JP2002196795A - Speech decoder, and speech coding and decoding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a speech decoder and a speech coding and decoding device, with which more improved decoded speech quality can be realized in the frame where the frame loss is detected. SOLUTION: When the preceding frame is the loss frame and the coded data of the loss frame is received behind time by the receiver buffer 101 in the present time, the preceding frame coded data and the correctly received present frame coded data of the present time are inputted into a loss frame decoding part 105. In the loss frame decoding part 105, the speech is decoded using the inputted preceding frame and the present frame coded data one by one, and the decoded speech signal to the present frame coded data is outputted. Moreover, state data required of the decoding is updated and held in a state data holding part 103. In a changeover part 106, the changeover according to the frame loss information is performed, and the decoded speech signal of the frame is outputted.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a speech decoding apparatus used for applications such as a mobile communication system or a wired communication system which encodes and transmits a speech signal, and more particularly to a speech decoding apparatus which decodes coded data when coded data is lost. The present invention relates to an audio decoding device and an audio encoding / decoding device having a frame compensation function for suppressing deterioration of audio quality.

[0002]

2. Description of the Related Art Conventionally, in the field of digital mobile communication and wire communication, a voice coding device for compressing voice information and coding at a low bit rate has been used for effective use of radio waves and wire lines. . In such a communication system, especially in a system that performs transmission on a packet-by-packet basis, a packet received on the receiving side (speech decoding side) may be lost or may arrive with a delay in time to perform speech decoding. In some cases, necessary data cannot be obtained (frame loss), and at that time, frame compensation is performed to suppress deterioration of decoded voice quality.

[0003] As such a conventional technique, ITU-T
Recommendation G. 729 ("Coding of speech at 8kbit / s using
conjugate-structure algebraic-code-excited linear
-prediction (CS-ACELP) "), there is an error compensation method described in the CS-ACELP encoding method.

FIG. 11 is a block diagram showing a configuration of a speech decoding apparatus having a frame compensation function of the CS-ACELP coding system. In this figure, audio decoding is 10 ms.
Is performed in frame units, and the presence / absence of frame loss is notified to the speech decoder in frame units.

[0005] First, a frame in which no frame loss is detected is separated by a data separation unit 1101 into parameters necessary for decoding. An adaptive excitation is generated by adaptive excitation codebook section 1103 and a fixed excitation is generated by fixed excitation codebook section 1104 using the lag parameters decoded by lag parameter decoding section 1102, and decoded by gain parameter decoding section 1105. A driving sound source is generated by the multiplication by the multiplication unit 1106 and the addition by the addition unit 1107 using the gain.
A decoded speech is generated via the LPC synthesis filter 1109 and the post filter 1110 using the LPC parameters decoded by the C parameter decoding unit 1108.

On the other hand, for a frame in which a frame loss is detected, an adaptive excitation is generated using a lag parameter of a previous frame in which no error is detected as a lag parameter. Generate a fixed excitation by giving a random fixed excitation code to the
A driving sound source is generated using a value obtained by attenuating the adaptive sound source gain and the fixed sound source gain of the previous frame as a gain parameter, and the LPC parameter of the previous frame LPC is used as an LPC parameter.
LPC synthesis and post-filter processing are performed using the parameters to obtain decoded speech. Thus, the frame compensation processing is performed when the frame is lost.

[0007]

However, the above conventional speech decoding apparatus has the following problems.
That is, since the frame compensation processing is performed using the past parameters from the lost frame, the past state data (the adaptive codebook, the LSP parameter, the gain parameter, etc.) necessary for the decoding in the next normal frame after the lost frame are decoded. MA prediction filter state) may differ greatly from the original value (when there is no frame loss), thereby deteriorating the quality of decoded speech.

If the lost frame includes a transitional section such as a rising edge of the speech, the section is not reproduced, and the quality of the decoded speech deteriorates. As described above, there is a limit to the improvement of the decoded voice quality degradation at the time of frame compensation.

