JP6713424B2 - Audio decoding device, audio decoding method, program, and recording medium - Google Patents

Description

The present invention relates to voice/acoustic signal (hereinafter, also simply referred to as voice) communication using a digital communication network, and particularly to a voice encoding technique for encoding an input voice and a voice decoding technique for generating a voice from a received voice code. Regarding

The frequency band of voice that can be transmitted by a conventional telephone system represented by an analog telephone is approximately 300 Hz to 3.4 kHz. This is determined by the International Telecommunication Union Telecommunication Standardization Sector (ITU-T) in consideration of the balance between the voice quality required to convey a message and the amount of information required for transmission, and is widely used worldwide. It is adopted by. In general, voice with a frequency band upper limit of 4 kHz or less is called a narrow band signal (or narrow band voice, also called telephone voice), and voice with a frequency of more than 4 kHz and about 7 kHz is called a wide band signal (or wide band voice). When voice is expressed by a pulse code modulation (PCM) method of a digital signal, it is desirable to sample a narrow band signal at 8 kHz and a wide band signal at 16 kHz according to the sampling theorem. For these reasons, a signal sampled at 8 kHz may be called a narrow band signal, and a signal sampled at 16 kHz may be called a wide band signal.

Due to recent developments in acoustic technology and advances in digital signal processing technology, the quality of sound of equipment used in daily life has been improved. In such a situation, there is a demand for a wider band in the voice of the telephone.

In order to efficiently transmit a voice signal using a digital communication network, a voice coding method is used. Speech coding for narrowband signals (also called narrowband speech coding) includes international standard methods such as ITU-T G.711 and ITU-T G.726. In addition, there are international standard methods such as ITU-T G.711.1 and ITU-T G.722 for speech coding for wideband signals (also referred to as wideband speech coding). A terminal (hereinafter, terminal) that performs voice communication includes an encoding device and a decoding device compatible with any one or more audio encoding methods. When the terminal supports a plurality of voice coding systems, the coding system used for the communication is switched at the start of communication. Conventionally, a call control protocol called SIP or H.323 (also called signaling) is used to switch the coding method, and the communication methods commonly used by communication terminals are determined by a predetermined priority order. Had chosen based on. For example, if both terminals are compatible with G.711.1 and G.711, communication is performed with G.711.1, which is wideband voice coding, and one is G.711.1 and G.711 and the other is G.722 and G. When .711 is supported, both terminals are compatible with wideband voice, but G.711 is used for voice encoding and communication is performed with narrowband voice.

The reason why the call control protocol is used to switch the coding method at the start of communication is that there is no compatibility between the code methods, but the call control protocol switching is to establish the voice communication between terminals. Is complicated and causes connection problems. Also, with IP phones that use the Internet as a communication network, it is possible to switch the encoding method by the call control protocol relatively freely, but it is not possible to use the conventional intra-company communication network or inter-communication carrier interconnection network. In the voice communication that passes through, if there is a facility on the communication path that allows only G.711, there is a problem that only G.711 can be used even if the terminal supports a plurality of encoding methods.

With respect to this problem, Patent Document 1 describes that wideband speech coding having complete compatibility with G.711 can be realized. If the wideband speech coding method is completely compatible with G.711, the switching procedure of the coding method is extremely simplified, and even if there is equipment on the communication path that only G.711 can pass through, wideband speech can be transmitted. Can be passed.

With reference to FIG. 1, a speech coding apparatus described in Patent Document 1 is shown. The voice input to the voice encoding device is accumulated in the input buffer 81, is divided into frames having a length of about 10 milliseconds to 20 milliseconds, and is sent to the band division filter 82. The band division filter 82 divides the input voice into a low frequency voice and a high frequency voice. The low frequency speech is sent to the low frequency speech encoding unit 83, and the high frequency speech is transmitted to the high frequency speech encoding unit 84. The high frequency speech encoding unit 84 encodes the high frequency speech to generate a high frequency code, and sends the high frequency code to the low frequency speech encoding unit 83. The low-frequency speech encoding unit 83 receives the low-frequency speech and the high-frequency code, and embeds the high-frequency code as a bit string of 1 or 0 in the LSB (Least Significant Bit) or MSB (Most Significant Bit) of the G.711 code. A low-pass code is generated and the low-pass code is sent to the packet composing unit 85. The packet composing unit 85 receives the low band code from the low band speech coding unit 83 and composes a packet using the low band code. The packet sending unit 86 receives the packet information created by the packet composing unit 85 and sends it as a voice packet to the packet communication network.

Referring to FIG. 2, a speech decoding device described in Patent Document 1 is shown. The voice packet output from the voice encoding device is received by the packet receiving unit 91 of the voice decoding device and accumulated in the reception buffer 92. The audio packet output from the reception buffer 92 is decoded by the low frequency audio decoding unit 94. Further, the high band code extraction unit 95 extracts a high band code from the speech code. The high frequency audio decoding unit 96 decodes the high frequency audio component from the extracted high frequency code. The checksum detection unit 93 determines whether or not the high-frequency code is embedded in the LSB or MSB of the low-frequency code in the voice code output from the reception buffer 92, and if it is embedded, the switch 97 is used. Is set on the high-frequency speech decoding unit 96 side, and the high-frequency speech component is sent to the band synthesis filter 98. If the checksum detector 93 determines that the high frequency band code is not embedded in the LSB or MSB of the low frequency band code, the switch 97 is set to the high frequency side-free side. That is, the high frequency sound component is not generated. The band synthesizing filter 98 synthesizes the output of the low band audio decoding unit 94 and the output of the high band audio decoding unit 96 into a wide band audio signal and outputs it.

Japanese Patent No. 47588687

However, Patent Document 1 only describes a part of the configuration for realizing wideband speech coding having complete compatibility with G.711. Specifically, it is described that the high-frequency speech encoding unit 84 simply encodes the high-frequency speech, and the high-frequency speech decoding unit 96 simply decodes the high-frequency speech component from the high-frequency code. Only that is mentioned. In order to realize wideband speech coding that is completely compatible with G.711, the quality of the wideband speech reproduced from the decoding device must be sufficiently good, and at least the narrowband decoded by the G.711 method must be used. It is necessary to be able to reproduce a wideband voice of higher quality than voice. In addition, according to the same idea, it is expected that wideband speech coding having complete compatibility with G.726 is realized, but Patent Document 1 does not specifically describe how to realize compatibility with G.726. ..

In view of the above points, an object of the present invention is to provide a voice encoding technique capable of improving the quality of wideband voice reproduced in voice communication of wideband voice.

