JPH113098A - Method and device of encoding speech - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method of encoding speech permitting to calculate a pitch cycle of a speech signal by a less calculation amount and express it by a less information amount. SOLUTION: An input speech signal from an input terminal 21 is divided into frames by a frame sub-frame composition part 22 and each frame is further divided into sub-frames, and then, pitch cycles of at least two of a present frame, a past frame to the present frame, and a future frame are determined by a pitch cycle analysis part 23. From these pitch cycles, pitch cycles of each sub-frame are determined by interpolating by a sub-frame pitch cycle extraction part 24. And, this sub-frame pitch cycle information is outputted from an output terminal 25.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech encoding method for compressing and encoding a speech signal, and more particularly to a process for obtaining a pitch period in a speech signal encoding process.

[0002]

2. Description of the Related Art A technique for compressing and encoding an audio signal at a low bit rate with high efficiency is an important technique for effective use of radio waves and reduction of communication costs in mobile communication such as a car telephone or in a company communication. .

[0003] CELP is a speech encoding system capable of performing speech synthesis of excellent quality at a bit rate of 8 kbps or less.
(Code Excited Linear Prediction) method is known. The CELP method is described in “Code-Excited Linear Prediction (CELP) High-Level” by MR Schrodeder and BSAtal.
Quality Speech at Very low Bit Rates ”, Proc.ICASS
P; 1985, pp. 937-939 (Reference 1), it has attracted attention as a method of synthesizing high-quality speech, and various studies have been made on improving quality and reducing the amount of calculation.

[0004] As one of the components for realizing speech coding by the CELP method, there is an adaptive codebook. The adaptive codebook performs pitch prediction analysis of an input speech signal by closed loop operation or analysis by synthesis (Analysis by Synthesis). In general, in pitch prediction analysis using an adaptive codebook, a search for a pitch period is performed in a search range (128 candidates) of 20 to 147 samples to find a pitch period that minimizes distortion with respect to a target signal. It is often transmitted as coded data of bits.

In the conventional CELP system described above, the pitch period is determined by a closed loop operation in units of subframes. Therefore, when the search range of the pitch period is as large as 128 candidates as described above, the amount of calculation becomes enormous. Further, in such a direct pitch period search method, the pitch period information requires 7 bits per subframe,
If one frame is composed of four subframes, as many as 28 bits are required per frame.

[0006] Originally, the pitch cycle of the voice signal fluctuates largely in many parts, and it is not necessary to perform a full search for each subframe. Utilizing such a property of the pitch period, the amount of calculation and the number of bits can be reduced. From such a viewpoint, a method using a differential pitch expression that limits the search range of the pitch cycle has been reported.

One of them is to search all candidates for odd-numbered sub-frames and search only candidates near odd-numbered sub-frames for even-numbered sub-frames when searching for a pitch period. JPCampbell, Jr. et al. “An Expandable Erro
r-Protected 4800 bps CELP coder (USFederal Standa
rd 4800 bps Voice Coder) ”, Proc. ICASSP; 1989, pp. 73
5-738 (Reference 2). According to this method, for odd subframes, a full search of 128 candidates,
For example, when the pitch period is searched for even 32 subframes, the information amount of the pitch period can be reduced to 24 bits per frame.

However, according to this method, if a value that is different from the actual pitch period is selected for the pitch period obtained in the odd-numbered sub-frames, it affects the next sub-frame and quality degradation is perceived. Problems arise.
In fact, if pitch analysis is performed in a short section such as a subframe, a pitch cycle different from the actual one is likely to be obtained.

A method of stably obtaining the pitch cycle of a frame, expressing each sub-frame as a variation from the frame pitch cycle, and searching for a pitch cycle only in the vicinity thereof is described by M. Yong et al. In "Efficient Encoding of
the Long-Term Predictor inVector Excitation Coders
According to this method, information on the pitch period of a frame is represented by 7 bits, and information on the pitch period of a subframe is represented by 5 bits. [Boston, MA: Kluwer, 1991, pp. 329-338 (Reference 3)] Expressed in bits.

The advantage of this method is that the search for the pitch period of all the subframes needs to be performed for only 32 candidates, and the amount of calculation can be greatly reduced. However,
This method has a problem that significant deterioration occurs when the pitch period fluctuates in a frame. This problem will be described with reference to FIG.

As shown in FIG. 1A, when the actual pitch period fluctuates in a frame, the pitch period of the frame is obtained as an average period in the frame. The pitch period of each subframe is set to a search range near the pitch period of the frame. However, if the search range is not large enough, the actual pitch period cannot be obtained as shown in FIG. In addition, if the search range is enlarged, the number of bits required to represent the information of the pitch period becomes necessary, and the amount of calculation and the effect of bit reduction are reduced.

[0012]

As described above, in the CELP system, which is a conventional speech coding method, the pitch period is obtained by a closed-loop search in units of subframes, so that the amount of calculation required to obtain the pitch period is small. There is a problem that it becomes enormous and the number of bits of pitch period information increases.

In the method described in Reference 2 for obtaining the pitch period by limiting the search range, the amount of calculation for obtaining the pitch period and the number of bits of the pitch period information are reduced. There is a problem that it is difficult to obtain.

Further, as described in Reference 3, in the method of searching the pitch period of a subframe as a variation from the pitch period of the frame only in the vicinity of the pitch period of the frame, it is possible to greatly reduce the amount of calculation. However, when the pitch period fluctuates within the frame, there is a problem that the pitch period of the sub-frame cannot be correctly obtained.

The present invention has been made in order to solve such problems of the prior art, and it is possible to correctly obtain a frame period of an audio signal with a small amount of calculation and express a pitch period with a small amount of information. It is an object of the present invention to provide a speech encoding method capable of performing the above.

[0016]

SUMMARY OF THE INVENTION In order to solve the above-mentioned problems, the present invention divides an input speech signal into frames of a predetermined length and includes a process for obtaining a pitch period of the input speech signal. The process is characterized in that it includes a process of obtaining a pitch period of a future frame in time with respect to a current frame to be coded, and a process of coding this pitch period.

Also, the present invention provides a speech code including a process of dividing an input speech signal into frames of a predetermined length, further dividing the speech signal of each frame into subframes, and finding a pitch period of the speech signal. In the encoding process, a pitch period prediction value of a subframe in the current frame is obtained using a pitch period of at least two frames of a past frame and a future frame for the current frame to be encoded and the current frame. The pitch period of a subframe in the current frame is obtained using the predicted pitch period value.

As described above, according to the present invention, the pitch period of the future frame is obtained with respect to the current frame, and is used together with the pitch period of the current frame and the pitch period of the past frame to obtain the pitch period of the subframe in the current frame. The pitch period prediction value is obtained, and the pitch period of the subframe in the current frame is obtained using the pitch period prediction value. Therefore, even if the pitch period fluctuates within the frame, the pitch period of the subframe can be accurately determined. It can be obtained with a good and small amount of calculation, and can be represented with a small amount of information.

Referring to FIG. 1B, as shown in FIG. 1, the actual pitch period changes within a frame, as in FIG. 1A. In the present invention, for example, interpolation is performed between the pitch cycle of the current frame and the pitch cycle of the past frame, a subframe pitch cycle prediction value is obtained for each subframe, and a search range of the pitch cycle of the subframe is determined in the vicinity thereof. . For this reason, the actual pitch period is included in the search range, and an accurate search for the pitch period can be performed.

Further, since the predicted value of the subframe pitch period approximates the actual pitch period fairly accurately, there is no problem even if the search range of the pitch period of the subframe is narrowed to, for example, eight candidates. If the search range of the subframe pitch period is 8 candidates, the pitch period of the frame is represented by 7 bits and 3 bits per subframe. Therefore, if one frame is composed of 4 subframes, the pitch period of the subframe is In contrast to the requirement of 28 bits per frame, it can be represented by an information amount of 7 bits + 3 bits * 4 = 19 bits. Further, since the search range of the subframe pitch period is as small as eight candidates, it greatly contributes to the reduction of the calculation amount.

In the present invention, the pitch period of the sub-frame in the current frame obtained as described above may be encoded, or a case where a pitch filter for enhancing the pitch period component of the input audio signal is provided. The transfer function of the pitch filter may be determined using the pitch period of the subframe in the current frame obtained as described above. The pitch filter is known as a component of, for example, an auditory weight filter or a post filter.

Further, the present invention has an adaptive codebook storing a plurality of adaptive vectors generated by repeating past drive signal sequences at a cycle included in a predetermined range, and is extracted from the adaptive codebook. In a speech encoding process including a process of searching an adaptive vector having a cycle in which an error between a signal obtained by passing an adaptive vector through a predetermined filter and a target vector is minimized from a predetermined search range, an input audio signal is determined in advance. After the audio signal of each frame is further divided into sub-frames, the current frame to be encoded and at least two of the past and future frames with respect to the current frame are encoded. The pitch period prediction value of the subframe in the current frame is obtained using the pitch period of the current frame, and the current frame And determines the search range for the sub-frame in the arm.

