JPH0744197A - Speech decoding device - Google Patents

Abstract

PURPOSE:To improve the aural speech quality by reducing the intermission of data due to decoding when an error is detected. CONSTITUTION:The number of limits N of a counter 12 and the number N of sent codes stored in a data buffer 13 are determined; when an error detector 11 decides that there is no error, sent codes from a transmission side are decoded by an ADPCM decoder 15 supplied with the codes through a selection switch 14 and the decoded codes are sent out as output data from a selection switch 17. If an error is detected, on the other hand, the counter 12 counts up in order of 1, 2, 3...N and data in buffers 2, 3...N in a data buffer 13 are passed through the selection switch 14 and decoded by the ADPCM decoder 15. When the counter 12 exceeds N, the decoded data whose amplitude is suppressed by a suppressing device 16 are sent out as output data from the selection switch 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention utilizes digital mobile communication and the like, and is recommended by the International Telegraph and Telephone Advisory Committee Recommendation G. Adaptive Differential Pulse Code Modulation (hereinafter referred to as ADPCM)
The audio decoding device for decoding the transmission code by the system audio encoding.

[0002]

2. Description of the Related Art FIG. 2 is a block diagram showing the structure of a conventional speech decoding apparatus. In FIG. 2, this speech decoding device includes an error detector 2 for detecting an error in a transmission code obtained by multiplexing a speech code of a predetermined fixed length from the transmitter side,
It has an ADPCM decoder 3 for decoding a transmission code. Further, this speech decoding device is supplied with error-time data 4 such that the amplitude in the output data becomes “0” when the error is detected by the error detector 2, and the decoding from the ADPCM decoder 3 or Error data 4
It has a selection switch 5 for transmitting the output data selected.

Next, the operation of this conventional configuration will be described. On the transmitter side, a transmission code in which a predetermined fixed length voice code is multiplexed is generated and transmitted. In the voice decoding device on the receiver side, the error detector 2 determines whether or not an error is detected for each transmission code transmitted from the transmission side, and switches the selection switch 5. When the error detector 2 does not detect an error, the received transmission code is decoded by the ADPCM decoder 3, and this decoding is sent as output data through the selection switch 5. When the error detector 2 detects an error, the error code 4 is output through the selection switch 5 without decoding the transmission code in which the error is detected so that the amplitude of the output data becomes “0”. .

[0004]

However, in the above-mentioned conventional speech decoding apparatus, when the error is detected by the error detector 2, the amplitude in the output data is set to "0". When this is done, there is a drawback in that audio interruptions are felt and the audio quality on hearing is degraded.

The present invention solves such a conventional problem. When an error is detected, data interruption in decoding is reduced, the continuity of voice is maintained, and the error continues. An object of the present invention is to provide an excellent voice decoding device in which abnormal noise is prevented from being generated when the voice is detected, and the audible voice quality is improved when voice is output.

[0006]

In order to achieve the above object, a speech decoding apparatus according to the present invention includes an error detection code for adding an error detection code and detecting an error when decoding a transmission code having a speech code. Means, counting means for counting the number of consecutive detected errors, data storage means for storing past transmission codes, and decoding means for decoding the transmission codes,
When the detected error is within the count limit number, it is replaced with the same number of accumulated past transmission codes as the count limit number of the counting means, and is decoded and output.

In this structure, the transmission code to which the error detection code is added is generated by the voice coding process by the waveform coding method of the adaptive differential pulse code modulation, and the decoding means is adapted by the adaptive differential pulse code modulation. Is used to decode the transmission code.

In addition to the above configuration, there is provided suppress means for suppressing the signal amplitude of decoding, and selecting means for selecting and outputting decoding from the decoding means or the suppress means or decoding in which the signal amplitude is suppressed. However, when the detected error exceeds the count limit, the decoding signal amplitude is suppressed and output.

[0009]

With this structure, the speech decoding apparatus of the present invention accumulates the transmission code transmitted in the past, which has the speech code added with the error detection code. Then, the transmission code in which an error has occurred is replaced with the accumulated transmission code, decoded, and output. If an error is continuously detected, the decoded signal amplitude is suppressed and output. Therefore, when an error is detected, data interruption in decoding is reduced. That is, the continuity of the voice is maintained by utilizing the correlation in the close interval of the voice, that is, the characteristic that the waveforms are similar. Furthermore, the generation of abnormal noises when an error is continuously detected is prevented, and the sound quality in the sense of hearing is improved when the sound is output.