In order to solve the above-mentioned problems, a voice encoding device according to a first aspect of the present invention is based on a band dividing unit that band-divides an input voice into a low-frequency voice and a high-frequency voice, and a decoded low-frequency voice. A high-frequency speech encoding unit that encodes high-frequency speech to generate a high-frequency code, a low-frequency speech encoding unit that encodes low-frequency speech and generates a low-frequency code in which a high-frequency code is embedded, and a low-frequency speech encoding unit. Includes a low-band speech decoding unit that decodes the code to generate decoded low-band speech, and a code sending unit that outputs the low-band code as a speech code.The low-band code is compatible with the ITU-T G.726 system. It has sex.

A speech decoding apparatus according to a second aspect of the present invention is a code receiving section for receiving a speech code output by the speech encoding apparatus according to the first aspect, and a low-frequency band for decoding a speech code to generate a decoded low frequency speech. A voice decoding unit, a high band code extracting unit that extracts a high band code embedded in the voice code, and a high band voice decoding unit that decodes the high band code based on the decoded low band voice to generate decoded high band voice. And a band synthesizing unit for synthesizing the decoded low-frequency speech and the decoded high-frequency speech and outputting the decoded speech.

According to the present invention, in audio encoding, high-frequency audio can be encoded with a small number of bits while minimizing the loss of information necessary for reproducing wideband audio. Further, in speech decoding, high-quality wideband speech can be reproduced by generating high-frequency speech that is auditorily less deteriorated in quality. That is, in voice communication of wideband voice, the quality of reproduced wideband voice can be improved.

FIG. 1 is a diagram illustrating a functional configuration of a conventional speech encoding device. FIG. 2 is a diagram illustrating a functional configuration of a conventional speech decoding device. FIG. 3 is a diagram illustrating a functional configuration of the speech encoding apparatus according to the embodiment. FIG. 4 is a diagram illustrating a functional configuration of the speech decoding device according to the embodiment. FIG. 5 is a diagram exemplifying a processing procedure of the speech encoding method according to the embodiment. FIG. 6 is a diagram illustrating a processing procedure of the speech decoding method according to the embodiment. FIG. 7 is a diagram illustrating a functional configuration of the high frequency audio encoding unit according to the embodiment. FIG. 8 is a diagram illustrating a functional configuration of the coefficient encoding unit according to the embodiment. FIG. 9 is a diagram exemplifying the functional configuration of the high frequency audio decoding unit according to the embodiment. FIG. 10 is a diagram illustrating a functional configuration of the coefficient decoding unit according to the embodiment. FIG. 11 is a diagram exemplifying the functional configuration of the low-frequency speech encoding unit according to the embodiment. FIG. 12 is a diagram exemplifying the functional configuration of the low-frequency speech encoding unit of Modification A. FIG. 13 is a diagram exemplifying the functional configuration of the low-frequency speech encoding unit of Modification A. FIG. 14 is a diagram exemplifying the functional configuration of the low-frequency speech decoding unit of the embodiment. FIG. 15 is a diagram exemplifying the functional configuration of the low-frequency speech decoding unit of modification B. FIG. 16 is a diagram exemplifying the functional configuration of the low-frequency speech decoding unit of modification B. FIG. 17 is a diagram exemplifying the functional configuration of the high frequency audio encoding unit according to the modification C. FIG. 18 is a diagram for explaining the effect of the modification C. FIG. 19 is a diagram illustrating a functional configuration of the designated band signal extraction unit of the modification C.

Hereinafter, embodiments of the present invention will be described in detail. In the drawings, components having the same function are denoted by the same reference numerals, and duplicate description will be omitted.

The symbol " ^- " used in the sentence should be written directly above the character just before it, but it is written immediately after the character due to the limitation of text notation. In the mathematical formula, these symbols are described at their original positions, that is, directly above the characters.

In the embodiments of the present invention, a speech coding apparatus that encodes input speech into a speech code and outputs the speech code, and a speech decoding apparatus that decodes the speech code output by the speech coding apparatus and outputs decoded speech. .. After converting one frame of input voice into a voice code and outputting the voice code, the voice encoding device processes the input voice of the next frame, and repeats this in the time period of the frame. The voice decoding device, after processing the voice code for one frame and outputting the decoded voice, processes the voice code of the next frame and repeats this in the time period of the frame.

As shown in FIG. 3, the speech coding apparatus according to the embodiment includes an input buffer 11, a band division filter (also referred to as a band division unit) 12, a low band speech coding unit 13, a high band speech coding unit 14, and a delay unit. 15, a low-frequency speech decoding unit 16 and a code transmitting unit 17 are provided. The speech coding method of the embodiment is realized by the processing of each step described later by this speech coding apparatus.

As shown in FIG. 4, the voice decoding device according to the embodiment includes a code receiving unit 21, a low band voice decoding unit 22, a high band code extraction unit 23, a delay unit 24, a high band voice decoding unit 25, and a band synthesis filter ( 26). The speech decoding method of the embodiment is realized by the processing of each step described later by this speech decoding device.

The voice encoding device and the voice decoding device are, for example, a special program loaded into a known or dedicated computer having a central processing unit (CPU), a main storage device (RAM: Random Access Memory), and the like. It is a special device configured. Each device executes each process under the control of the central processing unit, for example. The data input to each device and the data obtained by each process are stored in, for example, the main storage device, and the data stored in the main storage device is read as needed and used for other processes. .. Further, at least a part of each processing unit of each device may be configured by hardware such as an integrated circuit.

The processing procedure of the audio encoding method according to the embodiment will be described with reference to FIG.

In step S11, voice is input to the voice encoding device. The input voice x is stored in the input buffer 11, divided into frames each having a length of about 10 milliseconds to 20 milliseconds, and sent to the band division filter 12. The input sound x is wide band sound and the sampling frequency is 16 kHz. The input voice x is divided by the band division filter 12 into a low frequency voice x _L and a high frequency voice x _H having a sampling frequency of 8 kHz. The low-frequency speech x _L is sent to the low-frequency speech coding unit 13, and the high-frequency speech x _H is sent to the high-frequency speech coding unit 14. As the band division filter 12, a quadrature mirror filter (QMF) used in G.711.1 or G.722 can be used. Or, by using an appropriate low-pass filter and high-pass filter, the input sound x is low-pass filtered and the signal decimated to 1/2 sample is set as the low-pass sound x _L , and the input sound x is high-pass filtered to 1/ the thinned signal to 2 number of samples may be the high-band speech x _H.

In step S12, the high-band speech encoding unit 14, the decoded low-band speech x received from the low-band speech decoding section 16 to be described later ^- encodes the high-band speech x _H using the _L, and the high frequency encoding c _H It is sent to the delay unit 15. Details of the processing of the high frequency audio encoding unit 14 will be described later. The delay unit 15 has a memory for storing one frame of the high frequency code c _H , sends the high frequency code of one frame before to the low frequency speech encoding unit 13, and stores the received high frequency code. Since the delay unit 15 can be omitted, as will be described later, the high band code output from the high band speech coding unit 14 and the high band code one frame before output from the delay unit 15 Without distinction, it will be simply referred to as the high frequency code c _H.