In the present invention, when determining the pitch period of a frame, the analysis position of the pitch period may be determined adaptively for each frame. More specifically, the pitch period analysis position is determined by the magnitude of the short-term power of the speech signal, the prediction residual signal, or the prediction residual signal after passing through the low-pass filter. By doing so, the pitch period can be determined more accurately, and the quality of decoded speech can be improved.

Further, a method of obtaining the pitch period of the sub-frame in the current frame according to the continuity of the pitch period may be selected. For example, when it is determined that the pitch period is continuously changing, a predicted value of the sub-frame pitch period is obtained, and a sub-frame pitch period is obtained by searching the vicinity thereof. Conversely, when it is determined that the change in the pitch period is discontinuous, the subframe pitch period is obtained by a full search. By such adaptive processing, a search method for an optimal subframe pitch cycle is selected according to the continuity of the pitch cycle, so that the quality of decoded speech is improved.

Further, the apparatus further includes a relative pitch pattern codebook storing a plurality of relative pitch patterns representing variations of the pitch cycle of the plurality of subframes, and a change in the pitch cycle of the subframe is determined by a predetermined value from the relative pitch pattern codebook. It may be represented by one relative pitch pattern selected based on the index, thereby further reducing the number of bits of the information indicating the subframe pitch period.

That is, the relative pitch pattern codebook is:
For example, a relative pitch pattern having a high appearance frequency is stored as a vector. This is matched with a vectorized pitch period of a plurality of subframes, and the pitch period of the plurality of subframes is represented by the most suitable relative pitch pattern. For example, if three bits per subframe are required to individually represent the pitch period of a subframe, 12 bits are required per four subframes. This four-dimensional vector is represented by the size of 7 bits in the relative pitch pattern codebook. If represented by one relative pitch pattern having a length, this leads to a bit reduction of 5 bits per frame.

Further, in the present invention, a computer-readable program storing a program for performing a speech encoding process including a process of dividing an input speech signal into frames of a predetermined length and calculating a pitch period of the input speech signal is provided. A possible recording medium, in which a program for executing a process of obtaining a pitch period of a future frame in time with respect to a current frame to be encoded and a process of encoding the pitch period is recorded. A recording medium is provided.

According to the present invention, an audio signal includes a process of dividing an input audio signal into frames of a predetermined length, further dividing the audio signal of each frame into subframes, and obtaining a pitch period of the audio signal. A computer-readable recording medium on which a program for performing an encoding process is recorded, wherein a current frame to be encoded and a pitch of at least two frames of a past frame and a future frame with respect to the current frame. Using the period, determine the predicted pitch period of the subframe in the current frame,
A recording medium is provided that records a program for executing a process of obtaining a pitch period of a subframe in the current frame using the pitch period predicted value.

Further, in the present invention, there is provided an adaptive codebook storing a plurality of adaptive vectors generated by repeating a past drive signal sequence at a cycle included in a predetermined range, and is extracted from the adaptive codebook. A program for performing a speech encoding process including a process of searching an adaptive vector having a cycle that minimizes an error between a signal obtained by passing an adaptive vector through a predetermined filter and a target vector from a predetermined search range is recorded. A computer-readable recording medium, which divides an input audio signal into frames of a predetermined length, further divides the audio signal of each frame into subframes, and then encodes the current frame and the current frame to be encoded. Using the pitch periods of at least two of the past and future frames in the current frame. Determine the pitch period prediction value of the sub-frame, a recording medium recording a program for executing processing for determining the search range for the sub-frame of the current frame by using the pitch period prediction value is provided.

[0030]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, an embodiment of a speech encoding method according to the present invention will be described.

(First Embodiment) FIG. 2 shows a first embodiment of the present invention.
It is a block diagram for demonstrating the basic operation | movement of the pitch period analysis in the speech encoding method which concerns on embodiment. In FIG. 1, digitized audio signals (hereinafter referred to as input audio signals) are sequentially input to an input terminal 11. The input audio signal is divided by the frame composing unit 12 into units called frames of a predetermined length (frame formation). The pitch period of the voice signal framed by the frame forming unit 12 is analyzed by a pitch period analyzing unit 13, and pitch period information is output from an output terminal 14.

Next, the operation of the frame forming section 12 will be described with reference to FIG. For convenience, the operation of the subframe forming unit in the frame / subframe forming unit 22 (FIG. 4) according to the second embodiment described later will also be described.

FIG. 3 shows a frame structure and a sub-frame structure of an input audio signal in the present embodiment. In FIG.
The symbol f (m) is attached to specify the frame. Here, a frame with m = 0 represents a frame to be coded from now (referred to as a current frame), and m ≦ −1.
Represents a past frame for which encoding has already been completed. Furthermore, fr represents a frame temporally future with respect to the current frame (this is referred to as a prefetch unit).

Assuming that the input audio signal is represented by s (n) and s (0) is the audio signal located at the head of the current frame, the following is defined between s (n) and f (m) and fr. The relationship holds.

F (m) = {s (n); n = m * NF〜 (m + 1) * NF−1} (1) (m ≦ 0) fr = ｛s (n); n = NF〜 NF + ND-1｝ (2) Here, NF represents the frame length, and ND represents the length of the prefetch section. In the present embodiment, a description will be given assuming that NF = 160 and ND = 120.

The subframe of the frame f (0) is defined as s
Assuming f (k), the following relationship holds between s (m) and sf (k). sf (k) = {s (n); n = k * NSF to (k + 1) * NSF-1} (0 ≦ k ≦ K−1) (3) where NSF is a subframe length, and K is Each represents the number of subframes. In this embodiment, NSF = 40, K = 4
The description is made as follows. N is a variable representing a sample, m
Is a variable representing a frame, and k is a variable representing a subframe.

After the frame and the subframe are constructed in this manner, the pitch period analysis section 13 performs a pitch period analysis. The pitch period analysis unit 13 is characterized in that the analysis of the pitch period is performed with the analysis center placed at the prefetch unit fr. The definition of the center of the analysis will be described later.

Existing techniques can be applied to the analysis of the pitch period. For example, several pitch analysis methods are introduced in "Sound Digital Signal Processing" published by Corona (Reference 4). In the present embodiment, the following pitch period analysis method is used.

First, in an LPC analyzer (not shown),
The input speech signal s (n) is analyzed to obtain LPC coefficients, and the LPC coefficients are used to construct a prediction filter for a transfer function A (z) represented by the following equation.

[0040]

(Equation 1)

Here, α (i) represents an LPC coefficient, and IP represents an analysis order. In the present embodiment, description will be made on the assumption that IP = 10. The prediction residual signal e (n) is obtained according to the following equation by passing the speech signal s (n) through the prediction filter of the transfer function A (z).

[0042]

(Equation 2)

Next, a windowing signal u (n) is obtained by multiplying the prediction residual signal e (n) by a Hamming window, and this windowing signal u (n) is obtained.
And a pitch period is extracted using a correlation value ρ (t) shown in the following equation.

[0044]

(Equation 3)

Here, NW represents a pitch period analysis length, and NC represents a pitch period analysis position. In this embodiment,
The description will be made on the assumption that NW = 160 and NC = 200. Also,
The analysis range of the correlation value ρ (t) is {20 ≦ t ≦ 147}.

Then, t when the correlation value ρ (t) becomes the maximum is output from the output terminal 14 as information on the pitch period T of the frame.
Output from The information of the pitch period T is encoded, for example, and transmitted to a decoder (not shown).

When a symmetric window such as a Hamming window or a rectangular window is used as the window function, the pitch analysis position can be considered as the center of the window function. When an asymmetric window having different left and right window shapes is used, the analysis position cannot be considered to be the center as in the case of the symmetric window. In this case, in the present embodiment, the position where the amplitude value of the window function is maximum is regarded as the center.

As described above, in the present embodiment, in order to obtain the pitch cycle of the future frame with respect to the current frame, this is used together with the pitch cycle of the current frame and the pitch cycle of the past frame as described later, and interpolation is performed. Obtain the pitch period predicted value of the subframe in the current frame, and by using this pitch period predicted value to obtain the pitch period of the subframe in the current frame, even if the pitch period fluctuates within the frame, The pitch period of the subframe can be obtained with high accuracy and with a small amount of calculation, and can be represented with a small amount of information.

(Second Embodiment) FIG. 4 shows a second embodiment of the present invention.
It is a block diagram for demonstrating the basic operation | movement of a sub-frame pitch extraction in the speech encoding method which concerns on embodiment. In FIG. 1, a digitized input audio signal to an input terminal 21 is transmitted to a frame / subframe forming unit 22.
Are divided into frames, and each frame is further divided into subframes. This frame / subframe forming unit 2
The audio signal framed in 2 and the audio signal subframed are input to a pitch period analysis unit 23 and a subframe pitch period extraction unit 24. The pitch period analysis unit 23 analyzes the pitch period T, and information on the pitch period T is input to the subframe pitch period extraction unit 24.