[0010]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a speech decoding apparatus of the present invention will be described below in detail with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of a speech decoding apparatus of the present invention. In FIG. 1, the speech decoding apparatus includes an error detector 11 for detecting an error in a transmission code obtained by multiplexing a speech code of a predetermined constant length from a transmitter side, and an error detected by the error detector 11. Has a counter 12 for counting the number of consecutive transmissions, a data buffer 13 for accumulating past transmission codes, and a selection switch 14 for selecting and outputting the transmission codes from the data buffer 13 according to the value of the counter 12. . Further, this speech decoding device includes an ADPCM decoder 15 for decoding the transmission code selected by the selection switch 14, a suppress device 16 for suppressing the signal amplitude of the decoding from the ADPCM decoder 15, and an ADPCM decoder. 15 or suppressor 1
6 and the selection switch 17 for selecting the decoding in which the signal amplitude is suppressed.

Next, the operation of the configuration of this embodiment will be described. By specifying in advance how many past transmission codes are used, the limit number N of the counter 12 and
The number N of transmission codes accumulated in the data buffer 13 is determined. First, when the error detector 11 determines that there is no error, the counter 12 is “0”, and the transmission code from the transmitter side is transmitted through the selection switch 14 to the ADPCM decoder 15
And then decrypt it. Error detector 1 from voice transmission data
When 1 detects an error, the counter 12 becomes "1" and the selection switch 14 selects the buffer 1 in the data buffer 13. The transmission code stored in the buffer 1 is decoded by the ADPCM decoder 15, and this decoding is sent as output data through the selection switch 17.
When an error is continuously detected, the counter 12 is 2, 3, ...
The transmission code of the buffer 2, the buffer 3 ... The buffer N in the data buffer 13 is selected by the selection switch 14, respectively. After this, the ADPCM decoder 15
Will be decrypted with. In this case, if an error is subsequently detected and the counter 12 exceeds N, if a large number of past transmission codes are successively decoded, a waveform different from the waveform to be originally decoded, that is, the past accumulated waveform is accumulated. The transmission code is decoded. At this time, the transmission code is directly decoded through the selection switch 14, but this decoding is performed by the suppress device 16
The signal amplitude is suppressed at. Then, the decoding in which the amplitude of this signal is suppressed is sent as output data through the selection switch 17. In the decoding in which the amplitude is suppressed, since the amplitude of the signal is small, it is possible to prevent the generation of abnormal noise and suppress the deterioration of the audio quality in the sense of hearing.

[0013]

As is apparent from the above description, the speech decoding apparatus of the present invention accumulates the transmission code transmitted in the past, replaces the transmission code in which an error has occurred with the accumulated transmission code, and decodes it. Since the data is output after being output, there is less interruption of data in decoding. That is, there is an effect that the continuity of the voice is maintained by using the correlation between the voices in the close interval, that is, the characteristic that the waveforms are similar, and the voice quality in the sense of hearing is improved when the voice is output. Further, when an error is continuously detected, the decoding signal amplitude is suppressed and output, so that the generation of abnormal noise when an error is continuously detected is prevented, and the audio output is performed. In addition, it has the effect of improving the audio quality on hearing.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration in an embodiment of a speech decoding apparatus of the present invention.

FIG. 2 is a block diagram showing the configuration of a conventional speech decoding device.

[Explanation of symbols]

 11 error detector 12 counter 13 data buffer 14 selection switch 15 ADPCM decoder 16 suppressor 17 selection switch

Claims (3)

[Claims] 1. An error detection means for adding an error detection code and detecting an error when decoding a transmission code having a voice code, a counting means for counting the number of consecutive detected errors, and a past transmission code. A data storage means for storing
Decoding means for decoding the transmission code, and when the detected error is within the counting limit number, it replaces with the accumulated past transmission codes of the same number as the counting limit number of the counting means, and decodes and outputs. A voice decoding device characterized by. 2. A transmission code to which an error detection code is added,
2. The speech decoding apparatus according to claim 1, wherein the speech decoding processing is performed by a speech coding process using a waveform coding method of adaptive differential pulse code modulation, and the decoding means decodes a transmission code by adaptive differential pulse code modulation. . 3. Detecting errors are counted by suppressing means for suppressing the signal amplitude of decoding, and selecting means for selecting and outputting decoding from the decoding means or the suppressing means or decoding in which the signal amplitude is suppressed. The speech decoding apparatus according to claim 1, wherein when the limit number is exceeded, a decoding signal amplitude is suppressed and output.