In step S13, the low-frequency speech encoding unit 13 can use the same configuration as the low-frequency speech encoding unit 83 included in the conventional speech encoding device. That is, the low-frequency speech x _L and the high-frequency code c _H are received, and the low-frequency code c _{L in} which the high-frequency code is embedded as a bit string of 1 or 0 in the LSB or MSB of the G.711 code is output. The output of the low-frequency speech encoding unit 13 is sent to the code transmitting unit 17 and also to the low-frequency speech decoding unit 16.

In step S14, the low-frequency speech decoding unit 16 decodes the low-frequency code c _L received from the low-frequency speech encoding unit 13 and sends the decoded low-frequency speech x ⁻ _L to the high-frequency speech encoding unit 14. The low-frequency audio decoding unit 16 can use the same configuration as the low-frequency audio decoding unit 94 included in the conventional audio decoding device.

In step S15, the code sending unit 17 sends the low band code c _L received from the low band speech coding unit 13 to the communication network as a speech code.

The voice code c _L transmitted from the voice encoding device has complete bit compatibility with G.711, and when the voice decoding device corresponding to the conventional G.711 system receives the voice code c _L , Narrowband audio can be reproduced by the G.711 decoding system, and when the audio decoding device of the present invention receives the audio code c _L , wideband audio can be reproduced by the audio decoding method described later. Further, the voice code c _L can pass through the existing communication network compatible only with G.711.

The processing procedure of the speech decoding method according to the embodiment will be described with reference to FIG.

In step S21, the code receiving unit 21 receives the voice code c _L from the communication network and sends it to the low frequency voice decoding unit 22 and the high frequency code extracting unit 23.

In step S22, the low-frequency speech decoding unit 22 decodes the speech code c _L by the G.711 method and sends the decoded low-frequency speech x ⁻ _L to the delay unit 24. The delay unit 24 has a memory for storing one frame of the decoded low band speech x ^- _L , sends the decoded low band speech of one frame before to the high band speech decoding unit 25 and the band synthesis filter 26, and receives the received decoded low band. Memorize the voice. Since the delay unit 24 can be omitted, as will be described later, the decoded low-frequency voice output from the low-frequency voice decoding unit 22 and the decoded low-frequency voice one frame before output from the delay unit 24. Are not distinguished, and are simply referred to as decoded low frequency speech x ⁻ _L .

In step S23, the high frequency code extracting unit 23 extracts the high frequency code c _H from the speech code c _L. The configuration of the high band code extraction unit 23 can be the same as that of the conventional high band code extraction unit 95. That is, the bit string of 1 or 0 embedded in the LSB or MSB of the G.711 code is returned to the high frequency code c _H. The high band code c _H is sent to the high band speech decoding unit 25.

In step S24, the high-band speech decoding unit 25, the decoded low-band speech x ^- using the _L decodes the high frequency code c _H, decoded high-band speech x ^- Send _H to the band synthesis filter 26. Details of the process of the high frequency audio decoding unit 25 will be described later.

In step S25, the band synthesis filter 26, the decoded low-band speech x ^- _L and the decoded high-band speech x ^- from _H wideband decoded speech x ^- a synthesized and output. As the band synthesis filter 26, a quadrature mirror filter (QMF) used in G.711.1 or G.722 can be used as in the band division filter 12.

The voice decoding device is configured to include the checksum detection unit 93 and the switch 97 as described in Patent Document 1, and determines whether the high frequency code c _H is embedded in the received voice code c _L. It is also possible to perform a switching process of making a determination and outputting a wideband voice if embedded, and outputting a narrowband voice if not embedded.

The delay unit 15 included in the voice encoding device and the delay unit 24 included in the voice decoding device may be omitted. Since the speech coding apparatus has a feedback structure that decodes the low-band code c _{L in} which the high-band code c _H is embedded and codes the high-band speech x _H , if the delay unit 15 is omitted, in the speech coding apparatus decoded low-band speech x ^- _L and the decoded low-band speech x in the audio decoding device ^- and _L can not be matched. However, the difference is such that it is indistinguishable in terms of hearing, and there are few practical problems. If each delay unit is omitted, the delay time in voice communication can be shortened by one frame.

Hereinafter, a detailed configuration of the high frequency audio encoding unit 14 included in the audio encoding device will be described. As shown in FIG. 7, the high frequency audio encoding unit 14 includes a band division filter (also referred to as a high frequency band division unit) 31 _H , a band division filter (also referred to as a low frequency band division unit) 31 _L , and a power calculation unit 32. _H , 32 _L , linear prediction units 33 _H , 33 _L , a relative gain calculation unit 34, a coefficient coding unit 35, a gain coding unit 36, and a multiplexer (also referred to as a multiplexing unit) 37.

The high-band speech encoding unit 14, the high-band speech x _H and the decoded low-band speech x ^- _L is input. Taking the case where the sampling frequency of the input audio is 16 kHz and the frame length is 10 ms as an example, the high-frequency audio x _H and the decoded low-frequency audio x ^- _L both have a sampling frequency of 8 kHz and a frame length of 10 ms. Therefore, the number of samples in one frame is 80 samples.

The band division filter 31 _L divides the decoded low band speech x ^- _L into LL band speech x ^- _LL and LH band speech x ^- _LH each having a sampling frequency of 4 kHz. The band-division filter 31 _L may be the same as the band-division filter 12 of the audio encoding device, or may be a band-division filter having a different number of taps or characteristics from the band-division filter 12. Since the LL band voice x ^- _LL is not used in the high band voice encoding unit 14, the band division filter 31 _L may be configured to output only the LH band voice x ^- _LH . The LH band speech x ^- _LH is input to the linear prediction unit 33 _L and the power calculation unit 32 _L.

The linear prediction unit 33 _L applies a linear prediction analysis to the LH band speech x ⁻ _LH and outputs a p-th order LH band linear prediction coefficient a _LH (i) (where i=1, 2,..., P). To do. Here, a value of about 4 to 10 is generally used for p. Note that the p-th order linear prediction coefficient is a set of p values, but in the following, the index i will be omitted and simply expressed as a _LH , unless a linear prediction coefficient at a specific i is shown. .. a _LH can be regarded as a vector and is also called a linear prediction coefficient vector.

Power calculating portion 32 _L is, LH band speech x ^- calculating the power P _LH of one frame of _LH. At this time, the average power including the preceding and following frames, for example, the power for three frames including the signal one frame before and the signal one frame later, or 1/3 thereof may be set as the power for one frame. The same applies to the calculation of power for one frame.