In the subframe pitch period extracting section 24,
On the basis of the pitch period T obtained by the pitch period analysis unit 23, a pitch period ST (k) for each subframe is obtained,
Information on the subframe pitch period ST (k) is output from the output terminal 25.

Next, the subframe pitch period extracting unit 24, which is a feature of this embodiment, will be described with reference to FIG. FIG. 5 is a block diagram showing the configuration of the sub-frame pitch period extracting unit 24. The sub-frame pitch period predicting unit 24 uses the information of the pitch period T from the pitch period analyzing unit 23 of FIG. Value calculation unit 102
Calculates the subframe pitch period prediction value for each subframe. The sub-frame pitch period calculation unit 104 receives the sub-frame pitch period prediction value obtained by the sub-frame pitch period prediction value calculation unit 102 and inputs the sub-frame pitch period prediction value from the frame / sub-frame configuration unit 22 shown in FIG. The sub-frame pitch period ST (k) is calculated for each sub-frame using the audio signal, and information on the sub-frame pitch period ST (k) is output to the output terminal 105.

The method of calculating the subframe pitch period will be described in more detail with reference to FIG. First, the pitch period determined by the pitch period analyzer 23 in the current frame f (0) is defined as T (0), and the pitch period determined in the past frame f (-1) is defined as T (-1). Subframe pitch period extraction unit 24
Then, a sub-frame pitch period prediction value STP (k) is obtained for each sub-frame using these two pitch periods. In the present embodiment, a description will be given of a method of obtaining a sub-frame pitch cycle predicted value STP (k) by interpolation from T (-1) and T (0). From the nature of the audio signal, it can be considered that the original subframe pitch period ST (k) is near the predicted subframe pitch period value STP (k). The reason is that the pitch cycle of the audio signal has a small temporal variation.

Next, the subframe pitch period prediction value ST
A method of obtaining the subframe pitch period ST (k) using P (k) will be described with reference to FIG. As described above, since it is assumed that the subframe pitch period ST (k) is near the predicted subframe pitch period value STP (k), only the vicinity can be considered as a range for pitch period extraction. . In FIG. 7, the 0th subframe sf
Subframe pitch period predicted value STP (0) in (0)
And the range NR of the correlation value calculation for extracting the subframe pitch period ST (0).

As described above, in this embodiment, the range of ± NR / 2 around the predicted subframe pitch period STP (0) is used as the range for calculating the correlation value for obtaining the subframe pitch period ST (0). . That is, STP (0) −
The correlation value ρ (t) defined by the above equation (6) is calculated only for the period included in the range of NR / 2 ≦ t ≦ STP (0) + NR / 2-1, and this is the maximum value. S when taking
Information on the relative pitch period ΔT (0) with respect to TP (0) is output from the output terminal 25. Therefore, subframe sf (0)
Can be obtained as the sum STP (0) + ΔT (0) of the predicted subframe pitch period STP (0) and the relative pitch period ΔT (0).
Similarly, the relative pitch period ΔT (k) of another subframe sf (k) is obtained, and this information is output from the output terminal 25. In this embodiment, NR = 8.

According to the present embodiment, the following effects can be obtained. First, since the range of the correlation value calculation for obtaining the sub-frame pitch period can be NR, a large reduction in calculation amount can be realized. If the sub-frame pitch period is obtained for each sub-frame by equation (6), when calculating for all pitch period candidates (128 candidates), about 40,000 multiply-accumulate operations are required. In contrast, when the present embodiment is applied, the correlation value calculation range is set to N
If R = 3, the product-sum operation is about 4,000 times, and about 9
0% calculation amount can be reduced. However, this is the result calculated under the pitch period analysis length NW = 160.

Conventionally, the amount of information required to represent the subframe pitch period is calculated in the entire range of 128 candidates for each subframe. If one frame is composed of four subframes, 7 bits * 4 = 28 bits were required. On the other hand, in this embodiment, 7 bits are required to represent the frame pitch period T (0), and 3 bits per subframe are required to represent the relative pitch period ΔT (k). The amount of information required is 7 bits per frame + 3
Bit * 4 = 19 bits, which leads to a bit number reduction of about 35%. Note that the pitch cycle T (-
Since 1) can be substituted by the pitch period T (0) in the processing of one frame before, it is not necessary to express this using a new bit.

Next, the processing procedure for calculating the sub-frame pitch period in this embodiment will be described with reference to the flowchart of FIG. First, in step S11, the input audio signal is framed and subframed. Next, step S12
With the center of the analysis in the current frame, the pitch period T (0)
Is calculated. Next, in step S13, a predicted sub-frame pitch period value STP (k) for each sub-frame is obtained using the pitch period T (0) of the current frame and the pitch period T (-1) of the past frame. Then, after setting the counter k to 0 in step S14, the relative pitch period ΔT (k) of the subframe sf (k) is calculated in step S15, and the counter k is incremented by 1 in step S16. In step S17, the counter k is K (K is the number of subframes).
It is determined whether or not they match. If k does not match K, the process returns to step S15 to continue the process, and if they match, the process ends.

(Third Embodiment) A third embodiment of the present invention will be described with reference to FIG. The difference between this embodiment and the second embodiment is that the predicted value of the subframe pitch period STP (k)
Is obtained by using the pitch cycle Tr of the pre-read unit fr.

The center of the sub-frame is set to the past frame f (-
If it is between the analysis position of the pitch period T (-1) of 1) and the analysis position of the pitch period T (0) of the current frame f (0), the predicted subframe pitch period STP (k) of the subframe
Is calculated using T (-1) and T (0). Similarly, the center of the subframe is the pitch period T (0) of the current frame f (0).
Is located between the analysis position of the subframe and the analysis position of the pitch period Tr of the prefetch unit fr, the predicted subframe pitch period value STP (k) of the subframe is obtained using T (0) and Tr. In the present embodiment, the sub-frame pitch cycle predicted value STP (k) is calculated using linear interpolation.

After the sub-frame pitch period predicted value STP (k) is obtained in this manner, the relative pitch period ΔT (k) is obtained for each sub-frame in the same manner as in the second embodiment.

In the present embodiment, the information necessary to represent the subframe pitch period ST (k) is the pitch period T (0) of the current frame, the pitch period of the look-ahead fr obtained in the processing one frame before. Since Tr can be used instead, only the pitch period Tr of the prefetch unit fr and the relative pitch period ΔT (k) for each subframe are required.

Next, the processing procedure in this embodiment will be described with reference to the flowchart shown in FIG. The processing in steps S21, S24, S25, S26 and S27 in FIG. 10 is performed in steps S11, S14, S15, S1 in FIG.
6 and S17, respectively, and the description is omitted here. A characteristic part of the processing procedure of FIG. 10 is that the pitch period Tr is calculated with the analysis center placed in the prefetch unit fr in step S22, and the subframe pitch period predicted value STP (k) is calculated in step S23. It is determined using the pitch period Tr of fr and the pitch period T (0) of the current frame f (0), or the pitch period T (0) of the current frame f (0) and the past frame f (-1). The point is that either one of the values obtained using the pitch period T (-1) is used.

In the present embodiment, each of the predicted sub-frame pitch periods STP (k) has a pitch period T (-1) of the past frame f (-1) and a pitch period T (0) of the current frame f (0). )
, Or as an interpolated value of T (0) and the pitch period Tr of the prefetch unit fr, so that a more accurate predicted value can be obtained.

For example, when the actual pitch period fluctuates as shown in FIG. 11, the pitch period T (-1) of the past frame and the pitch period T (0) of the current frame as in the second embodiment. ) Based on the subframe pitch period predicted value STP
In the method of obtaining (k), there may be a problem that the reliability of the predicted subframe pitch period STP (k) is reduced. In contrast, in the present embodiment, the prefetch unit f
Since the pitch period Tr of r accurately represents the actual pitch period, the interpolated value can also accurately represent the actual pitch period.

(Fourth Embodiment) A fourth embodiment of the present invention will be described with reference to FIG. The difference between this embodiment and the third embodiment is that the subframe pitch period prediction value STP
When (k) is obtained, the pitch period Tr of the pre-read unit fr and the pitch period T (0) of the current frame f (0) are used.

The processing procedure of this embodiment will be described with reference to the flowchart shown in FIG. The processes in steps S31 and S34 to S37 in FIG. 13 are the same as the processes in steps S11 and S14 to S17 in FIG.