SUMMARY OF THE INVENTION The present invention has been made in view of the above points, and an object of the present invention is to provide a speech decoding apparatus and a speech encoding / decoding apparatus capable of realizing improved decoded speech quality in a frame in which a frame loss is detected. The purpose is to provide.

[0010]

SUMMARY OF THE INVENTION A speech decoding apparatus according to the present invention uses a reception buffer for receiving and storing coded data in frame units, and using the coded data when the coded data of the current frame is correctly received. Speech decoding means for performing normal speech decoding, loss frame compensation means for performing loss frame compensation when encoded data of the current frame cannot be received due to loss, and the reception buffer delays the encoded data in time. And a lossy frame decoding unit for performing speech decoding using the encoded data and the encoded data of the current frame when receiving, and updating state data necessary for decoding by the speech decoding unit. .

According to this configuration, when encoded data of a lost frame is received with a time delay, after outputting a decoded speech by frame compensation, the encoded data is used to correspond to a past lost frame. By performing audio decoding of the section and updating the state data necessary for audio decoding, and then performing audio decoding of the current normal frame, decoding using state data closer to the original value becomes possible. It is possible to suppress the deterioration of the voice quality.

A speech decoding apparatus according to the present invention comprises a receiving buffer for receiving and storing coded data in frame units, and a normal speech decoding apparatus using coded data when coded data of the current frame is correctly received. Voice decoding means for performing loss-frame compensation when the encoded data of the current frame cannot be received due to loss, and when the reception buffer receives the encoded data with a time delay. Loss frame decoding means for performing speech decoding using the encoded data and the encoded data of the current frame and updating state data necessary for decoding by the speech decoding means, and a current frame decoded by the lost frame decoding means. Windowing addition means for performing windowing addition using a decoded audio signal and a decoded audio signal of the current frame decoded by the audio decoding means. Employs a configuration having a.

According to this configuration, when the encoded data of the lost frame is received with a time delay, after outputting the decoded speech by frame compensation, the section corresponding to the past lost frame is used by using the encoded data. To update the state data by performing audio decoding, and then perform the audio decoding after updating the state data by the frame compensation processing of the lost frame and the decoded audio signal obtained by performing the audio decoding of the current normal frame. By performing windowed addition using the obtained decoded audio signal of the current normal frame, smoother decoding can be performed.

A speech encoding / decoding device according to the present invention comprises: a speech mode determining means for determining a speech mode representing characteristics of an input speech on a frame basis; Audio encoding means for outputting encoded data together with the audio mode information obtained by the judging means, a reception buffer for receiving and storing encoded data in frame units, and a code when the encoded data of the current frame is correctly received. Speech decoding means for performing normal speech decoding using encoded data, loss frame compensation means for performing lost frame compensation when encoded data of the current frame cannot be received due to loss, and the reception buffer converts encoded data. When receiving with a time delay, speech decoding is performed using the encoded data and the encoded data of the current frame. Loss frame decoding means for updating state data necessary for decoding by the decoding means, time axis compression means for time axis compressing the loss frame decoded by the loss frame decoding means and the decoded audio signal of the current frame, According to the audio mode obtained by the mode determination means, either the decoded audio signal obtained by the time axis compression means or the decoded audio signal of the current frame obtained by the lost frame decoding means is selected, and the selected decoding is performed. A window adding means for performing window adding using the audio signal and the decoded audio signal of the current frame decoded by the audio decoding means is adopted.

According to this configuration, when the encoded data of the lost frame is received with a time delay, after outputting the decoded speech by frame compensation, the encoded data and the encoded data of the current normal frame are used. If the mode information determined by the encoding side indicates a transient section, the obtained decoded signal is compressed in the time domain by performing speech decoding of the section between the past lost frame and the current normal frame. As a result, it is possible to suppress the deterioration of the decoded voice quality due to the frame loss in the transient section.

A base station apparatus according to the present invention employs a configuration including any one of the above-described speech decoding device, speech decoding device, and speech encoding / decoding device.

The communication terminal device of the present invention employs a configuration including any of the above-described speech decoding device, speech decoding device, or speech encoding / decoding device.