The band division filter 31 _H divides the high frequency sound x _H into an HL band sound x _HL and an HH band sound x _HH each having a sampling frequency of 4 kHz. The band-division filter 31 _H may be the same as the band-division filter 12 of the speech coding apparatus, or may be a band-division filter having a different number of taps or characteristics from the band-division filter 12. Since the HH band speech x _HH is not used in the high band speech coding unit 14, the band division filter 31 _H may be configured to output only the HL band speech x _HL . The HL band speech x _HL is input to the linear prediction unit 33 _H and the power calculation unit 32 _H.

The linear prediction unit 33 _H applies a linear prediction analysis to the HL band speech x _HL and outputs a p-th order HL band linear prediction coefficient a _HL (i) (where i=1, 2,..., P). .. Below, like the LH band linear prediction coefficient a _LH , the index i is omitted and simply expressed as a _HL . Like a _LH , a _HL can be regarded as a vector and is also called a linear prediction coefficient vector.

The power calculator 32 _H calculates the power P _HL for one frame of the HL band voice x _HL .

The relative gain calculator 34 calculates the relative gain G _HL defined by the following equation. The relative gain G _HL is HL band speech x _HL of LH band speech x ^- is the relative gain for _LH, LH band speech x ^- power of the signal obtained by multiplying the relative gain G _HL to each sample of _LH is, HL band speech x _HL It becomes the same as the power P _HL .

The coefficient encoding unit 35 encodes the HL band linear prediction coefficient a _HL with M ₁ bits using the LH band linear prediction coefficient a _LH , and sends the coefficient code c ₁ to the gain encoding unit 36 and the multiplexer 37. How to determine M ₁ will be described later.

The gain encoding unit 36 encodes the relative gain G _HL with M ₂ bits using the LH band linear prediction coefficient a _LH and the coefficient code c ₁ , and sends the gain code c ₂ to the multiplexer 37. The method of determining M ₂ will be described later.

Explain how to determine M ₁ and M ₂ . According to Patent Document 1, even if a high band code of 16 bits per 160 samples of low band speech, that is, 8 bits per 80 samples is embedded in a low band code, the subjective quality of decoded low band speech is It is said that it does not deteriorate in comparison. Therefore, when the frame length is 10 milliseconds (80 samples), it is preferable to determine M ₁ and M ₂ so that M ₁ +M ₂ ≦8. As an example, M ₁ =4 and M ₂ =4.

Coefficient coding section 35, the the LH band linear prediction coefficients a _LH and HL band linear prediction coefficients a _HL using a correlation, to encode the HL band linear prediction coefficients a _HL. For example, the value of the HL band linear prediction coefficient a _HL may be estimated from the value of the LH band linear prediction coefficient a _LH , and the error between the HL band linear prediction coefficient a _HL and the estimated value a′ _HL may be encoded. Note that the estimation uses a statistical method using a voice database.

As shown in FIG. 8, the coefficient coding unit 35 includes an LSP conversion unit 351, an LSP conversion unit 352, an LSP estimation unit 353, and an error coding unit 354. The LSP conversion unit 351 converts the HL band linear prediction coefficient a _HL into an HL band line spectrum pair (hereinafter, the line spectrum pair is referred to as LSP) f _HL . The LSP is a kind of linear prediction parameter, and the p-th order linear prediction coefficient and the p-th order LSP can be mutually converted. Regarding the notation of the LSP, the index i (i=1, 2,..., P) is omitted as in the notation of the linear prediction coefficient, and when the index i is omitted, it can be regarded as a vector. The LSP conversion unit 352 converts the LH band linear prediction coefficient a _LH into the LH band LSPf _LH . The LSP estimation unit 353 estimates the value of the HL band LSPf _HL using the LH band LSPf _LH . For the estimation rule, a statistical method using a voice database can be used. For example, a conversion function may be defined, or the correspondence between the distribution of the LH band LSPf _{LH and} the distribution of the HL band LSPf _HL can be determined. It may be statistically examined and defined. The error encoding unit 354 encodes the error between the HL band LSPf _HL and the estimated value f′ _HL of the HL band LSP using, for example, the vector quantization method, and outputs the coefficient code c ₁ .

The gain encoding unit 36 encodes the relative gain G _HL by utilizing the correlation between the combination of the LH band linear prediction coefficient a _LH and the coefficient code c ₁ and the relative gain G _HL . For example, to estimate the value of the relative gain G _HL from a combination of LH band linear prediction coefficients a _LH and coefficient code c _1, the error between the relative gain G _HL and the estimated value G _'HL on a logarithmic scale (or decibels) It is good to encode. Note that the estimation may be performed by using a statistical method using a voice database.

The multiplexer 37 receives the coefficient code c ₁ output from the coefficient encoding unit 35 and the gain code c ₂ output from the gain encoding unit 36, and outputs the high frequency code c _H.

Details regarding speech analysis, including linear predictive analysis, are provided in Reference 1 below.
[Reference 1] Sadahiro Furui, "Digital Audio Processing," Tokai University Press, pp. 60-98

Hereinafter, a detailed configuration of the high frequency audio decoding unit 25 included in the audio decoding device will be described. As shown in FIG. 9, the high frequency audio decoding unit 25 includes a demultiplexer (also referred to as a code separation unit) 40, a band division filter (also referred to as a band division unit) 41, a power calculation unit 42, a linear prediction unit 43, and an inverse filter. 44, duplication unit 45, coefficient decoding unit 46, relative gain decoding unit 47, synthesis filter 48, power calculation unit 49, gain calculation unit 50, multiplication unit (also referred to as HL band multiplication unit) 51, relative gain prediction unit 52, coefficient A prediction unit 53, a random number unit 54, a synthesis filter 55, a power calculation unit 56, a gain calculation unit 57, a multiplication unit (also called an HH band multiplication unit) 58, and a band synthesis filter (also called a band synthesis unit) 59 are provided.

The decoded low-frequency speech x ^- _L and the high-frequency code c _H are input to the high-frequency speech decoding unit 25. The high frequency code c _H is input to the demultiplexer 40. The decoded low frequency sound x ⁻ _L is input to the band division filter 41.

The band division filter 41 has the same configuration as the band division filter 31 _L of the high frequency speech encoding unit 14, and the decoded low frequency speech x ⁻ _L is the LL band speech x ⁻ _LL and the LH band speech x ⁻ with a sampling frequency of 4 kHz. Split into _LH and. Since the LL band speech x ^- _LL is not used in the high band speech decoding unit 25, the band division filter 41 may be configured to output only the LH band speech x ^- _LH . The LH band speech x ^- _LH is input to the linear prediction unit 43 and the power calculation unit 42.

Linear prediction unit 43, LH band speech x ^- by applying linear prediction analysis to _LH, and outputs a p-th order LH band linear prediction coefficients a _LH. The LH band linear prediction coefficient a _LH is input to the inverse filter 44, the coefficient decoding unit 46, the relative gain decoding unit 47, and the coefficient prediction unit 53.