In step S32, a pitch period Tr is calculated with the analysis centered on the prefetch section fr. next,
In step S33, the sub-frame pitch cycle predicted value STP (k) is calculated using the pitch cycle Tr of the prefetch unit fr and the pitch cycle T (0) of the current frame f (0).

In the present embodiment, the sub-frame pitch period predicted value STP (k) can be obtained by interpolation from the pitch period Tr of the prefetch section fr and the pitch period T (0) of the current frame f (0). Compared with the third embodiment, there is an advantage that the length of the prefetch unit fr can be shortened, and thus the delay can be reduced.

(Fifth Embodiment) A fifth embodiment of the present invention will be described with reference to FIG. The difference between this embodiment and the fourth embodiment is that the subframe pitch period prediction value STP
When (k) is obtained, the pitch period Tr of the pre-read unit fr and the pitch period T (-1) of the past frame f (-1) are used.

According to the present embodiment, the following effects can be obtained. Accurate analysis of the pitch period is generally difficult and there is no established method to date. In the present embodiment, the method based on the correlation analysis represented by the equation (6) is described. However, also in this case, a correct pitch period cannot always be obtained, and a pitch period equal to an integer fraction of the actual pitch period is calculated. In some cases, the pitch period may be obtained, or conversely, an integral multiple of the pitch period may be obtained.

As shown in FIG. 14, the current frame f
If the analysis of the pitch period T (0) of (0) fails and deviates greatly from the actual pitch period, the pitch period Tr of the prefetch unit fr and the pitch period T (-) of the past frame f (-1)
Interpolation is performed using 1) and the subframe pitch period prediction value ST
If P (k) is obtained, the pitch period can be expressed efficiently as in the case where the pitch period T (0) of the current frame f (0) is accurately obtained.

The processing procedure of this embodiment will be described with reference to the flowchart shown in FIG. The processing of steps S41, S42, S45, S47, S48, S49 and S50 in FIG.
22, S23, S24, S25, S26, and S27 are the same as those described above, and thus description thereof is omitted here. A characteristic part of the present embodiment is the processing of steps S43, S44 and S46, and these processings will be described.

In steps S43 and S44, the average value of the pitch period Tr obtained by the prefetch unit and the pitch period T (-1) of the past frame is calculated in step S43, and the average value and the pitch period T ( 0) is compared with the following equation (7) (step S44).

[0074]

(Equation 4)

If equation (7) holds, the current frame f
Assuming that the pitch period T (0) of (0) is reliable, the pitch period T (0) of the current frame f (0) is calculated to calculate the predicted subframe pitch period STP (k) as shown in step S45. Is used. Conversely, when the equation (7) does not hold, as shown in step S46, the sub-frame pitch is calculated using only the pitch cycle Tr of the prefetch unit fr and the pitch cycle T (-1) of the past frame f (-1). Calculate the cycle prediction value STP (k).

(Sixth Embodiment) A sixth embodiment of the present invention will be described with reference to FIG. 16, an input terminal 201, a frame / subframe forming unit 202, and a pitch period analyzing unit 203 are denoted by reference numerals 101 to 101 in FIG.
Since it is the same as the element with 103, the description is omitted here.

The feature of this embodiment lies in that the present invention is applied to a speech coding method using an adaptive codebook. The adaptive codebook is a codebook that stores a plurality of adaptive vectors generated by repeating a past drive signal sequence at a cycle included in a predetermined range.

Subframe pitch period predicted value calculating section 20
In step 4, a predicted subframe pitch period value STP (k) is calculated for each subframe based on the pitch period obtained by the pitch period analysis unit 203. Any of the methods described with reference to FIGS. 6, 9, 12, and 14 can be applied to the method of calculating the predicted subframe pitch period STP (k). Further, the effect of this is as described above. However, in the method described with reference to FIG. 6, the center of the pitch period analysis in the pitch period analysis unit 203 is on the current frame f (0), and FIG.
In the methods described with reference to FIGS. 12 and 14, there is a difference that the center of the pitch period analysis is located on the prefetch unit fr. In the present embodiment, an example based on the method described in FIG. 6 will be described.

Subframe pitch period predicted value calculating section 20
4, the pitch period T obtained by the pitch period analysis unit 203
A subframe pitch period predicted value STP (k) is obtained by interpolation from (0) and the pitch period T (-1) of the f (-1) frame.

The search range determining section 205 determines a search range for obtaining the subframe pitch period ST (0), that is, a search range of the adaptive vector for the adaptive codebook 206. Specifically, the subframe pitch period prediction value STP
The search range is ± NR / 2 around (0). That is, STP (0) −NR / 2 ≦ t ≦ STP (0) + NR / 2
The search is performed only for the periods included in the range of -1. The search is performed in the following manner.

The adaptive vector q corresponding to the pitch period t included in the search range determined by the search range determination unit 205
(t, n) is extracted from the adaptive codebook 206, and is passed through a perceptually weighted synthesis filter 207 to generate a synthesized signal p (t, n). Then, the degree of contribution cntb (t) between the synthesized signal p (t, n) and the target signal r (n) obtained by passing the input audio signal through the perceptual weight filter 208 is calculated by the subtractor 209 via the subtractor 209. At 210, it is calculated according to the following equation.

[0082]

(Equation 5)

Among all the pitch period t candidates included in the search range, the pitch period t when the contribution cntb (t) is the maximum is defined as the subframe pitch period ST (0), and the subframe pitch period ST ( 0) is subtracted from the predicted sub-frame pitch period STP (0) to obtain a relative pitch period ΔT
(0). The process of calculating the relative pitch period ΔT (0) is performed for all subframes sf (k), and the relative pitch period ΔT (k) of all subframes is determined.

Finally, information on the pitch period T (0) obtained by the pitch period analysis section 203 is output from the output terminal 212,
From the output terminal 211, the degree of contribution cntb is calculated by the distortion calculator 210.
Information on the relative pitch period ΔT (k) when (t) is maximum is output.

Next, the processing procedure of this embodiment will be described with reference to FIGS.
This will be described with reference to the flowchart shown in FIG. Figures 17-1
8, steps S101, S102, S103, S
The processing of steps 104, S114, and S115 is the same as the processing of steps S11, S12, S13, S14, S16, and S17 in FIG. Hereinafter, processing unique to the present embodiment will be described.

In step S105, the subframe sf
Determine the search range of (k). The search range is determined using the predicted subframe pitch period STP (k) as described above. In step S106, a variable cntbmax and an initial value of the subframe pitch period ST (k) are given. c
A sufficiently large value that can be an initial value is given to ntbmax, and a minimum pitch period is given to ST (k). Step S
At step 107, an adaptive vector q (t, n) corresponding to the cycle t included in the search range is selected from the adaptive codebook, and step S
At 108, the adaptive vector q (t, n) is passed through a perceptually weighted synthesis filter to generate a synthesized signal p (t, n). Step S
At 109, the contribution cntb (t) between the previously obtained target signal r (n) and the composite signal p (t, n) is calculated. At step S110, the contribution cntb (t) at the pitch cycle t is calculated.
And the size of cntbmax.

Here, if cntb (t)> cntbmax holds, then in step S111, cntbma
x is given cntb (t) and ST (k) is given t, and the flow advances to step S112. If cntb (t)> cntbmax does not hold, the process proceeds to step S112.

In step S112, it is determined whether or not all candidates for the pitch period t included in the search range have been searched.
If so, the process proceeds to step S113; otherwise, the process returns to step S107, and the process is continued using the pitch period t for which search has not been performed. Step S113
Then, the relative pitch period ΔT (k) of the subframe sf (k)
Is calculated using the sub-frame pitch period ST (k) and the sub-frame pitch period predicted value STP (k), and step S1 is performed.
Proceed to 14.

(Seventh Embodiment) A seventh embodiment of the present invention will be described with reference to the flowcharts of FIGS.
The present embodiment differs from the sixth embodiment in that the method described with reference to FIG. 9 is applied to the calculation of the predicted value of the subframe pitch period.

According to the present embodiment, since the accuracy of the predicted sub-frame pitch period STP (k) with respect to the actual pitch period is improved as described above, the sub-frame sf
(k), the search range of the subframe pitch period ST (k) can be reduced, and as a result, the number of bits and the amount of calculation required to express the relative pitch period ΔT (k) can be reduced. is there. 19 to 20, the processing in steps S201 and S204 to S215 is performed in steps S101 and S10 in FIGS.
Since the processing is the same as that of steps S4 to S115, the description is omitted here.

The characteristic processing in this embodiment is the processing in step S
202 and S203. In step S202, the pitch period is analyzed by placing the center of the pitch period analysis in the look-ahead section. In step S203, the calculation of the sub-frame pitch cycle predicted value STP (k) is determined using the pitch cycle Tr of the prefetch unit fr and the pitch cycle T (0) of the frame f (0), or the frame f (0) Pitch period T (0)
And using the pitch period T (-1) of the frame f (-1).