A speech decoding method according to the present invention comprises a receiving buffering step of receiving and storing encoded data in units of frames, and a normal speech decoding using encoded data when the encoded data of the current frame is correctly received. A voice decoding step of performing, and a lost frame compensation step of performing lost frame compensation when the encoded data of the current frame cannot be received due to loss, and when the reception buffer receives the encoded data with a time delay. A lost frame decoding step of performing speech decoding using the encoded data and the encoded data of the current frame and updating state data necessary for decoding in the speech decoding step.

According to this method, when encoded data of a lost frame is received with a time delay, a decoded speech is output by frame compensation, and a section corresponding to a past lost frame is used by using the encoded data. Then, the state data necessary for the sound decoding is updated, and then, by performing the sound decoding of the current normal frame, decoding using the state data closer to the original value becomes possible. It is possible to suppress the deterioration of the voice quality.

A recording medium according to the present invention is a computer-readable recording medium storing an audio decoding program, the audio decoding program receiving and storing encoded data in frame units, A procedure for performing normal speech decoding using the encoded data when the encoded data is correctly received; a procedure for performing loss frame compensation when the current frame encoded data cannot be received due to loss; And a procedure for performing speech decoding using the encoded data and the current frame encoded data when receiving the data with a time delay and updating state data necessary for decoding, and a program for executing the program. It is readable.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The essence of the present invention is that when encoded data of a lost frame is received with a time delay, a decoded speech is output by frame compensation, and the past loss is output using the encoded data. That is, speech decoding is performed for a section corresponding to a frame to update state data necessary for speech decoding, and then speech decoding for a current normal frame is performed.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(Embodiment 1) FIG. 1 is a block diagram showing a configuration of a speech decoding apparatus according to Embodiment 1 of the present invention. In this figure, the speech decoding device of the present embodiment
It comprises a reception buffer 101, a speech decoding unit 102, a state data holding unit 103, a lost frame compensation unit 104, a lost frame decoding unit 105, and a switching unit 106.

The receiving buffer 101 receives encoded data that has been encoded in frame units on the transmission side and transmitted in packet units, outputs the received encoded data, and converts the received packet-based data in frame units. And outputs information (frame loss information) as to whether or not a frame loss has occurred.

When the current frame coded data is correctly received, the voice decoding section 102 performs normal voice decoding using the coded data. The state data holding unit 103 holds state data used by the audio decoding unit 102 and the lost frame decoding unit 105. This state data depends on the specific configuration of the audio decoding unit 102.
In the case of a P-type decoder, there are an adaptive codebook, an LSP / gain parameter MA prediction filter state, and the like.

The lost frame compensator 104 performs lost frame compensation when the current frame coded data cannot be received due to loss. When the reception buffer 101 receives the encoded data with a time delay, the lost frame decoding unit 105 performs the audio decoding using the encoded data and the current frame encoded data, and performs the decoding in the audio decoding unit 102. Update the necessary status data. Switching unit 106
Performs switching between the output of the audio decoding unit 102, the lost frame compensation unit 104, and the lost frame decoding unit 105 according to the frame loss information.

The speech decoding apparatus having such a configuration performs different operations depending on the frame loss information for the current frame. First, when the encoded data for the current frame is correctly received without loss or delay, the audio decoding unit 102 performs normal audio decoding using the encoded data, and outputs the decoded audio signal. At this time, state data necessary for decoding is updated and held in the state data holding unit 103.

Next, if the encoded data for the current frame cannot be received correctly due to a lost frame, ie, loss or delay, the lost frame compensator 104 performs lost frame compensation, and the resulting decoded speech signal Is output. This compensation processing is optional, and compensation processing using past encoding parameters and decoded data according to a conventional method may be performed.

Next, the previous frame is a lost frame,
If the coded data of the lost frame is received with a time delay at the present time, the coded data of the previous frame and the coded data of the current frame correctly received at the current time are input to the lost frame decoding unit 105. .

