JPH11243376A - Sound decoding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To save power in a digital sound decoding device utilizing a radio line by reducing the deterioration of sound quality caused by continuing correction even for a frame with weak power at the time of correcting the deterioration of reproduced sound quality due to burst-like transmission error. SOLUTION: The number of error frames to be counted by an error frame number counter 24 is inputted to a control frame number arithmetic equipment 41, the frame affected by an error, which comes after the error frame, is added and a power correction control frame number (r) is calculated. In this case, power information (q) from a power calculator 22 is fetched and a correcting section corrected by a power corrector 23 is made to be short when power is smaller than a prescribed threshold value.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a decoding apparatus for voice coded communication used in digital communication, and more particularly to a voice decoding apparatus which takes measures against bursty transmission errors when a radio line is used as a propagation path. Things.

[0002]

2. Description of the Related Art Wireless voice communication is about to transition from an analog system to a digital system. Under such circumstances, there is an urgent need to establish a voice digital coded communication system with low power consumption and a small circuit configuration. The voice digitizing method is a method used in digital communication in a wired network, for example, an adaptive difference PCM (ADPC
The application of the M) method is most easily realized and has high consistency. However, in the case of wireless propagation, transmission errors often occur in the environment of the wireless section, and the communication quality is affected by the propagation path, and the quality of the reproduced voice is greatly degraded by the influence of the transmission errors, which causes discomfort to the user. It is also known that a phenomenon such as a loss of naturalness of sound quality occurs, which leads to deterioration of speech quality.

In the current ADPCM system, a system has been proposed in which a countermeasure is provided by adding a comparative attenuator. As for an error detection method, it is common to use an error detection code or an error correction code, but there is a drawback that information transmitted increases redundancy. In addition, there is a method of inferring from a state of occurrence of an error such as a code string for frame synchronization in transmission data for each frame. Here, considering application to a cordless telephone, the telephone is considered to be used under a slow fading environment, and a so-called burst error in which transmission errors continuously occur is predicted. Such errors can be improved to some extent by techniques such as retransmission, but the delay causes loss of naturalness, and retransmission techniques are not suitable for speech signal processing such as conversation. Although the use of an error correction code is conceivable, an increase in the amount of information cannot be avoided, and this is not appropriate. Furthermore, the above-mentioned speech codec with a comparative attenuator cannot be adapted when the speech coding method is different. Therefore, if a burst-like error countermeasure can be realized without information redundancy, its utility is great.

FIG. 4 is a block diagram of a speech coded communication apparatus to which the present invention is applied, wherein (A) is a transmitting section and (B) is a receiving section. First, in the transmitting section (A), a speech signal input from the microphone 1 is encoded by a speech encoder 2, and a multiplexer 3 encodes speech encoded data and control data output from a control data generator 5. Are multiplexed, and the multiplexed data is modulated by the modulation transmitter 4 to generate a modulated wave. The generated modulated wave is transmitted from the transmitting antenna 6.

In the receiving section (B), the modulated wave transmitted from the transmitting antenna to the radio transmission path is received by the receiving antenna 7,
The demodulator 8 demodulates and obtains the transmitted data sequence e. The output data e from the demodulator 8 is separated by the demultiplexer 9 into control data g and audio data h. The audio data h is decoded by the audio decoding device 10, and the decoded audio is output from the speaker 12. The controller 11 receives the data sequence e from the demodulator 8
Means for detecting a timing for separating control data g and audio data from the received signal f and outputting a frame pulse p, and detecting the timing (frame pulse p)
To the demultiplexer 9 and the audio decoder 10. Also,
The controller 11 has means for detecting the occurrence of a burst error from the received signal strength or the like, and outputs an error flag k to the speech decoding device 10 when the burst error is detected.