Power calculation unit 42, like the power calculating portion 32 _L of the high-band speech encoding unit 14, LH band speech x ^- calculating the power P _LH of one frame of _LH. The power P _LH is input to the gain calculator 50 and the gain calculator 57.

Inverse filter 44 is a FIR filter that the LH band linear prediction coefficients a _LH and filter coefficients, LH band speech x ^- seeking LH band linear prediction residual e _LH from _LH, and sends the replicated portion 45. Here, x ^- _LH (j) is the j-th sample of the LH band speech x ^- _LH , e _LH (j) is the j-th sample of the LH band linear prediction residual, and j=1 is the start sample of the current frame. , J=N represents the last sample of the current frame, then e _LH (j) is expressed by the following equation.

When one frame consists of 80 samples, N=80. When j-i is negative, the sample position in the past frame is represented as a relative sample position with the head sample of the current frame as a reference. When representing a set of sample values for one frame, the index j is omitted.

The duplication unit 45 duplicates the LH band linear prediction residual e _LH and outputs the HL band driving sound source e _HL as in the following equation. The HL band drive sound source e _HL is input to the synthesis filter 48.

The demultiplexer 40 divides the high frequency code c _H into a coefficient code c ₁ and a gain code c ₂ . The coefficient code c ₁ is input to the coefficient decoding unit 46, the relative gain decoding unit 47, the relative gain prediction unit 52, and the coefficient prediction unit 53. The gain code c ₂ is input to the relative gain decoding unit 47 and the relative gain prediction unit 52.

Coefficient decoding unit 46 decodes the coefficient code c ₁ using LH band linear prediction coefficients a _LH, HL band decoded linear prediction coefficients a ^- outputs the _HL. As shown in FIG. 10, the coefficient decoding unit 46 includes an LSP conversion unit 461, an LSP estimation unit 462, a reconstruction unit 463, and a coefficient conversion unit 464. The LSP conversion unit 461 and the LSP estimation unit 462 are the same as the LSP conversion unit 352 and the LSP estimation unit 353 of the coefficient coding unit 35. Reconstruction unit 463 uses the estimated value f _'HL coefficient code c ₁ and HL band LSP, the decoding method corresponding to the error encoding, HL band decoding LSPf ^- reconstructing the _HL. The coefficient conversion unit 464 converts the HL band decoded LSPf ^- _HL into the HL band decoded linear prediction coefficient a ^- _HL and outputs the coefficient. HL band decoded linear prediction coefficients a ^- _HL is input to the synthesis filter 48.

The relative gain decoding unit 47 decodes the gain code c ₂ using a combination of LH band linear prediction coefficients a _LH and coefficient code c _1, decoding relative gain G ^- seeking _HL. The decoded relative gain G ^- _HL is input to the gain calculation unit 50. As the decoding method, a method corresponding to the encoding method of the gain encoding unit 36 of the high frequency speech encoding unit 14 is used. For example, the relative gain G _HL of the relative gain G _HL is calculated from the combination of the LH band linear prediction coefficient a _LH and the coefficient code c ₁ . The decoded relative gain G ⁻ _HL can be obtained by a method of estimating the value and adding the error represented by the gain code c ₂ to the estimated value G′ _HL of the relative gain on a logarithmic scale or by multiplying by a linear scale.

The synthesis filter 48 is an IIR filter (also referred to as an AR filter) that uses the HL band decoding linear prediction coefficient a ⁻ _HL received from the coefficient decoding unit 46 as a filter coefficient, and the HL band driving sound source e _HL to the HL band synthetic speech y _HL. Is output. The HL band synthesized speech y _HL is input to the power calculation unit 49 and the multiplication unit 51.

The power calculator 49 calculates the power P _HL for one frame of the HL band synthesized speech y _HL . The power P _HL is input to the gain calculator 50.

Gain calculating section 50, decoding relative gain G ^- _HL, using the power P _LH, and the power P _HL, calculates a gain g _HL represented by the following formula. The gain g _HL is input to the multiplication unit 51.

Multiplication unit 51 multiplies the gain g _HL to HL band synthesized speech y _HL, decoding HL band speech x ^- calculating the _HL. The decoded HL band speech x ^- _HL is input to the band synthesis filter 59.

The relative gain prediction unit 52 predicts and obtains the predicted relative gain G ⁻ _HH using the coefficient code c ₁ and the gain code c ₂ . The predicted relative gain G ⁻ _HH is input to the gain calculator 57.

Coefficient prediction unit 53 uses the LH band linear prediction coefficients a _LH and coefficient code c _1, HH band linear prediction coefficients a ^- obtained by predicting _HH. HH band linear prediction coefficients a ^- _HH is input to the synthesis filter 55.

The random number unit 54 generates a Gaussian random number and outputs a random number signal sequence e _HH having a one-frame length. The random number signal sequence e _HH is input to the synthesis filter 55.

Synthesis filter 55, HH band linear prediction coefficients a ^- a IIR filter to _HH filter coefficients, and outputs the HH band synthesized speech y _HH from the random number signal sequence e _HH. The HH band synthesized speech y _HH is input to the power calculation unit 56 and the multiplication unit 58.

The power calculator 56 calculates the power P _HH for one frame of the HH band synthesized voice y _HH . The power P _HH is input to the gain calculator 57.

Gain calculator 57, prediction relative gain G ^- with _HH, power P _LH, and the power P _HH, calculates a gain g _HH represented by the following formula. The gain g _HH is input to the multiplication unit 57.

Multiplier 58 multiplies the gain g _HH to HH band synthesized speech y _HH, decoding HH band speech x ^- calculating the _HH. The decoded HH band speech x ^- _HH is input to the band synthesis filter 59.

The band synthesizing filter 59 is a band synthesizing filter corresponding to the band dividing filter 31 _H of the high-frequency speech encoding unit 14 (that is, as an inverse transform), and includes a decoded HL band speech x ^- _HL and a decoded HH band speech x ^- _HH. Is used to generate and output a decoded high frequency speech x ⁻ _H. The sampling frequency of the decoded HL band speech x ^- _HL and the decoded HH band speech x ^- _HH is 4 kHz, and the sampling frequency of the decoded high band speech x ^- _H is 8 kHz.

The points of the speech coding apparatus and speech decoding apparatus according to the present invention will be described.

In the speech coder, wideband speech is band-divided into low-frequency speech and high-frequency speech, low-frequency speech is further converted into LL band signals and LH band signals, and high-frequency speech is further measured into HL band signals and HH. Band signal is divided into band signal and band signal. That is, the wideband voice is divided into four bands of the LL band, the LH band, the HL band, and the HH band.