The lost frame decoding unit 105 performs speech decoding by sequentially using the input encoded data of the previous frame and the current frame ((a) and (b) of FIG. 2), and performs the decoding on the encoded data of the current frame. The decoded audio signal ((b) in FIG. 2) is output. The state data necessary for decoding is updated and held in the state data holding unit 103. The speech decoding process in the lost frame decoding unit 105 may be basically the same process as the speech decoding unit 102. Switching unit 10
In 6, the switching according to the above-mentioned frame loss information is performed, and the decoded audio signal of the frame is output.

Next, referring to the flowchart of FIG.
The operation of the speech decoding device according to the present embodiment will be described.

First, in step 101, the transmitting side receives and stores encoded data that has been encoded in frames and transmitted in units of packets, and further converts the accumulated data in units of packets into encoded data in units of frames. The frame is decomposed and output together with information indicating whether a frame loss has occurred (frame loss information).

Next, in step 102, frame loss information for the current frame is determined. In this determination, if it is determined that the coded data for the current frame has been correctly received without loss or delay, the process proceeds to step 103, where normal voice decoding is performed using the coded data, and output as a decoded voice signal. On the other hand, if it is determined that the coded data for the current frame cannot be correctly received due to a lost frame, that is, a loss or delay, the process proceeds to step 104, where lost frame compensation is performed, and a decoded speech signal obtained thereby is output. .

The previous frame is a lost frame,
If it is determined that the encoded data of the lost frame has been received with a time delay at the current time, step 105
To perform speech decoding by sequentially using the encoded data of the previous frame and the encoded data of the current frame correctly received at the present time, and output a decoded audio signal for the encoded data of the current frame. After outputting the decoded audio signal, the state data necessary for decoding is updated in step 106. After performing the processing of step 103, step 104, or step 106, the decoded audio signal is output as an output signal in step 107.

As described above, according to the speech decoding apparatus according to the present embodiment, when the encoded data of the lost frame is received with a time delay, the decoded speech is output by frame compensation, and then the decoded speech is output. Speech decoding of the section corresponding to the past lost frame is performed using the encoded data to update the state data necessary for speech decoding, and then the speech decoding of the current normal frame is performed. Can be decoded using state data close to the value of, and it is possible to suppress deterioration of the decoded voice quality.

(Embodiment 2) FIG. 4 is a block diagram showing a configuration of a speech decoding apparatus according to Embodiment 2 of the present invention. In this figure, the same parts as those in FIG. 1 described above are denoted by the same reference numerals.

In this figure, the speech decoding apparatus according to the present embodiment includes a window adding section 107 in addition to the same configuration as the speech decoding apparatus according to the first embodiment. In the receiving buffer 101, the data is encoded in frame units on the transmission side,
Receives encoded data transmitted in packet units. Then, the packet data is decomposed into coded data in frame units and output together with information on whether or not a frame loss has occurred (frame loss information). The subsequent operation differs depending on the frame loss information for the frame at the current time.

First, when the encoded data for the current frame is correctly received without any loss or delay, the audio decoding section 102 performs normal audio decoding using the encoded data and outputs the decoded audio signal as a decoded audio signal. I do. At this time, state data necessary for decoding is updated and held in the state data holding unit 103.

Next, if the coded data for the current frame cannot be received correctly due to a lost frame, ie, loss or delay, loss frame compensation is performed in the loss frame compensator 104, and the resulting decoded speech signal is Output. This compensation process is optional,
Compensation processing using past encoding parameters and decoded data according to a conventional method may be performed.

Next, the previous frame is a lost frame,
If the coded data of the lost frame is received with a time delay at the current time, the previous frame coded data and the correctly received current frame coded data of the current time are sent to the lost frame decoding unit 105. input.

The lost frame decoding unit 105 performs speech decoding by sequentially using the input previous frame and the current frame encoded data (FIGS. 5A and 5B), and decodes the decoded speech for the current frame encoded data. A signal ((b) in FIG. 5) is output. Further, the state data necessary for decoding is updated and held in the state data holding unit 103. The speech decoding process in the lost frame decoding unit 105 may be basically the same as the speech decoding unit 102. Also, the audio decoding unit 102 performs audio decoding using the current frame coded data correctly received at the present time, and outputs a decoded audio signal ((c) in FIG. 5).