[0006] One of the inventors of the present invention has previously proposed a measure for preventing the deterioration of the quality of the received / reproduced voice affected by the above-mentioned radio transmission path (see Japanese Patent Application No. 4-95975).
FIG. 5 is a block diagram of the embodiment proposed earlier. 5 is a detailed configuration example of the speech decoding device 10 in FIG. This speech decoding apparatus comprises a correction frame number counter 24 for counting an error flag k from the controller 11 and a control frame number calculator 25 for calculating the number of frames to be power-corrected from the output of the error frame number counter 24. An interval calculator 26, and a data replacement unit 20 for replacing the coded data h in the error interval with specific data such as pseudo noise.
And a speech decoder 21, a power calculator 22 for obtaining the power of the decoded voice, and a frame r of the correction section obtained by the correction section calculator 26 based on the power information q from the power calculator 22. And a power corrector 23 that performs power correction on the decoded voice and outputs the reproduced voice.

In the configuration example shown in FIG. 5, the data replacing unit 20 replaces the data of the erroneous frame, and
By performing power correction on the corrected frame (the number of frames r) obtained by the correction section calculator 26, deterioration of the reproduced sound quality of the erroneous frame and the error-free frame section which is still affected by the error is reduced. Has been corrected.

FIG. 6 is a block diagram showing an example of the control frame number calculator 25 in FIG. FIG. 6A shows an example in which the multiplier 31 is used to calculate the control frame number r. The multiplier 31 multiplies the error frame number output from the error frame number counter 24 by the weighting coefficient stored in the storage unit 32. , And calculates and outputs the number r of control frames. The output r of the control frame number calculator 25 is, for example, α × N if the number of error frames is N frames and the weighting coefficient is α.

FIG. 6B shows an example in which an adder 33 is used to calculate the number of control frames. The number of error frames output from the error frame counter 24 is stored in the storage unit 32. The weighting coefficient is added by the adder 33 to calculate and output the control frame number r. The output r of the control frame number calculator 25 is, for example, N + β if the number of error frames is N frames and the weighting coefficient is β.

[0010]

However, in the configuration example shown in FIG. 5 proposed earlier, the power correction processing by the power corrector 23 is performed regardless of the power of the reproduced sound when there is a transmission error. Therefore, when the power of the decoded speech is small, the sound quality is not degraded due to the transmission error, and the power correction section is uselessly lengthened despite the small number of frames without transmission error, which is affected by the transmission error. There is a problem of excessive processing that the deterioration prevention processing is performed up to the section where there is no
For this reason, the naturalness of the reproduced sound is spoiled, and at the same time, the power consumption is increased, which has the drawback of hindering a reduction in power consumption.

An object of the present invention is to reduce the influence of transmission errors on reproduced sound and eliminate excessive processing in an inappropriate power correction section, which is a problem of the previously proposed circuit configuration. Another object of the present invention is to provide a speech decoding device with a burst error countermeasure function that can simultaneously reduce power consumption.

[0012]

A speech decoding apparatus according to the present invention receives encoded digital speech data obtained by receiving and demodulating a radio wave transmitted by a radio line, and outputs an error-free frame to the digital speech data. Is output, and when an error frame flag generated when a burst-like error is detected from the received demodulated signal is input, the digital voice data of the frame is replaced with a predetermined data sequence and output. A speech decoder that decodes the output of the data replacement unit, a power calculator that calculates the power of the decoded speech output from the speech decoder and outputs power information, and the error frame flag is input. An error frame number counter for counting and outputting a frame pulse indicating a frame division of the received demodulated signal, A control frame number calculator for calculating and outputting the number of control frames in the power correction section from the number of error frames output from the data and the number of subsequent frames affected by errors, and a decoded speech output from the speech decoder And a power corrector that performs power correction on the power correction section according to the power information and outputs a reproduced voice signal, wherein the control frame number calculator has a size of the power information. A plurality of weighting coefficient storage means for storing a corresponding weighting coefficient; and a threshold value storage means for storing a threshold value for selecting a weighting coefficient corresponding to the magnitude of the power information from the plurality of weighting coefficient storage means. A comparator for comparing the power information output from the power calculator with the threshold value, and the plurality of weighting coefficient storage means based on an output of the comparator. Selection means for selectively outputting a weighting coefficient corresponding to said power information from,
Multiplication or addition means for outputting the control frame number by multiplying or adding the weighting coefficient selected by the selection means and the output of the error frame number counter, and when the power information is larger than the threshold value, A power correction is performed by selecting a weighting coefficient for which the power correction section is long, and a power correction is performed by selecting a weighting coefficient for which the power correction section is short when the power information is smaller than the threshold. It is characterized by having been done.