(Eighth Embodiment) The eighth embodiment will be described with reference to the flowcharts of FIGS. The present embodiment is different from the seventh embodiment in that the method described with reference to FIG. 12 is applied to the subframe pitch cycle predicted value calculation unit.

According to the present embodiment, the size of the prefetch unit can be reduced as described above, so that there is an effect that the delay can be reduced. In FIGS. 21 to 22, the processes in steps S301, S302, and S304 to S315 are performed in steps S201, S2 in FIGS.
02 and the processing in S204 to S215, and a description thereof is omitted here.

The characteristic processing in this embodiment is step S3
In step S303, the sub-frame pitch cycle predicted value STP (k) is calculated using the pitch cycle Tr of the prefetch unit fr and the pitch cycle T (0) of the frame f (0).

(Ninth Embodiment) A ninth embodiment of the present invention will be described with reference to the flowcharts of FIGS.
The present embodiment is different from the eighth embodiment in that the method described with reference to FIG. 14 is applied to the subframe pitch cycle predicted value calculation unit.

According to the method of the present embodiment, even if the analysis of the pitch period T (0) in the frame f (0) fails as described above and the pitch period T (0) greatly deviates from the actual pitch period, the prefetch unit fr Pitch period Tr and frame f obtained by
By performing interpolation with the pitch period T (-1) of (-1), the predicted value of the subframe pitch period STP (k) can be accurately obtained, and the pitch period T (0) of the frame f (0) is The pitch period can be expressed efficiently as in the case where it is accurately obtained.

In FIGS. 23 and 24, step S40
1, S402 and S407 to S418 are the same as those in FIG.
Steps S301, S302 and S in 1 to 22
The processing is the same as the processing in steps S304 to S315, and
4 are performed in steps S403 to S406 in FIG.
5, the description is omitted here.

(Tenth Embodiment) A tenth embodiment of the present invention will be described with reference to FIG. 25, an input terminal 301, a frame / subframe forming unit 302, a pitch period analyzing unit 303, a subframe pitch period calculating unit 304, a search range determining unit 305, an adaptive codebook 306, a perceptual weighting synthesis filter 307, a perceptual weight filter 30
8, subtractor 309, distortion calculator 310, output terminal 311,
Reference numeral 312 is the same as the element denoted by reference numerals 201 to 212 in FIG. The feature of this embodiment is that the pitch cycle analysis position in the pitch cycle analysis unit 303 is determined by the pitch cycle analysis position determination unit 313.

The effects of the present embodiment will be described with reference to FIG. As shown in the figure, when the characteristics of the input audio signal change within a frame, if the pitch period analysis position is fixed, an accurate pitch period may not be obtained. In the case of FIG. 26, the fact that the signal of the pitch analysis target when the pitch period analysis position is fixed does not include a period component causes an inability to obtain an accurate pitch period. However, if the pitch analysis position can be made variable, the pitch period analysis position can be set at a position where a waveform having a pitch period component exists as shown in FIG. 26, so that an accurate pitch period can be obtained. . However, in this case, additional information for representing the pitch period analysis position is required.

In FIG. 25, pitch period analysis section 313
Determines the pitch period analysis position. In the present embodiment, using the short-term power of the input audio signal or the power of the prediction residual signal after passing through the low-pass filter as an index,
A method of setting a pitch period analysis position in a portion where the power becomes large is used. The information of the pitch cycle analysis position is sent to the pitch cycle analysis unit 303 and output from the output terminal 314 at the same time. The pitch cycle analysis unit 303 analyzes the pitch cycle according to the information on the pitch cycle analysis position transmitted from the pitch cycle analysis position determination unit 313.

(Eleventh Embodiment) An eleventh embodiment of the present invention will be described with reference to FIG. 27, an input terminal 401, a frame / subframe forming unit 402,
Pitch cycle analysis section 403, subframe pitch cycle calculation section 404, search range determination section 405, adaptive codebook 406,
Perceptual weighting synthesis filter 407, perceptual weight filter 4
08, a subtractor 409, a distortion calculator 410, an output terminal 41
1, 412 are the same as the elements denoted by reference numerals 201 to 212 in FIG.

This embodiment is characterized in that the pitch period analysis unit 4
The point is that the continuity determination unit 413 determines whether or not the change in the pitch period obtained in step 03 is continuous, and the search range of the adaptive codebook 406 is determined according to the determination result. That is, in the present embodiment, when it is determined that the change in the pitch period is continuous, the predicted value of the subframe pitch period is determined as described above, and the search range determination unit 405 determines the predicted value of the adaptive codebook 406 based on the value. Determine the search range. If it is determined that they are not continuous, the full code search unit 414 searches the adaptive codebook 406
The search is performed for all the candidates in.

This embodiment has the following effects.
When the pitch period obtained by the pitch period analysis unit 403 is continuously changing, the method of interpolating the pitch period and searching only the periphery thereof as described above works effectively. However, the pitch period of an audio signal does not always change stably, and the pitch period may change greatly, or there may be no pitch component such as a silent part. In such a case, if the search range is limited by forcibly interpolating based on the assumption that the pitch period is continuously changing, serious quality degradation is caused. When the pitch cycle is not continuous as described above, in the present embodiment, the above-described problem can be avoided by switching to search all candidates in the adaptive codebook 406.

The same effect can be obtained by switching whether or not to search all candidates according to the strength of the periodicity at the pitch period T obtained by the pitch period analysis unit 403.
That is, when the variable representing the strength of the pitch periodicity as represented by ρ (T) in equation (6) is equal to or smaller than a predetermined threshold, the all search unit 414 searches all candidates. By switching the switch 415 to the above, it is possible to avoid deterioration in a region where no pitch period exists.

The continuity determining unit 413 determines the continuity of the pitch period using the pitch period obtained by the pitch period analyzing unit 403. In the present embodiment, a case will be described in which the pitch period T (0) obtained in the frame f (0) and the pitch period T (-1) obtained in the frame f (-1) are used.
r pitch period Tr and frame f (0) pitch period T
When (0) is used, or when the continuity of the pitch period Tr of the prefetch unit fr and the pitch period T (-1) of the frame f (-1) is used, the pitch period Tr of the prefetch unit fr and the frame f
The present invention can also be applied to the case where the continuity of the pitch period T (0) of (0) and the pitch period T (-1) of the frame f (-1) is used. The continuity of the pitch period is determined according to the following equation (9).

[0106]

(Equation 6)

When the equation (9) is satisfied, it is determined that the pitch period is continuously changing, and the switch 415 selects the output of the search range determination unit 405 and performs the limited search as described above. Adaptive codebook 406 for range only
Search for. If the equation (9) does not hold, it is determined that the pitch period is not continuously changing, and the switch 415
Selects the full search unit 414 and searches all candidates in the adaptive codebook 406. The determination result of the continuity determination unit 414 is output from the output terminal 416. Further, if the output from the continuity determining unit 414 indicates that the pitch period is not continuously changing, the pitch period obtained by the pitch period analyzing unit 403 does not need to be output from the output terminal 412. .

(Twelfth Embodiment) A twelfth embodiment of the present invention will be described with reference to FIG. 28, an input terminal 501, a frame / subframe forming unit 502,
The pitch period analyzer 503, the subframe pitch period calculator 504, the adaptive codebook 506, the perceptual weighting synthesis filter 507, the perceptual weight filter 508, the subtractor 509, the distortion calculator 510, and the output terminals 511 and 512 are shown in FIG. Are the same as the elements denoted by reference numerals 201 to 204 and 206 to 212, and the description thereof is omitted here.

The feature of the present embodiment is that a relative pitch pattern codebook 505 composed of a plurality of sets of relative pitch period patterns (referred to as relative pitch patterns) representing variations in subframe pitch periods of a plurality of subframes is provided. The relative pitch period of the plurality of subframes is regarded as a vector, and the relative pitch pattern in the relative pitch pattern codebook 505 most similar to the vector is selected.

According to the present embodiment, the following effects can be obtained. When the relative pitch period ΔT (k) is scalar-quantized for each subframe, for example, as described above, the search range NR = 8
, Three bits are needed to represent it.
That is, to represent the information of the relative pitch period ΔT (k),
Only 3 bits × 4 subframes = 12 bits per frame are required.

On the other hand, in the present embodiment, there is provided a relative pitch pattern codebook 505 having, as a relative pitch pattern, a pattern of a relative pitch cycle having a high appearance frequency.
Here, suppose four subframes (sf (0) to sf (3))
) Represents one vector (four dimensions) with the relative pitch periods ΔT (0) to ΔT (3), and if the relative pitch pattern codebook 505 is composed of 128 (7 bits) types of relative pitch patterns, there are originally four relative pitch patterns. Twelve bits were required to represent the relative pitch period ΔT (k) of the subframe.
It has the advantage of being represented in bits and achieving significant bit reduction.