The window adding section 107 uses the output signal from the lost frame decoding section 105 (FIG. 5B) and the output signal from the speech decoding section 102 (FIG. 5C) to obtain the frame. At the start end, the output of the audio decoding unit 102 is subjected to windowing addition such that the output of the lost frame decoding unit 105 becomes dominant at the end end, and the resultant is output as a decoded audio signal. Then, the switching unit 106 performs switching according to the above-described frame loss information, and outputs a decoded audio signal of the frame.

Next, referring to the flowchart of FIG.
The operation of the speech decoding device according to the present embodiment will be described.

First, in step 201, the transmitting side receives encoded data encoded in frames and transmitted in packet units, and further decomposes the received packet-based data into encoded data in frame units. Information indicating whether a frame loss has occurred (frame loss information)
Output with Next, in step 202, frame loss information for the current time frame is determined.

In this determination, if it is determined that the coded data for the current frame has been correctly received without any loss or delay, the process proceeds to step 203, where normal voice decoding is performed using the coded data and output as a decoded voice signal. I do. On the other hand, if it is determined that the encoded data for the current frame cannot be correctly received due to a lost frame, that is, a loss or delay, the process proceeds to step 204,
Loss frame compensation is performed, and the resulting decoded speech signal is output.

The previous frame is a lost frame,
If it is determined that the encoded data of the lost frame has been received with a time delay at the current time, step 205
To perform speech decoding by sequentially using the encoded data of the previous frame and the encoded data of the current frame correctly received at the present time, and output a decoded audio signal for the encoded data of the current frame. Then, in step 206, state data necessary for decoding is updated.

After updating the data, step 207
, Normal speech decoding is performed using the coded data of the current frame correctly received at the present time, and a decoded speech signal is output. Next, in step 208, window addition is performed using the output signal in step 204 and the output signal in step 205, and the resultant signal is output as a decoded speech signal. After performing the processing of step 203, step 204, or step 208, the decoded audio signal is output as an output signal in step 209.

As described above, according to the speech decoding apparatus according to the present embodiment, when encoded data of a lost frame is received with a time delay, a decoded speech is output by frame compensation, and The state data is updated by performing speech decoding of a section corresponding to a past lost frame using the encoded data at the time, and thereafter, a decoded speech signal obtained by performing speech decoding of a current normal frame and a lost frame of a lost frame are obtained. Windowed addition is performed using the decoded audio signal of the current normal frame obtained by performing audio decoding after updating the state data by frame compensation processing, so that smoother decoding can be performed. Becomes

(Embodiment 3) FIG. 7 is a block diagram showing a configuration of a speech encoding / decoding apparatus according to Embodiment 3 of the present invention. Note that, in this figure, the same reference numerals are given to portions common to FIG. 1 or FIG. 4 described above.

In this figure, the encoding side of the speech encoding / decoding apparatus of the present embodiment is
, An audio encoding unit 111, and a multiplexing unit 112, and the decoding side includes a reception buffer 101, a demultiplexing unit 1
13, the audio decoding unit 102, and the state data holding unit 103
A lost frame compensating unit 104, a lost frame decoding unit 105, a time axis compressing unit 114, a switching unit 115,
A switching unit 106 and a window adding unit 107 are provided.

The voice coding unit 111 performs voice coding on input voice in frame units and outputs coded data. The audio mode determination unit 110 determines the audio mode of the input audio in frame units. Here, the audio mode indicates a characteristic that the input audio in a frame unit is a stationary section or a transient section such as a rising section. The multiplexing unit 112 multiplexes the encoded data output from the audio encoding unit 111 and the audio mode obtained by the audio mode determination unit 110, and outputs the multiplexed data as encoded data including audio mode information.

The receiving buffer 101 receives encoded data that has been encoded in frame units on the transmitting side and transmitted in packet units. Further, the received data is decomposed into encoded data in frame units and output together with information (frame loss information) indicating whether or not a frame loss has occurred. The separation unit 113 separates the encoded data including the audio mode information into audio mode information and encoded audio data. The subsequent operation differs depending on the frame loss information for the current frame.