[0013]

FIG. 1 is a block diagram showing an embodiment of the present invention. The difference from the previously proposed configuration of FIG.
The control frame number calculator 41 in the correction section calculator 40. The power frame q is calculated by adding the power information q from the power calculator 22 to the calculation of the control frame r. To eliminate over-treatment.

Therefore, the speech decoding apparatus of the present invention comprises a data replacement unit 20 for replacing the coded data h in the error frame section with specific data such as pseudo noise, a speech decoder 21,
A power calculator 22 for calculating the power of the decoded voice i;
An error frame number counter 24 for counting the error flag k from 1; a control frame number calculator for calculating the number r of frames to be power-corrected from the output of the error frame number counter 24 and the power information q from the power calculator 22 41, and a correction section calculator 40 composed of 41 and a frame of the correction section (the number of frames r) obtained by the correction section calculator 40
A power corrector 23 performs power correction on the decoded voice i based on the power information q from the power calculator 22 and outputs a reproduced voice.

FIGS. 2 and 3 are block diagrams showing a detailed example of the control frame number calculator 41 which is a main part of the embodiment of the present invention shown in FIG. 1, and a storage means for storing weighting coefficients. Is an example when there are two. FIG. 2 shows a control frame calculator 4 using a multiplier 31 to calculate the number r of control frames.
1 is a block diagram of the first embodiment. The multiplier 31 calculates the number of error frames output from the error frame number counter 24 and the weighting coefficient 1 stored in the storage means 51 or the weighting coefficient 2 stored in the storage means 52. It is configured to calculate the number of control frames r by multiplication. Power calculator 22
Is compared with the threshold information stored in the storage means 54 by the comparator 53, and based on the result, a switch (SW) 55 is switched to select a coefficient by which the number of error frames is multiplied.

To give a concrete example, the output of the control frame number calculator 41 is as follows: the number of error frames is N frames, the weighting coefficient 1 is α, the weighting coefficient 2 is 0.0, and the power information q
Is 1.5, and the threshold value stored in the storage means 54 is 0.9.
Then, the number of control frames becomes α × N, and if the power information q is 0.5, the number of control frames becomes 0. in this way,
When the power information q is larger than the threshold, the power correction is performed for a certain fixed number of frames. When the power information q is smaller than the threshold, the power correction section is controlled so as to be shorter (shorter).

Next, FIG. 3 is a block diagram of a control frame number calculator 41 using an adder 33 for calculating the control frame number r. The weighting coefficient 1 stored in the means 51 or the weighting coefficient 2 stored in the storage means 52 is added by the adder 33 to calculate the control frame number r. The comparator 53 compares the power information q output from the power calculator 22 with the threshold information stored in the storage unit 54, and based on the result, a switch (SW) 55
Is switched to select a coefficient to be added to the number of error frames.

To give a concrete example, the output of the control frame number calculator 41 is as follows: the number of error frames is N frames, the weighting coefficient 1 is β, the weighting coefficient 2 is 0.0, and the power information q
Is 1.5, and the threshold value stored in the storage means 54 is 0.9.
Then, the number r of control frames becomes N + β, and the power information q
Is 0.5, the control frame number r becomes N. As described above, when the power information is larger than the threshold value, the power correction is performed for a certain fixed number of frames. When the power information is smaller than the threshold value, the power correction section is controlled so as to be small (short).

In the case where a means for storing an arbitrary number (here, m) of weighting coefficients is provided, m storage means for storing m threshold values are compared with m threshold values. The configuration can be easily made by providing a comparator for taking the AND of the result.