In order to embed high-frequency audio information in the low-frequency code without degrading the quality of the decoded low-frequency audio, it is necessary to encode the high-frequency audio with as few bits as possible. Therefore, the spectrum envelope information and the power information of the HL band are encoded with a small number of bits so as not to deteriorate the quality of the decoded low frequency speech, and embedded in the low frequency code. In order to encode such information with a small number of bits, the correlation between parameters is used as much as possible. At this time, the information of the HH band is not sent.

The speech decoding apparatus extracts the HL band spectrum envelope information and the power information from the low band code, and generates the HL band signal and the HH band signal. Generally, a speech coding method using linear prediction requires spectrum envelope information, excitation information that drives a synthesis filter, and information that represents power, but a speech coding apparatus drives excitation information that drives a synthesis filter. Since it is not transmitted, it is necessary to generate pseudo sound source information for driving the synthesis filter from other information obtained by the speech decoding device. Therefore, it is considered that the linear prediction residual signal of the LH band is the same as the sound source information that drives the synthesis filter of the HL band, and the linear prediction residual signal of the LH band is driven to drive the synthesis filter of the HL band. Generate a signal. Further, regarding the HH band, since information is not sent from the voice encoding device, a signal in the HH band is pseudo-generated from the information of the LH band and the HL band obtained by the voice decoding device. Specifically, the spectrum envelope information of the HH band and the information indicating the power are predicted from the information of the LH band and the HL band by a statistical method, and the synthesis filter is driven by a Gaussian random number.

With the above method, high-frequency speech is represented by 8 bits per 10 milliseconds, and the speech decoding apparatus can reproduce wide-band speech with sufficiently good audibility. It should be noted that although the reproduced wideband sound has good audible quality, the SN ratio with the input sound, particularly the SN ratio in the high frequency range, is not high. Even if the SN ratio is not high, the auditory sense is good because the human predictive characteristics are linearly predicted when the spectral envelope and power are reproduced in a state where the power is close to the input voice in the high frequency range. That is, it is insensitive to the fine structure and phase of the spectrum. In addition, it is possible to perform highly reproducible encoding with a high-frequency spectrum envelope and bits with low power, and in particular, for the HH band, it is possible to reproduce the spectrum envelope and power without sending information. This is achieved by utilizing the fact that it has a high correlation with the low-frequency spectrum envelope and power.

[Modification A]
The above-described embodiment is an example of a wideband speech coding apparatus and a wideband speech decoding apparatus having complete compatibility with G.711, but as shown below, a wideband speech system having complete compatibility with G.726. It can be modified into a voice encoding device and a wideband decoding device.

FIG. 11 is a configuration example in which the voice encoding unit described in Patent Document 1 is used as the low-frequency voice encoding unit 13 of the embodiment. As shown in FIG. 11, the low-frequency audio encoding unit 13 includes an audio signal buffer 131, a bit analysis unit 132, a switch 133, a G.711 all code search unit 134, a G.711 even code search unit 135, and a G.711. An odd code search unit 136 and a voice code buffer 137 are provided. The audio signal buffer 131 divides the low-frequency audio x _L at regular time intervals called a frame and outputs it. The bit analysis unit 132 decomposes the high-frequency code c _H into a bit string for each frame, and controls the switch 133 according to the value of each bit. Among the sample points of the frame, at the sample point corresponding to each bit of the predetermined bit string, if the value of the bit is 0, the G.711 even code search unit 135 corresponds to the even code of G.711. A code having a quantized value close to the input voice sample value is searched from the 128 codes and the G.711 voice code whose LSB represents the value of the bit is output according to the search result. If the value of the bit is 1 at the sample point corresponding to each bit, the G.711 odd code search unit 136 converts the 128 code corresponding to the G.711 odd code into the input audio sample value. A code having a close coded value is searched for, and a G.711 voice code whose LSB represents the value of the bit is output according to the search result. Except for the sample points corresponding to the bits, the G.711 all-code search unit 134 searches for a code having a quantized value close to the input voice sample value from all 256 G.711 codes and follows the search result. Outputs G.711 voice code. The voice code buffer 137 collects the G.711 voice code for one frame and outputs it as a voice code (also a low frequency code) c _L.

FIG. 12 is a modified example of the low-frequency speech coding unit for realizing wideband speech coding that is completely compatible with G.726. The low-frequency audio encoding unit 13A of the modified example includes an audio signal buffer 131, a bit analysis unit 132, an audio code buffer 137, and further includes a G.726 compatible encoder 140, like the low-frequency audio encoding unit 13. .. The G.726 compatible encoder 140 includes a switch 143, a G.726 all code search unit 144, a G.726 even code search unit 145, a G.726 odd code search unit 146, an adaptive prediction unit 148, and a difference signal calculation unit 149. Prepare G.726 is called an Adaptive Differential Pulse Code Modulation (ADPCM) method. In G.726, the adaptive prediction unit 148 predicts the current voice sample value using the quantized value of the past voice sample, and the difference signal calculation unit 149 determines that the current voice sample sent from the voice signal buffer 131. By calculating a difference signal between the value and the predicted value of the voice sample sent from the adaptive prediction unit 148, the difference is quantized with 2 to 5 bits per sample. Generally, a mode in which quantization is performed with 4 bits per sample (also called 32 kbit/s mode) is often used in digital cordless telephones, etc. Therefore, in the following description, a mode with 4 bits per sample will be described as an example. To do.

Among the sample points of the frame, at the sample point corresponding to each bit of the predetermined bit string, if the value of the bit is 0, the G.726 even code search unit 145 corresponds to the even code of G.726. A code having a quantized value close to that of the difference signal is searched from among the 8 codes, and a G.726 voice code whose LSB represents the value of the bit is output according to the search result. At the sample point corresponding to each bit, if the value of the bit is 1, G.726 odd code search unit 146, from the 8 codes corresponding to the odd code of G.726, to the difference signal of the quantization value It searches for a close code and outputs the G.726 voice code whose LSB represents the value of the bit according to the search result. Except for the sample points corresponding to the respective bits, the G.726 all-code search unit 144 searches for all 16 codes of G.726 for a code having a quantized value close to the difference signal, and according to the search result, G.726. The voice code of is output. The G.726 voice code is sent to the voice code buffer 137 and also to the adaptive prediction unit 148.

Details of G.726 other than the above are described in ITU-T G.726. In addition, since G.726 has a small number of quantization bits per sample, if a low-frequency code with a high-frequency code embedded is generated, and the decoded low-frequency speech obtained by decoding the low-frequency code is noisy for G.711. There are more problems than. As a method of reducing this noise sensation, as shown in FIG. 13, a “method of effectively reducing the harshness of quantization noise” described in Reference 2 can be combined.
[Reference 2] Japanese Patent No. 5014493

The method of dividing the low-frequency speech code to be transmitted by dividing the entire code into half codes such as even-numbered code and odd-numbered code and switching the search range, and embedding the high-frequency code as a bit string of 1 or 0 in LSB or MSB is G.711. Not limited to G.726, it can be widely applied to existing speech coding methods, and by replacing G.726 in the above configuration example with another existing speech coding method, it is completely compatible with the existing speech coding method. It is possible to realize a wideband speech coding apparatus and a wideband speech decoding apparatus having the above.