The relative pitch pattern codebook 505 has J types of representative relative pitch patterns pv (j) obtained in advance as shown in the following equation. pv (j) = {v (j, k); k = 0−K−1} (10) (0 ≦ j ≦ J−1) The adder 513 predicts the subframe pitch period as shown in the following equation. The value STP (k) and the j-th relative pitch pattern pv (j) are added to obtain a set Tx (j) of pitch periods of K subframes. Tx (j) = STP (k) + v (j, k) (11) Based on the pitch period set Tx (j) thus obtained, when the contribution over the K subframes is maximized In a closed loop. At this time, the internal state of adaptive codebook 506 is updated for each subframe. At this time, the number of subframes is not limited to K, and a relative pitch period with respect to the number of subframes of 2 or more and K or less can be obtained. Also, in order to reduce the amount of calculation, first, the candidates of the relative pitch pattern are limited by a simple index having a small amount of calculation (for example, a numerator term of the degree of contribution), and an accurate contribution is obtained for the remaining candidates. Finally, one relative pitch pattern may be determined.

(Thirteenth Embodiment) A thirteenth embodiment of the present invention will be described with reference to FIG. 29, an input terminal 601, a frame / subframe forming unit 602,
Pitch cycle analyzer 603, subframe pitch cycle calculator 604, relative pitch pattern codebook 605, adaptive codebook 606, perceptual weighting synthesis filter 607, perceptual weight filter 608, subtractor 609, distortion calculator 610, output terminal 611 , 612 and the adder 613 are the same as the elements denoted by reference numerals 501 to 513 in FIG.

A feature of the present embodiment is that the relative pitch pattern is determined in consideration of the influence of the random codebook 614 together with the adaptive codebook 606. With such a configuration, a more accurate relative pitch pattern can be determined.

At this time, the gain by which the adaptive vector is multiplied by the multiplier 615 is given from the terminal 616. This gain is an ideal gain gopt expressed by the following equation, or
Here, an element closest to the ideal gain gopt in an adaptive vector gain codebook (not shown) is used.

[0116]

(Equation 7)

Here, r (n) is a target vector, and p (t, n)
Represents a signal obtained by passing the adaptation vector at the pitch period t through the perceptual weighting synthesis filter 607.

Similarly, the gain by which the noise vector is multiplied by the multiplier 617 is given from a terminal 618. As this gain, an ideal gain bopt expressed by the following equation or an element closest to the ideal gain bopt in a noise vector gain codebook (not shown) is used.

[0119]

(Equation 8)

Here, t (n) is a target vector, and e (i, n)
Represents a signal obtained by passing the noise vector at the index i through the perceptual weighting synthesis filter 607.

When the relative pitch pattern at which the contribution over a plurality of subframes is maximized is determined,
The output terminal 611 outputs the index of the relative pitch pattern at that time. The output terminal 612 outputs information on the pitch period obtained by the pitch period analysis unit 603.

At this time, the internal state of adaptive codebook 606 is updated for each subframe. The number of subframes at this time is not limited to K, and a relative pitch period for the number of subframes of 2 or more and K or less can be obtained. Also, in order to reduce the amount of calculation, candidates of relative pitch patterns are first limited by a simple index with a small amount of calculation (for example, a numerator term of the degree of contribution), and an accurate degree of contribution is obtained for the remaining candidates. Finally, one relative pitch pattern may be determined.

(Fourteenth Embodiment) A fourteenth embodiment of the present invention will be described with reference to FIG. In the present embodiment, the CE
This is an example in which the present invention is applied to an LP encoder (speech coding apparatus). Further, in this embodiment, one representative configuration is obtained by combining some of the embodiments described above, but the present invention is not limited to this.
Various combinations can be applied to the CELP scheme.

In FIG. 30, a digitized audio signal is input from an input terminal 501, and a frame / subframe forming unit 502 forms a frame and a subframe. After that, the audio signal divided into frames and subframes is provided to an LPC coefficient analysis unit 521,
The LPC coefficient is calculated by the LPC analysis. This LP
The C coefficient is used when determining the transfer functions of the perceptual weight filter 508 and the perceptual weighted synthesis filter 507.

LP calculated by LPC coefficient analysis section 521
The C coefficient is quantized by an LPC coefficient quantization unit 522, and an index obtained by the quantization is provided to a multiplexer 523 and multiplexed with other information described later. The LPC coefficient decoded after quantization is used when determining the transfer function of the perceptually weighted synthesis filter 507.

The input voice signal is also supplied to a pitch period analysis position determination section 525. Pitch cycle analysis position determination unit 5
In step 25, a predetermined number of powers of a predetermined short section of the audio signal are obtained, and an index representing an area where the short section power is the largest is provided to the pitch period analysis unit 503 and the multiplexer 523. Pitch cycle analyzer 503
Then, the pitch period is analyzed using the audio signal, the prediction residual signal, the prediction residual signal after passing through the low-pass filter, and the like, centering on the analysis position determined by the pitch period analysis position determination unit 525.

The continuity between the pitch period obtained here and the pitch period obtained in the previous frame processing is judged by the continuity judgment section 526, and the judgment result is given to the multiplexer 523. If the result of the determination is continuous, the pitch period obtained by the pitch period analysis unit
3, and a sub-frame pitch period predicted value calculation unit 504 obtains a pitch period predicted value for each sub-frame. Search range determining section 530 determines a pitch period search range for each subframe based on the predicted subframe pitch period value. In this case, the adaptive vector of the pitch period included in the search range determined by the search range determination unit 530 is selected as a candidate.

When the result of determination by the continuity determining unit 526 is not continuous, the switch 531 is switched so that the full search unit 527 is selected, and all the adaptive vectors included in the adaptive codebook 528 are selected as candidates. In this case, the pitch cycle obtained by the pitch cycle analysis unit 503 does not need to be given to the multiplexer 523.

The multiplier 532 includes an adaptive vector selected from the adaptive codebook 528 and an adaptive vector gain codebook 5.
The product with the adaptive vector gain selected from 33 is taken. Similarly, the multiplier 534 generates the noise vector selected from the noise codebook 529 and the noise vector gain codebook 53.
5 and the product of the noise vector and the noise vector gain. The adder 536 calculates the sum of the signals obtained from the multipliers 532 and 534, and generates a drive vector.

The drive vector thus generated is passed through a perceptually weighted synthesis filter 507 to generate a synthesized vector. A subtractor 509 obtains a difference between a target vector obtained by passing the audio signal through an audibility weighting filter 508 and a synthesized vector, and calculates a distortion calculating unit 511 based on the difference signal.
Find the distortion with. An adaptive vector, an adaptive vector gain, a noise vector, and a combination of a noise vector gain when this distortion is minimized are efficiently searched.

For example, as shown in the flowchart of FIG. 31, there is a method in which an adaptive vector, an adaptive vector gain, a noise vector, and a noise vector gain are sequentially obtained for each subframe in this order. In addition, as shown in the flowchart of FIG. 32, in a configuration in which the adaptive vector gain and the noise vector gain are simultaneously optimized by vector quantization for each subframe, the adaptive vector, the noise vector, and the gain vector are obtained in this order. There are ways.

That is, in FIG. 31, first, at step S60
After setting the counter k to 0 at 1, the process proceeds to step S602.
In steps S603, S604, and S605, an adaptive vector, an adaptive vector gain, a noise vector, and a noise vector gain are sequentially obtained, and in step S606, the counter k is incremented by one. In step S607, it is determined whether the counter k matches K. If they do not match, the process returns to step S602 to continue the process. If they match, the process ends.

In FIG. 32, first, the counter k is set to 0 in step S701, and then in steps S702 and S7.
03, S704: adaptive vector, noise vector, gain vector (adaptive vector gain and noise vector gain)
Are sequentially obtained, and the counter k is incremented by 1 in step S705. In step S706, the counter k is K
It is determined whether or not they match. If they do not match, the process returns to step S702 to continue the process, and if they match, the process ends.

In this way, the adaptive vector, adaptive vector gain, noise vector, and index representing the noise vector gain when the distortion is minimized are assigned to the multiplexer 52.
Give to 3. When the continuity is determined to be continuous by the continuity determination unit 526, the index of the adaptive vector is represented as a relative value to the subframe pitch period predicted value.

