JPH04234800A - System for silent reproduction in compressed voice - Google Patents

Abstract

PURPOSE:To prevent deterioration in articulation by compensating the reproduction of the part of the ending of a word even when a voice signal below a silence deciding threshold is cut when transmitted. CONSTITUTION:Silence deciding means 203 and 303 in composing and decomposing circuits 201 and 301 connected to voice signal digital compressing devices 10 and 40 compare the power value in a digitally compressed signal with a predetermined value to decide a silent section. Voiced section final end storage means 204 and 304 output successively (n) voiced compensation signals, which are generated by reducing the power value of the voiced section stored in the voiced section end signal storage means 204 and 304 to 1/2n and have the same feature parameters, as a substitute for voiceless deleted signals at the end of a voiced section when the voiced section end signal in the digitally compressed signal is stored and the signal after voiceless deletion is decomposed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a compressed audio silent reproduction system, and more particularly to a compressed audio silent reproduction system for compensating for clarity when transmitting digitally compressed audio analog signals.

0002

[Prior Art] Conventionally, in this type of compressed audio silence playback system, audio signals below a threshold value for which silence has been determined are not transmitted in order to increase the multiplexing effect in transmission, and during playback, signals in silent sections are not transmitted. In this method, the power value was set to 0 on the digital compressed signal, or a low-level white noise equivalent was reproduced as a silent signal. Therefore, the ending of the word was cut off as being below the threshold.

FIG. 5 is a diagram showing the signal transmission state and reproduction state in the conventional system.

In FIG. 5, the decomposed digital signals obtained by decompressing and decompressing the input analog audio signal are digitally compressed signals d1, d2, d4, ~d7, d11, d1 with a period T.
2 and a silence signal Z having a power value of 0 are arranged. Also,
In the reproduced analog signal, the section of the silent signal Z following the digitally compressed signal d2 and the section of the two silent signals Z following the digitally compressed signal d7 are cut portions c without being compensated for the presence of sound.

[0005]

This conventional compressed audio silent reproduction method has the disadvantage that the clarity of the reproduced audio deteriorates because the endings of words below the threshold are cut off.

An object of the present invention is to ensure that the characteristic parameters of n digitally compressed signals, especially at the end of words, of continuous speech in time series are almost the same, and that the power value of the signal tends to decrease. An object of the present invention is to provide a compressed speech silent reproduction method that improves the clarity of word endings.

[0007]

[Means for Solving the Problems] A compressed audio silent reproduction method of the present invention includes an audio signal digital compression device and an assembly/disassembly circuit connected to the audio signal digital compression device, the assembly/disassembly circuit being connected to the audio signal digital compression device. Silence determining means for determining a silent section by comparing a power value in the digitally compressed signal with a predetermined threshold; and a sound section for storing a final signal for a sound section in the digitally compressed signal. The section final signal storage means, when decomposing the signal in which silence has been deleted, 1/ (2n) Speech compensation means for sequentially outputting n (n is a positive integer) speech signal compensation signals having the same feature parameter reduced to (2n) in place of the silence-deleted signal. shall be.

Further, the digitally compressed signal may be a signal of a linear predictive encoding method or a multipulse encoding method.

[0009]

Embodiments Next, the present invention will be explained with reference to FIGS. 1 to 4.

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing signal waveforms of each part in FIG. 1, FIG. 3 is a diagram showing an example of a digital compressed signal format, and FIG. FIG. 3 is a flow diagram for explaining the operation of the decomposition circuit.

Input analog voice signal 5 from telephone 60
0 is input as a digital compressed signal 51 by the audio signal digital compression device 10 to the assembly/disassembly circuit 201 in the multiplex transmission device 20 . The digital compressed signal 51 is input to the assembly circuit 202 and is compared with a predetermined threshold value in the silence determination circuit 203 to determine whether there is any sound. transmitted to the station.

[0012] Then, at the other party's station, the active signal 53 of one call separated by the multiplex transmission circuit 306 in the multiplex transmission device 30 is transmitted.
is input to the separation circuit 304.

As shown in FIG. 4, the separation circuit 304 is activated at an audio signal generation period T, detects that the audio signal continuously sent from the multiplex transmission circuit 306 is interrupted, and determines that there is no audio. That is, if there is no received voice signal in step 1, check whether the count value = 0 or not (step 2)
, if No, the last signal of the sound section is stored in the memory 305. Next, the power values of the sound signals stored in the memory 305 as the first n signals of the silent section are sequentially 1/2, 1/
4, 1/8..., and sends out n pieces while silence continues to compensate for sound (step 3), and sends it to the audio signal digital compression device 40 as a signal 54 (step 4).
), the counter value is decremented by 1 (step 5).