In this modified example, when the bit of the high frequency code to be embedded is 0, the even code is searched, and when the bit of the high frequency code is 1, the odd code is searched. However, when 1 is set, the correspondence may be such that an even code is searched.

[Modification B]
FIG. 14 is a configuration example of the low-frequency speech decoding unit 16 in the speech encoding device of the embodiment and the low-frequency speech decoding unit 22 in the speech decoding device of the embodiment. An existing G.711 decoder can be simply used for the low-band speech decoding unit of the wideband speech coding apparatus and the wideband speech decoding apparatus which are completely compatible with G.711. FIG. 15 is a modification of the low-frequency speech coding unit of the wideband speech coding apparatus and the wideband decoding apparatus for realizing wideband speech coding completely compatible with G.726. The .711 decoder 162 is replaced with the existing G.726 decoder 163.

Since G.726 has a small number of quantization bits per sample, there is a problem that the sense of noise in decoded low-frequency speech increases more than in G.711. As a method of reducing this noise sensation, as shown in FIG. 16, a post filter 164 widely used as a narrow band speech encoding method can be used after the G.726 decoder 163. The post filter is widely used as a method for improving the quality of decoded speech in so-called "CELP system" such as ITU-T G.729 and PSI-CELP used in Japanese mobile phones. Therefore, detailed description is omitted here.

[Modification C]
FIG. 17 is a modification of the high frequency audio decoding unit 25 shown in FIG. The high-frequency speech decoding unit 25A of the modified example is different from the high-frequency speech decoding unit 25 in that it includes a designated band signal extraction unit 61, a linear prediction unit 62, and an inverse filter 63. In the high frequency speech decoding unit 25, the HL band signal is generated by driving the HL band synthesis filter with the LH band linear prediction residual signal. Here, the linear prediction residual signal of the LH band is a linear prediction residual signal of the 2 to 4 kHz band, but in the high frequency speech decoding unit 25A, instead of the 2 to 4 kHz band,
F to (F+2)kHz band However, 0≦F≦2
Drive the HL band synthesis filter with the linear prediction residual signal of Here, F is a predetermined constant. In the configuration of FIG. 17, the designated band signal extraction unit 61 extracts the F to (F+2) kHz band component of the decoded low frequency speech x ^- _L to obtain the designated band signal x ^- _LF of 4 kHz sampling. Linear prediction unit 62 is designated band signals x ^- by applying linear prediction analysis to _LF, and outputs a p-th order specified band linear prediction coefficients a _LF. The designated band linear prediction coefficient a _LF is input to the inverse filter 63. Inverse filter 63 is a FIR filter that the specified band linear prediction coefficients a _LF filter coefficients, designated subband signal x ^- asked for from the _LF band linear prediction residual e _LF, and sends to the replication section 45. The duplication unit 45 duplicates the designated band linear prediction residual e _LF and outputs the HL band driving sound source e _HL .

The high-frequency speech decoding unit 25 shown in FIG. 9 corresponds to F=2, but a case where F<2 is preferable will be described. FIG. 18(A) is an example of the frequency spectrum of the low-frequency speech x _L , and FIG. 18(B) is the frequency spectrum of the decoded low-frequency speech x ^- _L corresponding to the example of FIG. 18(A). Comparing FIG. 18(A) and FIG. 18(B), noise is superimposed over the entire band in FIG. 18(B) due to coding distortion, and the SN ratio in the 2 to 4 kHz band is, for example, in the 1 to 3 kHz band. It can be said that it is worse than the SN ratio. In such a case, rather than driving the synthesis filter in the HL band with the linear prediction residual signal in the 2 to 4 kHz band,
F to (F+2)kHz band, 0≤F<2
Driving the synthesis filter of the HL band with the linear prediction residual signal of 1 can reduce the noise feeling of the generated signal of the HL band.

[Modification D]
FIG. 19 is a configuration example of the designated band signal extraction unit 61. The designated band signal extraction unit 61 includes a Fourier transform unit 161, a frequency component cyclic unit 162, a Fourier inverse transform unit 163, and a band division filter (also referred to as a designated band division unit) 164. The Fourier transform unit 161 performs a Fourier transform on the decoded low frequency speech x ^- _L to convert it into the frequency domain decoded low frequency speech X ^- _L . The frequency component circulating unit 162 determines that the fkHz component of the frequency domain decoded low frequency speech X ^- _L is
When f≦2+F f'=f+2-F
When f>2+F f'=f-2-F
Compute the low-frequency speech X ⁻ _LS in the frequency domain cyclic decoding with the f'kHz component as Inverse Fourier transform unit 163, frequency domain cyclic decoded low-band speech X ^- the _LS and inverse Fourier transform, cyclic decoded low-band speech x ^- calculating the _LS. The band division filter 164, similar to the band division filter 41 of the high frequency audio decoding unit 25 shown in FIG. 9, inputs the cyclically decoded low frequency audio x ^- _LS into high frequency audio and low frequency audio each having a sampling frequency of 4 kHz. Divide into voice and output high frequency voice as specified band signal x ^- _LF .

Although the embodiments of the present invention have been described above, the specific configuration is not limited to these embodiments, and even if the design is appropriately changed without departing from the gist of the present invention, Needless to say, it is included in the present invention. The various kinds of processing described in the embodiments may be executed not only in time series according to the order described, but also in parallel or individually according to the processing capability of the device that executes the processing or the need.

[Program, recording medium]
When various processing functions in each device described in the above embodiment are realized by a computer, processing contents of functions that each device should have are described by a program. By executing this program on a computer, various processing functions of the above-described devices are realized on the computer.

The program describing the processing contents can be recorded in a computer-readable recording medium. The computer-readable recording medium may be, for example, a magnetic recording device, an optical disc, a magneto-optical recording medium, a semiconductor memory, or the like.

The distribution of this program is performed by, for example, selling, transferring, or lending a portable recording medium such as a DVD or a CD-ROM in which the program is recorded. Further, the program may be stored in a storage device of the server computer and transferred from the server computer to another computer via a network to distribute the program.