In the multiplexer 523, the multiplexing target differs depending on whether the determination result of the continuity determining unit 526 is continuous or not. That is, the continuity determination unit 5
26 are continuous, the LPC coefficient index obtained by the LPC coefficient quantization unit 522, the analysis position index obtained by the pitch period analysis position determination unit 525, the determination information obtained by the continuity determination unit 526, the pitch period The pitch period information obtained by the analysis unit 503, the relative value of the adaptive vector to the subframe pitch period predicted value, the index of the adaptive vector gain, the noise vector index, and the index of the noise vector gain are multiplexed and encoded from the encoded data output terminal 524. Output as coded data. On the other hand, when the determination result of the continuity determination unit 526 is not continuous, the LPC coefficient index obtained by the LPC coefficient quantization unit 522, the determination information obtained by the continuity determination unit 526, the index of the adaptive vector, the index of the adaptive vector gain, A noise vector index and a noise vector gain index are multiplexed and output terminal 524
Output from

(Fifteenth Embodiment) A fifteenth embodiment of the present invention will be described with reference to FIG. This embodiment shows an example in which the present invention is applied to a CELP decoder (speech decoding device). In the present embodiment, a decoder unit corresponding to a CELP encoder unit configured by combining some of the embodiments described above is shown. However, the present invention is not limited to this. The present invention can be applied to a decoder section corresponding to an encoder section.

The multiplexed coded data is input from an input terminal 801, and is converted by a demultiplexer 802 into indices of various parameters. In LPC coefficient decoding section 811, LP based on the LPC coefficient index
The C coefficient is decoded and supplied to the synthesis filter 812. The subframe pitch period generation unit 803 is given information related to the pitch period, and generates a pitch period used in the adaptive codebook 804. Details of the generation of the pitch period will be described later with reference to FIG.

An adaptive vector is obtained from adaptive codebook 804 using the generated pitch period as an index. The adaptive vector is multiplied by a multiplier 806 with the adaptive vector gain obtained from the adaptive vector gain codebook 805 using the adaptive vector gain index as an index. Similarly, a noise vector obtained by using the noise vector index as an index from the noise codebook 807 is added to the noise vector gain codebook 8 using the noise vector gain index as an index.
Multiplier 809 with the noise vector gain obtained from
Multiply by

The two signals after the multiplication are added to an adder 810.
And a driving signal is generated, and the driving signal is given to the synthesis filter 812 to generate a synthesized signal. The synthesized signal is input to a post-filter 813, subjected to auditory improvement such as formant emphasis, pitch emphasis, and gain adjustment, and then output from an output terminal 815.

Next, the subframe pitch period generation section 803 will be described with reference to FIG. Information for determining the continuity of the pitch period is input from an input terminal 901, and switches 904 and 911 are switched according to the information. When the pitch cycle continuity determination information indicates “discontinuous”, a pitch cycle index for each subframe is input from the input terminal 902, and the subframe pitch is decoded by the subframe pitch cycle decoding unit 910 via the switch 904. The period is decoded, and is output to the output terminal 912 via the switch 911.
Output from

When the pitch cycle continuity determination information indicates “continuous”, a pitch cycle analysis position index, a pitch cycle index, and a relative pitch cycle index are given from an input terminal 902, and the pitch is determined via a switch 904. The period analysis position decoding unit 905, the pitch period decoding unit 906, and the relative pitch period decoding unit 907 are provided. The predicted subframe pitch period is calculated by the predicted subframe pitch period calculation unit 908 using the pitch period analysis position and the pitch period obtained as a result. The subframe pitch cycle calculation unit 909 calculates a subframe pitch cycle using the predicted subframe pitch cycle value and the relative pitch cycle, and outputs the calculated subframe pitch cycle from the output terminal 912 via the switch 911.

FIG. 35 shows the configuration of a computer system for realizing the speech encoding / decoding method according to the present invention.
This computer system is, for example, a personal computer, and includes a CPU 1001 for performing arithmetic processing, an input unit 1002 such as a keyboard, a pointing device and a microphone, an output unit 1003 such as a display and a speaker, and a ROM 1004 serving as a main memory.
And a RAM 1005, a hard disk device 1006 as an external storage device, a floppy disk device 1007
And an optical disk device 1008 connected by a bus 1009.

Here, the hard disk device 1006, floppy disk device 1007 and optical disk device 10
08, a program for executing the audio encoding process described in the above-described embodiment and encoded data are stored. In accordance with this program, an audio encoding process is performed by the CPU 1001 on the input audio signal input from the input unit 1001, and the encoded data read from the recording medium is processed by the CPU 1001.
001 performs audio decoding processing, and outputs
Output from By doing so, the voice encoding / decoding processing of the present invention can be performed using a normal personal computer.

The present invention can be implemented with various modifications as described below. (1) In the above embodiment, interpolation or extrapolation was used as a method for obtaining the subframe pitch period predicted value.
However, the present invention is not limited to this. For example, it is possible to improve the accuracy of prediction by increasing the number of pitch periods used for prediction, or to perform more advanced prediction such as a higher-order function or a spline function. By doing so, the error between the measured value and the predicted value of the subframe pitch is reduced, and the subframe pitch period can be represented with a smaller amount of information.

(2) When quantizing the pitch period obtained by the pitch period analysis unit, the quantization is improved by predicting the pitch period obtained in the past and quantizing the difference from the predicted value. It can also be done.

(3) When determining the search range from the predicted value of the sub-frame pitch period, the search range is not set only in the vicinity of the predicted value of the sub-frame pitch period. It can also be set to be close to one-half.

(4) For the purpose of calculating the pitch period of the pitch filter, which is one component of the perceptual weight filter, the embodiments described above may be applied. Similarly, the embodiments described so far may be applied for the purpose of calculating the pitch period of a pitch filter, which is a component of a post filter not explicitly shown here.
In this case, since additional information is not required, the present invention can be easily applied to a speech encoding device.

(5) The speech coding method according to the above-described embodiment may be switched and used for each frame. In that case, additional information is required to indicate which embodiment of the method has been used.

[0149]

As described above, according to the speech encoding method of the present invention, it is possible to reduce the amount of calculation when searching for a subframe pitch period while maintaining the quality of synthesized speech. The number of bits for representing information on the subframe pitch period can be effectively reduced.

That is, in the speech encoding method of the present invention,
Using the pitch period of at least two of the current frame, the past frame, and the future frame,
The subframe pitch period prediction value of the subframe of the current frame is obtained by interpolation, and the subframe pitch period is determined using the subframe pitch period prediction value. Predicting the predicted value of the subframe pitch period from the pitch period enables highly accurate prediction, and reduces the amount of calculation for obtaining the subframe pitch period and the number of bits representing the information.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining a basic principle of the present invention.

FIG. 2 is a block diagram for explaining a basic operation of pitch period analysis in the speech encoding method according to the present invention.

FIG. 3 shows a frame structure and a subframe structure.

FIG. 4 is a block diagram illustrating a speech encoding method according to a second embodiment of the present invention.

FIG. 5 is a block diagram illustrating a configuration of a subframe pitch extraction unit according to a second embodiment.

FIG. 6 is a diagram for explaining a method of calculating a subframe pitch period in the second embodiment.

FIG. 7 is a diagram for explaining a method of calculating a subframe pitch period in the second embodiment.

FIG. 8 is a flowchart illustrating a processing procedure for calculating a subframe pitch cycle in the second embodiment.

FIG. 9 is a diagram illustrating a method for calculating a subframe pitch period according to a third embodiment of the present invention.

FIG. 10 is a flowchart for explaining a processing procedure for calculating a subframe pitch cycle in the third embodiment;

FIG. 11 is a view for explaining effects of the third embodiment;

FIG. 12 is a diagram for explaining a subframe pitch period calculation method according to a fourth embodiment of the present invention.

FIG. 13 is a flowchart for explaining a processing procedure for calculating a subframe pitch period in the fourth embodiment;

FIG. 14 is a diagram for explaining a subframe pitch period calculation method according to the fifth embodiment of the present invention.

FIG. 15 is a flowchart for describing a processing procedure for calculating a subframe pitch cycle in the fifth embodiment.

FIG. 16 is a block diagram illustrating a speech encoding method according to a sixth embodiment of the present invention.

FIG. 17 is a diagram showing a part of a flowchart for describing a processing procedure for calculating a subframe pitch cycle in the sixth embodiment.

FIG. 18 is a diagram showing another part of the flowchart for describing the processing procedure for calculating the subframe pitch period in the sixth embodiment.

FIG. 19 is a diagram showing a part of a flowchart for describing a processing procedure of calculating a subframe pitch period in the seventh embodiment of the present invention.

FIG. 20 is a diagram illustrating another part of the flowchart for describing the processing procedure of calculating the subframe pitch period in the seventh embodiment.

FIG. 21 is a diagram showing a part of a flowchart for describing a processing procedure of calculating a subframe pitch period in the eighth embodiment of the present invention.

FIG. 22 is a diagram illustrating another part of the flowchart for describing the processing procedure of calculating the subframe pitch period in the eighth embodiment.

FIG. 23 is a diagram showing a part of a flowchart for explaining a processing procedure of calculating a subframe pitch period in the ninth embodiment of the present invention;

FIG. 24 is a diagram showing another part of the flowchart for explaining the processing procedure of calculating the subframe pitch period in the ninth embodiment;

FIG. 25 is a block diagram illustrating a speech encoding method according to a tenth embodiment of the present invention.