If the encoded data for the current frame is correctly received without any loss or delay, the speech decoding section 102 performs normal speech decoding using the encoded data and outputs the decoded speech signal. . At this time, state data necessary for decoding is updated and held in the state data holding unit 103.

Here, the state data is the speech decoding unit 102
However, in the case of a CELP type decoder, there are an adaptive codebook, an LSP / gain parameter MA prediction filter state, and the like.

On the other hand, if the coded data for the current frame cannot be correctly received due to a lost frame, that is, a loss or delay, the lost frame compensator 104 performs loss frame compensation, and obtains the decoded speech signal obtained thereby. Output. This compensation processing is optional, and compensation processing using past encoding parameters and decoded data according to a conventional method may be performed.

The previous frame is a lost frame,
If the coded data of the lost frame is received with a time delay at the present time, the coded data of the previous frame and the coded data of the current frame correctly received at the current time are input to the lost frame decoding unit 105. .

The lost frame decoding unit 105 performs speech decoding by sequentially using the input encoded data of the previous frame and the current frame ((a) and (b) of FIG. 8), and decodes the encoded data of the current frame. The audio signal ((b) in FIG. 8) is supplied to the switching unit 115, and the decoded audio signal ((a) in FIG. 8) for the encoded data of the previous frame and the current frame.
And (b)) to the time axis compression unit 114. Also, the state data necessary for decryption is updated and the state data holding unit 10
3 is held. The speech decoding process in the lost frame decoding unit 105 may be basically the same process as the speech decoding unit 102.

The time axis compression section 114 decodes the decoded audio signal (see FIG.
(A) and (b)) are subjected to time axis compression so as to be a signal of one frame section, and the compressed signal ((d) in FIG. 8)
Is output to the switching unit 115. Also, the audio decoding unit 102 performs audio decoding using the coded data of the current frame correctly received at the current time point, and outputs a decoded audio signal ((c) in FIG. 8). When the sound mode information of the previous frame, which is the lost frame obtained from the separating unit 113, indicates the transient mode, the switching unit 115 outputs the output of the time axis compressing unit 114, , The output of the lost frame decoding unit 105 is output to the window adding unit 107.

The window addition unit 107 includes a switching unit 115
(B) or (d) of FIG. 8 and the output signal ((c) of FIG. 8) of the audio decoding unit 102
At the start of the frame,
Is added at the end end so that the output signal that has passed through the switching unit 115 becomes dominant, and is output as a decoded audio signal. Then, in the switching unit 106,
Switching according to the above-described frame loss information is performed, and a decoded audio signal of the frame is output.

Next, the operation of the speech decoding apparatus according to the present embodiment will be described. FIG. 9 is a flowchart showing the operation on the encoding side. In step 301, speech encoding is performed on input speech in frame units, and encoded data is output. Next, in step 302, the sound mode of the input sound for each frame is determined. Then, in step 303, the encoded data obtained in step 301 and the audio mode obtained in step 302 are multiplexed and output as encoded data including audio mode information.

FIG. 10 is a flowchart showing the operation on the decoding side. First. In step 304, the transmitting side receives the encoded data that is encoded in frame units and transmitted in packet units. Further, the received packet-based data is decomposed into frame-based encoded data, and information indicating whether or not a frame loss has occurred (frame loss information).
Output with Next, in step 305, the coded data including the voice mode information is separated into voice mode information and voice coded data.

Next, in step 306, frame loss information for the current frame is determined. First, if it is determined that the encoded data for the current time frame has been correctly received without loss or delay, step 307 is executed.
To perform normal speech decoding using the encoded data, and output the decoded speech signal.

On the other hand, if it is determined that the coded data for the current frame cannot be correctly received due to a lost frame, that is, a loss or delay, the process proceeds to step 308, where loss frame compensation is performed, and a decoded speech signal obtained thereby is output. .

The previous frame is a lost frame,
If it is determined that the encoded data of the lost frame has been received with a time delay at the current time, step 309 is executed.
To perform speech decoding using the encoded data of the previous frame and the encoded data of the current frame correctly received at the current time, and output a decoded audio signal for the encoded data of the previous frame and the current frame. Then, step 310
, The state data necessary for decryption is updated.