A computer that executes such a program first stores, for example, the program recorded in a portable recording medium or the program transferred from the server computer in its own storage device. Then, when executing the processing, this computer reads the program stored in its own recording medium and executes the processing according to the read program. As another execution form of this program, a computer may directly read the program from a portable recording medium and execute processing according to the program, and the program is transferred from the server computer to this computer. Each time, the processing according to the received program may be sequentially executed. In addition, a configuration in which the above-described processing is executed by a so-called ASP (Application Service Provider) type service that realizes a processing function only by executing the execution instruction and the result acquisition without transferring the program from the server computer to the computer May be Note that the program in this embodiment includes information that is used for processing by an electronic computer and that conforms to the program (such as data that is not a direct command to a computer but has the property of defining computer processing).

Further, in this embodiment, the present apparatus is configured by executing a predetermined program on the computer, but at least a part of the processing content may be implemented by hardware.

11 Input Buffer 12 Band Division Filter 13 Low-Band Speech Encoding Unit 14 High-Band Speech Encoding Unit 15 Delay Unit 16 Low-Band Speech Decoding Unit 17 Code Sending Unit 21 Code Receiving Unit 22 Low-Band Speech Decoding Unit 23 High-Band Code Extracting Unit 24 Delay Unit 25 High-Band Speech Decoding Unit 26 Band Synthesis Filter

Claims (5)

A low-pass code obtained by encoding the low-pass voice of the input voice, wherein a low-pass code obtained by embedding a high-pass code obtained by encoding the high-pass voice of the input voice based on the low-pass voice obtained by decoding the low-pass code. A code receiving unit for receiving as a voice code,
A low-frequency audio decoding unit that decodes the audio code to generate a decoded low-frequency audio,
A high-frequency code extraction unit that extracts the high-frequency code embedded in the voice code,
A high-frequency speech decoding unit that decodes the high-frequency code based on the decoded low-frequency speech to generate decoded high-frequency speech,
A band synthesizing unit for synthesizing the decoded low-frequency speech and the decoded high-frequency speech to output decoded speech,
Including,
The above high frequency speech decoding unit,
A band division unit that divides the decoded low-frequency voice into an LH band voice and an LL band voice,
A code separation unit that separates the voice code into a gain code and a coefficient code,
A coefficient decoding unit that decodes the coefficient code using the linear prediction coefficient of the LH band speech to obtain an HL band decoding linear prediction coefficient;
A relative gain decoding unit that decodes the gain code using the linear prediction coefficient of the LH band speech and the coefficient code to obtain a decoded relative gain;
A coefficient predicting unit that predicts an HH band linear prediction coefficient using the linear prediction coefficient of the LH band speech and the coefficient code;
A relative gain prediction unit that predicts and calculates a predicted relative gain using the gain code and the coefficient code,
A designated band signal extraction unit for obtaining a designated band signal by extracting F to (F+2) kilohertz band component of the decoded low frequency speech, where F is a constant of 0≦F≦2, which is predetermined.
A duplication unit that duplicates a designated band linear prediction residual obtained from the designated band signal using the linear prediction coefficient of the designated band signal as a filter coefficient to obtain an HL band driven sound source,
The HH band speech is multiplied by a gain calculated from the predicted relative gain based on the ratio of the power of the HH band speech obtained from Gaussian random numbers using the HH band linear prediction coefficient as a filter coefficient and the power of the LH band speech. An HH band multiplication unit for generating a decoded HH band voice,
The gain calculated from the decoding relative gain based on the ratio of the power of the HL band synthesized speech obtained from the HL band driven sound source with the HL band decoding linear prediction coefficient as a filter coefficient is used as the HL band. An HL band multiplication unit that multiplies the synthesized voice to generate a decoded HL band voice,
A band synthesizing unit for synthesizing the decoded HH band speech and the decoded HL band speech and outputting the decoded high band speech;
Including
Speech decoding device. The voice decoding device according to claim 1 , wherein
The designated band signal extraction unit,
A Fourier transform unit for converting the decoded low-frequency speech into the frequency domain to generate frequency-domain decoded low-frequency speech,
The f kilohertz component of the frequency domain decoded low-frequency speech is f'=f+2-F when f≤2+F, and f'=f-2-F when f>2+F. 'A frequency component cyclic unit for calculating low-frequency speech, which is composed of a frequency domain cyclic decoding composed of kilohertz components,
A Fourier inverse transform unit for converting the frequency domain cyclic decoding low-frequency speech into the time domain to generate cyclic decoding low-frequency speech,
A specified band division unit that divides the cyclic decoding low-frequency audio into high-frequency audio and low-frequency audio, and outputs the high-frequency audio as the specified band signal;
Including
Speech decoding device. The code receiving unit is a low band code obtained by coding the low band voice of the input voice, and outputs a high band code obtained by coding the high band voice of the input voice based on the low band voice obtained by decoding the low band code. the low-frequency encoding embedded received as the speech code,
The low-frequency audio decoding unit decodes the audio code to generate a decoded low-frequency audio,
High band sign extracting unit extracts the high frequency code embedded in the speech code,
A high-frequency audio decoding unit decodes the high-frequency code based on the decoded low-frequency audio to generate a decoded high-frequency audio,
The band synthesizing unit synthesizes the decoded low-frequency speech and the decoded high-frequency speech to output decoded speech ,
The above high frequency speech decoding unit,
Band-dividing the decoded low-frequency voice into LH band voice and LL band voice,
The voice code is separated into a gain code and a coefficient code,
Decoding the coefficient code using the linear prediction coefficient of the LH band speech to obtain an HL band decoding linear prediction coefficient,
Decoding the gain code using the linear prediction coefficient of the LH band speech and the coefficient code to obtain a decoded relative gain,
The HH band linear prediction coefficient is predicted and obtained using the linear prediction coefficient of the LH band speech and the coefficient code,
Predicting and obtaining a predicted relative gain using the gain code and the coefficient code,
Assuming that F is a predetermined constant of 0≦F≦2, F−(F+2) kilohertz band component of the decoded low frequency speech is extracted to obtain a designated band signal,
Using the linear prediction coefficient of the designated band signal as a filter coefficient, the designated band linear prediction residual obtained from the designated band signal is duplicated to obtain an HL band driven sound source,
The HH band speech is multiplied by a gain calculated from the predicted relative gain based on the ratio of the power of the HH band speech obtained from Gaussian random numbers using the HH band linear prediction coefficient as a filter coefficient and the power of the LH band speech. Generate decoded HH band audio,
The gain calculated from the decoding relative gain based on the ratio of the power of the HL band synthesized speech obtained from the HL band driven sound source with the HL band decoding linear prediction coefficient as a filter coefficient is used as the HL band. Generate a decoded HL band voice by multiplying the synthesized voice,
Synthesizing the decoded HH band sound and the decoded HL band sound, and outputting the decoded high band sound,
Speech decoding method. Program for causing a computer to function as the speech decoding apparatus according to claim 1 or 2. A computer-readable recording medium in which a program for causing a computer to function as the audio decoding device according to claim 1 or 2 is recorded.

Images