As described above, if the control frame r is calculated by the control frame number calculator of the present invention shown in FIGS. 2 and 3, when the power of the decoded voice is large, the correction section is almost the same as the conventional one. However, it can be seen that when the power of the decoded speech is small, the correction section is smaller than the conventional section, and the correction section can be a correction section that avoids excessive correction, reduces the deterioration of sound quality, and does not impair naturalness.

Further, by reducing the number of correction sections, it is possible to shift the unused storage means and the like in the power corrector to the low power consumption mode in other than the correction section, thereby enabling low power consumption.

[0022]

As described above in detail, according to the present invention, the correction section calculation of the conventional speech decoding apparatus with a burst error countermeasure function comprising a data replacement, a correction section calculator, a power corrector and the like. By improving the decoder, it is possible to realize a speech decoding device that can improve the naturalness of the reproduced sound and reduce the power consumption compared to the past, and achieve higher performance in a system with a wireless section where burst transmission errors are likely to occur. High quality calling is possible.

[Brief description of the drawings]

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

FIG. 2 is a diagram illustrating a first configuration example of a control frame number calculator showing a main part of the present invention;

FIG. 3 is a diagram illustrating a second configuration example of a control frame number calculator which is a main part of the present invention.

FIG. 4 is a block diagram of a transmitting apparatus and a receiving apparatus of a speech coded communication system to which the present invention is applied.

FIG. 5 is a block diagram of a speech decoding device with sound quality deterioration reduction proposed earlier.

FIG. 6 is a block diagram showing a detailed configuration example of the control frame number calculator proposed above.

[Explanation of symbols]

 Reference Signs List 1 microphone 2 audio encoder 3 multiplexer 4 modulation transmitter 5 control data generator 6 transmission antenna 7 reception antenna 8 demodulator 9 demultiplexer 10 audio decoder 11 controller 12 speaker 21 audio decoder 22 power calculator 23 Power Corrector 24 Error Frame Number Counter 25 Control Frame Number Calculator 26 Correction Section Calculator 30 Correction Section Calculator 31 Multiplier 32 Storage Means (Weighting Coefficient) 33 Adder 40 Correction Section Calculator 41 Control Frame Number Calculator 51 Storage means (weighting coefficient 1) 52 Storage means (weighting coefficient 2) 53 Comparator 54 Storage means (threshold) 55 Switch (SW)

Claims (1)

[Claims] 1. Encoded digital voice data obtained by receiving and demodulating a radio wave transmitted by a wireless line is input, and an error-free frame outputs the digital voice data. A data replacer that, when an error frame flag generated upon detection of the frame is input, replaces the digital audio data of the frame with a predetermined data sequence and outputs the data, and decodes the output of the data replacer. A speech decoder, a power calculator for calculating power of a decoded speech output from the speech decoder and outputting power information, and indicating a frame division of the received demodulated signal when the error frame flag is input. Error frame number counter that counts and outputs frame pulses, error frame number output from the frame number counter, and A control frame number calculator for calculating and outputting the number of control frames in the power correction section from the number of frames affected by the following errors, and the power for the power correction section of the decoded speech output from the speech decoder. A power corrector that performs power correction in accordance with information and outputs a reproduced audio signal, wherein the control frame number calculator stores a weighting coefficient corresponding to the magnitude of the power information. Weighting coefficient storage means, threshold value storage means for storing a threshold value for selecting a weighting coefficient corresponding to the magnitude of the power information from the plurality of weighting coefficient storage means, and output from the power calculator. A comparator for comparing power information with the threshold value, and outputting a weighting coefficient corresponding to the power information from the plurality of weighting coefficient storage means based on an output of the comparator. Selecting means for selecting and outputting, and multiplying or adding means for multiplying or adding the weighting coefficient selected by the selecting means and the output of the error frame number counter to output the control frame number, the power information Is larger than the threshold, power correction is performed by selecting a weighting coefficient that makes the power correction section longer, and when the power information is smaller than the threshold, a weighting coefficient that makes the power correction section shorter is selected to An audio decoding device characterized in that correction is performed.