Next, in step 311, the decoded speech signal corresponding to the encoded data of the previous frame and the current frame obtained in step 309 is subjected to time axis compression so as to be a signal of a section of one frame. Output. Next, in step 312, normal speech decoding is performed using the coded data of the current frame correctly received at the present time, and a decoded speech signal is output.

Next, in step 313, the voice mode information of the previous frame, which is the lost frame, obtained in step 305 is determined. If it is determined in this determination that the audio mode information is in the transient mode, the process proceeds to step 314, where windowing addition using the output of the process of step 311 and the output of the process of step 312 is performed, and the resultant is output as a decoded audio signal. . On the other hand, if it is determined that the voice mode information is not in the transient mode, the process proceeds to step 315, where windowing addition using the decoded voice signal for the encoded data of the current frame obtained in the process of step 309 and the output of the process of step 312 is performed. And outputs it as a decoded audio signal. Step 307, Step 308, Step 314
Alternatively, after performing the processing of step 315, step 316
Outputs the decoded audio signal as an output signal.

As described above, according to the speech decoding apparatus according to the present embodiment, when encoded data of a lost frame is received with a time delay, after outputting decoded speech by frame compensation, the encoded And using the encoded data of the current normal frame, perform speech decoding of the section of the past lost frame and the current normal frame, and when the mode information determined on the encoding side indicates that it is a transient section Since the obtained decoded signal is time-axis-compressed and used as decoded speech of a normal frame, it is possible to suppress deterioration of decoded speech quality due to frame loss in a transient section.

[0068]

As described above, according to the speech decoding apparatus of the present invention, when coded data of a lost frame is received with a time delay, the decoded speech is output by frame compensation, and After performing speech decoding for the section corresponding to the past lost frame and updating the state data necessary for speech decoding using the encoded data, the speech decoding of the current normal frame is performed to obtain a more original value. Thus, decoding using state data close to the above can be performed, and deterioration of decoded speech quality can be suppressed.

According to the speech decoding apparatus of the present invention, when encoded data of a lost frame is received with a time delay, a decoded speech is output by frame compensation, and then the encoded data is used. After updating the state data by performing speech decoding for the section corresponding to the past lost frame, the decoded speech signal obtained by performing the speech decoding of the current normal frame and the state data are updated by the frame compensation processing of the lost frame. Then, windowed addition is performed using the decoded audio signal of the current normal frame obtained by performing audio decoding after the decoding, so that smoother decoding can be performed.

Further, according to the speech encoding / decoding apparatus of the present invention, when encoded data of a lost frame is received with a time delay, after outputting decoded speech by frame compensation, the encoded data and Using the encoded data of the current normal frame, perform speech decoding of the section of the past lost frame and the current normal frame, and when the mode information determined on the encoding side indicates that it is a transient section By compressing the obtained decoded signal on the time axis and using it as decoded speech of a normal frame, it is possible to suppress degradation of decoded speech quality due to frame loss in a transient section.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a speech decoding device according to Embodiment 1 of the present invention.

FIG. 2 is a view for explaining processing when encoded data of a lost frame is received with a delay in the speech decoding apparatus according to the first embodiment.

FIG. 3 is a flowchart showing the operation of the speech decoding apparatus according to the first embodiment.

FIG. 4 is a block diagram showing a configuration of a speech decoding apparatus according to Embodiment 2 of the present invention.

FIG. 5 is a view for explaining processing when encoded data of a lost frame is received with a delay in the speech decoding apparatus according to the second embodiment.

FIG. 6 is a flowchart showing the operation of the speech decoding apparatus according to the second embodiment.

FIG. 7 is a block diagram showing a configuration of a speech encoding / decoding device according to Embodiment 3 of the present invention.

FIG. 8 is a view for explaining processing when encoded data of a lost frame is received with a delay in the speech encoding / decoding apparatus according to Embodiment 3.

FIG. 9 is a flowchart showing the operation of the speech encoding / decoding device according to the third embodiment.

