 1 is a block diagram for explaining a basic operation of a perceptual weighting filter used in a speech coding method according to an embodiment of the present invention.

FIG. 2 is a block diagram showing a configuration of a pitch information analysis unit in the same embodiment.

FIG. 3 is a flowchart showing a processing procedure of the embodiment.

FIG. 4 is a C to which the speech coding method according to the embodiment is applied.
Block diagram showing the configuration of a speech synthesizer by ELP method

FIG. 5 is a block diagram for explaining a basic operation of a post filter used in the speech decoding method according to the embodiment of the present invention.

FIG. 6 is a block diagram showing a configuration of a pitch information analysis unit in the same embodiment.

FIG. 7 is a flowchart showing a processing procedure according to the embodiment;

FIG. 8 is a C to which the speech decoding method according to the embodiment is applied.
Block diagram showing the configuration of a voice decoding device by the ELP method

FIG. 9 is a block diagram illustrating a basic operation of a post filter used in the speech decoding method according to the embodiment.

FIG. 10 is a block diagram showing a configuration of a pitch information analysis unit in the same embodiment.

FIG. 11 is a flowchart showing a processing procedure of the embodiment.

FIG. 12 is a block diagram showing the configuration of a CELP speech decoding apparatus to which a speech decoding method according to another embodiment of the present invention is applied.

[Explanation of symbols]

 101 ... Voice signal input terminal 102 ... LPC coefficient analysis section 103 ... Pitch information analysis section 104 ... Auditory weighting filter 105 ... Output terminal 106 ... LPC coefficient quantization section 107 ... Weighted synthesis filter 108 ... Adaptive codebook 109 ... Noise codebook 110, 111 ... Multiplier 112 ... Adder 113 ... Adaptive Vector Gain Codebook 114 ... Noise Vector Gain Codebook 115 ... Subtractor 116 ... Minimum Distortion Search Unit 117 ... Multiplexer 118 ... Output Terminal 201 ... Decoded Speech Signal Input Terminal 202 ... Drive signal input terminal 203 ... Pitch information analysis unit 204 ... LPC coefficient input terminal 205 ... Post filter 206 ... Post filter output terminal 207 ... Adaptive codebook pitch period input terminal 211 ... Codebook selection information input terminal 301 ... Voice signal input terminal 302 ... LPC coefficient input terminal 3 3 ... Prediction residual signal calculation unit 304 ... Pitch cycle analysis unit 305 ... Pitch filter coefficient input terminal 306 ... Output terminal 308 ... Output terminal 310 ... Switch 401 ... Bit stream input terminal 402 ... Demultiplexer 403 ... Adaptive codebook 404 ... Adaptation Vector gain codebook 405 ... Multiplier 406 ... Noise codebook 407 ... Noise vector gain codebook 408 ... Multiplier 409 ... Adder 410 ... LPC coefficient decoding unit 411 ... Synthesis filter 412 ... Fixed codebook 413 ... Judgment unit 414, 415 …switch

Claims (4)

[Claims] 1. A pitch cycle of an input speech signal is analyzed, the pitch cycle component of the input speech signal is applied to a pitch filter, and the input is made based on the speech signal passed through a perceptual weighting filter including the pitch filter. In a voice encoding method for searching and encoding a pitch period of a voice signal, a wide range of analysis of the pitch period given to the pitch filter is set with respect to the range of the pitch period that can be represented by the encoded data of the pitch period. Characteristic speech coding method. 2. A pitch cycle of an input speech signal is analyzed, the pitch cycle component of the input speech signal is given to a pitch filter which emphasizes the pitch cycle component, and the input is based on the speech signal passed through a perceptual weighting filter including this pitch filter. A voice coding method for searching for and coding a pitch period of a voice signal, comprising a pitch period (TL) that can be represented by the coded data.
Is set to TLL ≦ TL ≦ TLH, and the analysis range of the pitch period (TW) given to the pitch filter is TWL ≦ T.
When W ≦ TWH, TLL> TWL and TLH <
A speech encoding method characterized by satisfying at least one condition of TWH. 3. Speech decoding for analyzing a pitch period of a decoded speech signal obtained by decoding coded data and outputting the decoded speech signal through a post filter including a pitch filter for emphasizing a pitch period component. The speech decoding method, wherein the analysis range of the pitch period given to the pitch filter is set wider than the range of the pitch period that can be represented by the encoded data. 4. A speech decoding method in which a pitch period of a decoded speech signal obtained by decoding encoded data is analyzed, and the decoded speech signal is output through a post filter including a pitch filter that emphasizes a pitch period component. Method, a pitch period (TL) that can be represented by the encoded data.
Is set to TLL ≦ TL ≦ TLH, and the analysis range of the pitch period (TP) given to the pitch filter is TPL ≦ T.
When P ≦ TPH, TLL> TPL and TLH <
A speech decoding method characterized by satisfying at least one condition of TPH.