FIG. 26 is a view for explaining effects of the tenth embodiment;

FIG. 27 is a block diagram illustrating a speech encoding method according to an eleventh embodiment of the present invention.

FIG. 28 is a block diagram for explaining a speech encoding method according to a twelfth embodiment of the present invention.

FIG. 29 is a block diagram illustrating a speech encoding method according to a thirteenth embodiment of the present invention.

FIG. 30 is a block diagram illustrating a configuration of a speech encoding apparatus according to a fourteenth embodiment in which the speech encoding method of the present invention is applied to the CELP scheme.

FIG. 31 shows an adaptation vector according to a fourteenth embodiment;
Flowchart for explaining a procedure for obtaining an adaptive vector gain, a noise vector, and a noise vector gain

FIG. 32 shows an adaptation vector according to a fourteenth embodiment;
Flow chart for explaining a procedure for obtaining a noise vector and a gain vector

FIG. 33 is a block diagram showing a configuration of a speech decoding apparatus according to a fifteenth embodiment in which the speech decoding method of the present invention is applied to the CELP scheme.

FIG. 34 is a block diagram illustrating a configuration of a subframe pitch cycle generation unit according to the fifteenth embodiment.

FIG. 35 is a block diagram showing a configuration of a computer system for realizing a speech encoding / decoding method according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 11 ... Audio signal input terminal 12 ... Frame construction part 13 ... Pitch period analysis part 14 ... Pitch period information output terminal 21 ... Speech signal input terminal 22 ... Frame / subframe construction part 23 ... Pitch period analysis part 24 ... Subframe pitch period Extraction unit 25: Subframe pitch period information output terminal 201, 301, 401, 501, 601: Audio signal input terminal 202, 302, 402, 502, 602: Frame
Subframe configuration units 203, 303, 403, 503, 603... Pitch period analysis units 204, 304, 404, 504, 604... Subframe pitch period prediction value calculation units 205, 305, 405, 530. 306, 406, 506, 528, 606... Adaptive codebooks 207, 307, 407, 507, 607... Perceptual weighting synthesis filters 208, 308, 408, 508, 608 ... perceptual weight filters 209, 309, 409, 509, 609 ... Subtractors 210, 310, 410, 510, 610 Strain calculator 211, 311, 411, 511, 611 Relative pitch period information output terminal at minimum distortion 212, 312, 412, 512, 612 Pitch period information output Terminals 313, 525: Pitch period analysis position determination unit 314: Pitch Period analysis position information output terminals 413, 526: continuity determination unit 414, 527: full search unit 415, 531 ... switch 416: continuity determination result output terminal 505, 605: relative pitch pattern codebook 513, 536, 613, 619 ... Adder 521 LPC coefficient analyzer 522 LPC coefficient quantizer 523 Multiplexer 524 Coded data output terminal 529,614 Noise codebook 532, 534, 615, 617 Multiplier 533 Adaptive vector codebook 535 ... Adaptive vector gain codebook 801 ... Encoded data input terminal 802 ... Demultiplexer 803 ... Subframe pitch period generator 804 ... Adaptive codebook 805 ... Adaptive vector gain codebook 806,809 ... Multiplier 807 ... Noise codebook 808 ... Noise vector gain codebook 810 ... addition Device 811 LPC coefficient decoding section 812 synthesis filter 813 post filter 814 audio signal output terminal

Claims (14)

[Claims] An audio encoding method for dividing an input audio signal into frames of a predetermined length and performing an encoding process including a process of obtaining a pitch period of the input audio signal. A speech encoding method characterized by including a process of obtaining a pitch period of a frame in the future in terms of time and a process of encoding the pitch period. 2. An encoding process including a process of dividing an input speech signal into frames of a predetermined length, further dividing a speech signal of each frame into subframes, and finding a pitch period of the speech signal. In the speech coding method, a pitch period prediction value of a subframe in a current frame is obtained by using a pitch period of at least two frames of a past frame and a future frame for a current frame to be coded and a current frame. And calculating a pitch period of a subframe in the current frame using the pitch period prediction value. 3. The speech encoding method according to claim 2, wherein a pitch period of a subframe in the current frame is encoded. 4. A pitch filter for suppressing or emphasizing a pitch period component of an input speech signal, wherein a transfer function of the pitch filter is determined using a pitch period of a subframe in the current frame. The speech encoding method according to claim 2, wherein 5. An adaptive codebook storing a plurality of adaptive vectors generated by repeating past drive signal sequences at a period included in a predetermined range, and the adaptive vector extracted from the adaptive codebook is stored in an adaptive codebook. In a speech encoding method including a process of searching an adaptive vector having a cycle that minimizes an error between a signal obtained through a predetermined filter and a target vector from a predetermined search range, the input audio signal has a predetermined length. After the audio signal of each frame is further divided into subframes, the pitch period of the current frame to be encoded and at least two of the past and future frames with respect to the current frame Is used to determine the predicted pitch period of the subframe in the current frame. Using the predicted pitch period, the subframe in the current frame is calculated. Speech encoding method, characterized by determination of the search range for over arm. 6. The speech encoding method according to claim 2, wherein, when determining the pitch period of the frame, an analysis position of the pitch period is adaptively determined for each frame. 7. The speech coding method according to claim 2, wherein a method for obtaining a pitch period of a subframe in the current frame is selected according to the continuity of the pitch period. . 8. A relative pitch pattern codebook storing a plurality of relative pitch patterns representing a change in pitch period of a plurality of subframes, wherein a change in pitch period of the subframe is determined from the relative pitch pattern codebook. 3. A method according to claim 2, wherein the relative pitch pattern is represented by one relative pitch pattern selected based on the index.
The speech encoding method according to any one of claims 1 to 7. 9. A speech encoding apparatus which divides an input speech signal into frames of a predetermined length and performs an encoding process including a process of obtaining a pitch period of the input speech signal, comprising: A speech coding apparatus comprising: means for calculating a pitch cycle of a frame in the future in terms of time; and means for coding the pitch cycle obtained by the means. 10. An encoding process including a process of dividing an input audio signal into frames of a predetermined length, further dividing an audio signal of each frame into subframes, and obtaining a pitch period of the audio signal. In a speech coding apparatus, a pitch period prediction value of a subframe in a current frame is obtained by using a pitch period of at least two frames of a past frame and a future frame for a current frame to be coded and a current frame. And a subframe pitch period calculating unit that calculates a pitch period of a subframe in the current frame using the pitch period predicted value. 11. An adaptive codebook storing a plurality of adaptive vectors generated by repeating past drive signal sequences at a period included in a predetermined range, and extracting adaptive vectors extracted from the adaptive codebook. An audio encoding apparatus including a process of searching an adaptive vector having a cycle in which an error between a signal obtained through a predetermined filter and a target vector is minimized from a predetermined search range, the input audio signal having a predetermined length Means for dividing the audio signal of each frame into sub-frames, and the pitch of at least two of the past frame and the future frame with respect to the current frame to be encoded and the current frame. Subframe pitch period prediction value calculating means for obtaining a pitch period prediction value of a subframe in the current frame using the period; Speech coding apparatus characterized by having a search range determining means for determining the search range for the sub-frame in the current frame using the pitch period prediction value. 12. A computer-readable recording medium on which a program for performing an audio encoding process including a process of dividing an input audio signal into frames of a predetermined length and obtaining a pitch period of the input audio signal is recorded. A recording medium on which a program for executing a process of obtaining a pitch period of a future frame in time with respect to a current frame to be encoded and a process of encoding the pitch period is recorded. 13. A speech encoding process including a process of dividing an input speech signal into frames of a predetermined length, further dividing the speech signal of each frame into subframes, and finding a pitch period of the speech signal. A computer-readable recording medium on which a program to be executed is recorded, wherein a pitch of at least two of a past frame and a future frame is used for a current frame to be encoded and a current frame. A recording medium storing a program for obtaining a predicted pitch period of a subframe in a current frame, and executing a process of obtaining a pitch period of a subframe in the current frame using the predicted pitch period. 14. An adaptive codebook storing a plurality of adaptive vectors generated by repeating a past drive signal sequence at a cycle included in a predetermined range, and extracting adaptive vectors extracted from the adaptive codebook. A computer-readable recording of a program for performing a speech encoding process including a process of searching an adaptive vector having a cycle that minimizes an error between a signal obtained through a predetermined filter and a target vector from a predetermined search range. A recording medium, wherein an input audio signal is divided into frames of a predetermined length, and the audio signal of each frame is further divided into subframes. The pitch period of the subframe in the current frame is determined by using the pitch period of at least two of the current frame and the future frame. Determine the period predicted value, a recording medium recording a program for executing processing for determining the search scope for the subframes in the current frame by using the pitch period estimated value.