FIG. 10 is a flowchart showing the operation of the speech encoding / decoding device according to the third embodiment;

FIG. 11 is a block diagram showing a configuration of a conventional speech decoding device.

[Explanation of symbols]

 Reference Signs List 101 reception buffer 102 audio decoding unit 103 state data holding unit 104 lost frame compensation unit 105 lost frame decoding unit 106, 115 switching unit 107 windowed addition unit 110 audio mode determination unit 111 audio encoding unit 112 multiplexing unit 113 separation unit 114 Time axis compression unit

Claims (7)

[Claims] 1. A receiving buffer for receiving and storing coded data in frame units, a voice decoding means for performing normal voice decoding using the coded data when the coded data of the current frame is correctly received, Loss frame compensating means for performing loss frame compensation when encoded data of a frame cannot be received due to loss; and when the reception buffer receives encoded data with a time delay, the encoded data and the current frame are compared. A speech decoding apparatus comprising: speech decoding using encoded data; and a lost frame decoding unit that updates state data necessary for decoding by the speech decoding unit. 2. A receiving buffer for receiving and storing coded data in frame units, a voice decoding means for performing normal voice decoding using the coded data when the coded data of the current frame is correctly received, Loss frame compensating means for performing loss frame compensation when encoded data of a frame cannot be received due to loss; and when the reception buffer receives encoded data with a time delay, the encoded data and the current frame are compared. A lost frame decoding unit for performing audio decoding using the encoded data and updating state data required for decoding by the audio decoding unit; a decoded audio signal of the current frame decoded by the lost frame decoding unit; and the audio decoding unit And windowing addition means for performing windowing addition using the decoded audio signal of the current frame decoded by Voice decoding device. 3. A voice mode determining means for determining a voice mode representing characteristics of an input voice on a frame basis, and voice mode information obtained by performing voice coding on the input voice on a frame basis and obtaining said voice mode determining means. And a receiving buffer for receiving and storing coded data in frame units, and a normal voice decoding using the coded data when the coded data of the current frame is correctly received. Voice decoding means for performing loss-frame compensation when the encoded data of the current frame cannot be received due to loss, and when the reception buffer receives the encoded data with a time delay. Speech decoding is performed using the encoded data and the encoded data of the current frame, and state data necessary for decoding by the audio decoding means. Frame decoding means for updating the lost frame decoding means, a time axis compression means for time axis compressing the lost frame decoded by the lost frame decoding means and the decoded audio signal of the current frame, and a voice mode obtained by the voice mode determination means. And selecting either the decoded audio signal obtained by the time axis compression means or the decoded audio signal of the current frame obtained by the lost frame decoding means, and the selected decoded audio signal and the audio decoding means And a window addition means for performing window addition using the decoded speech signal of the current frame. 4. The speech decoding device according to claim 1, the speech decoding device according to claim 2, or the speech encoding / decoding device according to claim 3.
A base station device comprising any one of the decoding devices. 5. The speech decoding device according to claim 1, the speech decoding device according to claim 2, or the speech encoding / coding device according to claim 3.
A communication terminal device comprising any one of the decoding devices. 6. A reception buffering step of receiving and storing encoded data in frame units, and a speech decoding step of performing normal audio decoding using the encoded data when the encoded data of the current frame is correctly received. A lost frame compensation step of performing lost frame compensation when the encoded data of the current frame cannot be received due to loss; and, when the reception buffer receives the encoded data with a time delay, the encoded data and the current A voice decoding method for performing voice decoding using encoded data of the frame and updating state data necessary for decoding in the voice decoding step. 7. A recording medium which stores an audio decoding program and is readable by a computer, wherein the audio decoding program receives and accumulates encoded data in units of frames, and corrects the encoded data of the current frame. A procedure for performing normal speech decoding using the encoded data when received, a procedure for compensating for a lost frame when the encoded data of the current frame cannot be received due to loss, and a time delay for the encoded data. A step of performing speech decoding using the encoded data and the current frame encoded data and updating state data necessary for decoding when the received data is received.