[0014] Furthermore, if the sound signal starts before n signals are transmitted (Yes in step 1), the sound signal compensation operation is immediately stopped, the sound signal is stored in the memory 35 (step 6), and the sound signal is stored in the counter. The value=n is set (step 7), the signal 54 is sent to the audio signal digital compression device 40 (step 8), and the reproduced analog signal 55 is sent to the telephone 70.

Note that when the counter value is 0 in step 2 (Yes), a silence signal Z (shown in FIG. 2) with a power value of 0 is sent to the audio signal digital compression device 40 (step 9).

Next, the operation of this embodiment will be explained with reference to FIG.

In FIG. 2, the signals d1 to d12 of the signal 51 are signal sequences digitally compressed with a period T. Further, d2' of the signal 54 is a signal with the power value of the signal d2 reduced to 1/2, and d7' and d7'' are each a signal with the power value of the signal d7 reduced to 1/2.
/2 and 1/4, and Z is a silent signal with a power value of 0.

In this embodiment, an input analog audio signal 50 is digitally compressed by the audio signal digital compression device 10 and input as a signal 51 to an assembly/disassembly circuit 201, and a signal 52 is transmitted to the multiplex transmission device 30.

Note that the section of the signal 52 indicated by an arrow is a section in which no signal is transmitted as there is no sound. Upon receiving signal 52, multiplex transmission circuit 306 reproduces it as signal 53.

The signal 53 is the signal 5 in the assembly/disassembly circuit 301.
In the sections corresponding to d3 and d8 of signal 51, the power value is reduced to 1/2, and in the section corresponding to d9 of signal 51, the power value is reduced to 1/4.
Further, the section corresponding to d10 of the signal 51 is arranged as a silent signal Z, the signal 54 is input to the audio signal digital compression device 40, and the reproduced analog signal 55 is sent to the telephone 70. That is, the section corresponding to d2' and the sections corresponding to d7' and d7'' of the signal 54 become sections a and b for which the sound presence is compensated for in the reproduced analog signal 55.

FIG. 3 shows an example of the digitally compressed signal 51.
(a) is one signal of the linear predictive coding method, and (b) is one signal of the multipulse coding method. In Figure 3, S is a frame synchronization bit (1 bit)
, A is the signal power value (7 bits), P is the pitch information (6 bits), and P is the pitch information (6 bits).
bit), V represents voiced/unvoiced sound identification (1 bit), K represents a feature parameter (39 bits), and M represents multipulse information (152 bits). Note that P, V, K, and M are collectively referred to as feature parameters.

[0022]

Effects of the Invention As explained above, in the present invention, when reproducing the silence of a digital compressed signal, two signals are generated with the same characteristic parameter and only the power value as 1/(2n) of the final signal of the sound section. By continuing this signal to the final signal of the voiced section, even if the portion of the voiced signal below the silence determination threshold is cut during transmission, the reproduction of the final part of the word is compensated for, thereby preventing deterioration in clarity. has.

[Brief explanation of the drawing]

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

FIG. 2 is a diagram showing signal waveforms at various parts in FIG. 1;

FIG. 3 is a diagram showing an example of a digital compressed signal format.

FIG. 4 is a flow diagram for explaining the operation of the decomposition circuit in FIG. 1;

FIG. 5 is a diagram showing a signal transmission state and reproduction state in a conventional system.

[Explanation of symbols]

10, 40 Audio signal digital compression device 20, 3
0 Multiplex transmission device 60, 70 Telephone 80 Transmission line 201, 207, 301, 307 Assembly/disassembly circuit 202, 302 Assembly circuit 203, 303 Silence judgment circuit 204, 304 Disassembly circuit 205, 305 Sound section final storage circuit (memory) 206,306 Multiplex transmission circuit

Claims (3)

[Claims] 1. An audio signal digital compression device and an assembly/disassembly circuit connected to the audio signal digital compression device, the assembly/disassembly circuit predetermining the power value in the digitally compressed signal. Silence determining means for determining a silent section by comparing it with a threshold; sound section final signal storage means for storing the final signal of a speech section in the digitally compressed signal; and a signal from which silence has been removed. During the decomposition operation, the power value of the sound section final signal stored in the sound section final signal storage means at the end of the sound section is divided by 1/(2
n) with the same feature parameters reduced to n
and a voice compensation means for sequentially outputting the signal from which the silence has been removed as a voice signal compensation signal (wherein is a positive integer). 2. Claim 1, wherein the digitally compressed signal is a linear predictive coding signal.
Compressed audio silent playback method described. 3. The compressed audio silent reproduction method according to claim 1, wherein the digitally compressed signal is a signal of a multi-pulse encoding method.