JPH06104852A - Voice decoder - Google Patents

Abstract

PURPOSE:To provide the voice decoder able to prevent quality deterioration in a decoded voice signal caused attended with omission of transmission of a voice signal for a non-sound period. CONSTITUTION:The voice decoder provided with a transmission coding decoding means 53 decoding a transmission code sent for a voiced period and outputting a decoding signal, a noise generating means 55 generating a noise signal for the non-voice period, and a changeover switch 56 selecting and outputting either output of the transmission code decoding means 53 or the noise generating means 55 is provided with a noise power control means 59 controlling the power of a noise signal caused for the non-voice period in response to the power of a decoding signal for a voiced period. Thus, the power of the noise signal generated from the noise generating means 55 is controlled in response to the power of the decoding signal by the noise power control means 59, resulting that the continuity of the signal is kept for the transition from the voiced period to the non-voiced period.

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 digital mobile communication or the like, and more particularly, it is constructed so as to prevent deterioration of decoded speech quality due to low power consumption.

[0002]

2. Description of the Related Art In a digital mobile communication device, in order to reduce power consumption of a transmitter, a signal is transmitted only when an input signal is voiced, and the operation of a transmission amplifier is stopped when the input signal is silent. There is.

In this digital mobile communication device, as shown in FIG. 2, an input voice encoder 1 for encoding an input voice and a code representing a voiced / unvoiced voice of an input signal are provided on a transmitting side (a). Sound / silence determiner 3 for output and input speech encoder 1
And a signal assembler 2 that multiplexes the signals output by the voice / silence determiner 3, while the reception side (b) represents the voice signal code and voice / silence from the received transmission code. It is provided with a signal decomposer 4 for separating the code and a voice decoding device 5 for decoding the code of the transmitted voice signal in the voiced section and generating a signal that changes in the voiced section.

In this apparatus, the voice / non-voice determining unit 3 on the transmitting side determines the voice / non-voice of the input voice coded by the input voice encoder 1, and determines the start of the voice period and the silence period. To generate a code indicating the start of. The signal assembler 2 is
This code and the output code of the input speech encoder 1 are multiplexed to assemble a transmission code. The transmission amplifier that transmits this transmission code stops the operation after transmitting the code for the start of silence, and restarts the transmission with the code indicating the start of the voiced section.

On the receiving side, the signal decomposer 4 decomposes the transmission code into the code of the voice signal and the code indicating the start of voiced / silent, and inputs the coded voice to the voice decoding device 5. The voice decoding device 5
In the voiced section, the code of the transmitted voice signal is decoded,
Since the code of the voice signal is not transmitted in the silent section, a noise signal is generated and output from the terminal 9 as an alternative.

This noise signal causes the receiving party to recognize that the call is continuing. If it becomes completely silent, the receiver cannot distinguish whether the call is ended or is silent.

As shown in FIG. 4, the conventional voice decoding device 5 includes a transmission code decoder 53 for decoding the code of a voice signal, a noise generator 55 for generating noise, and a transmission code decoder 53 for outputting an output signal. Or switch 56 to switch to the output of noise generator 55
And the noise generator 55 according to the code indicating the start of voiced / silent
And a sound / silence indicator 54 for instructing the operation of the switch 56.

In the voiced section, the code of the voice signal sent from the signal decomposer 4 is input to the transmission code decoder 53 through the terminal 51, and the code indicating the start of the voiced section is the terminal.
Input to the voice / silence indicator 54 through 52. The transmission code decoder 53 decodes the received code into an audio signal and switches it.
The switch 56 outputs the voice signal decoded by the transmission code decoder 53 to the terminal 57 in response to an instruction from the voice / silence indicator 54.

Further, when the code indicating the start of the silent section is input to the voice / silence indicator 54, the voice / silence indicator 54 instructs the noise generator 55 to start the operation, and the noise generator which receives the instruction.
55 outputs a signal of constant power. At the same time, the sound / silence indicator 54 instructs the switch 56 to switch the output, and the switch 56 outputs the noise signal of constant power sent from the noise generator 55 to the terminal 57.

The waveform of the signal at this time is shown in FIG. FIG. 3A shows an audio signal input to the input terminal 6 on the transmitting side, and the voice / silence determiner 3 determines a voice section and a silence section based on this signal (b). As shown in FIG. 3C, the voice decoding device 5 outputs the waveform of the input voice signal (a) in the voiced section and outputs the noise signal of constant power output from the noise generator 55 in the silent section. To do.
The change in power per unit time at this time is shown in (d).

As described above, the conventional voice decoding device 5 artificially inserts a noise signal of a constant power in a silent section where no voice signal is sent, and informs the receiver that the call is continuing. I understand.

[0012]

However, in the conventional speech decoding apparatus 5, since the signals of the voiced section and the silence section are generated independently by the transmission code decoder 53 and the noise generator 55, respectively. As shown in A of 3 (d),
There is no continuity in the power at the transition from the voiced section to the silent section. Therefore, depending on how the output power of the silent section is set, the output level may drop sharply or a signal with an unnecessarily large power may be output as soon as the section moves from the voiced section to the silent section. This is a cause of deterioration of the quality of decoded speech.

The present invention solves such a conventional problem, and provides a speech decoding apparatus capable of preventing quality deterioration of decoded speech which is caused by omitting transmission of a speech signal in a silent section. The purpose is to do.

[0014]

Therefore, in the present invention, a transmission code decoding means for decoding a transmission code transmitted in a sound section and outputting a decoded signal, and a noise generating means for outputting a noise signal in a silent section. And a changeover switch for selecting and transmitting the output of either the transmission code decoding means or the noise generating means, in a voice decoding device, the power of the noise signal output in the silent section is set to the decoded signal in the voiced section. A noise power control means is provided for adaptively controlling the power.

[0015]

Therefore, the noise signal generated by the noise generating means is controlled by the noise power control means in accordance with the power of the decoded signal, and as a result, in the transition from the voiced section to the silent section, The continuity of the signal is maintained, and a natural-feeling call can be performed between the transmitting side and the receiving side.

[0016]

BEST MODE FOR CARRYING OUT THE INVENTION As shown in FIG. 1, a speech decoding apparatus according to an embodiment of the present invention includes a decoded signal power detector 58 for detecting the power of a decoded signal output by a transmission code decoder 53.
And a noise power controller 59 for controlling the output power of the noise generator 55. Other configurations are the same as those of the conventional device (FIG. 4).

This device operates as follows. When a signal indicating the start of the voiced section is input to the terminal 52, the voiced / silent indicator 54 instructs the switch 56 to switch the switch so as to output the signal on the transmission code decoder 53 side, and The transmission code decoder 53 decodes the code received from the terminal 51 into an audio signal. As a result, this decoded voice is output from the terminal 57 via the switch 56.

On the other hand, when the code indicating the start of the silent section is input to the sound / silence indicator 54 via the terminal 52, the sound / silence indicator 54 instructs the noise generator 55 to start operation. Further, the switch 56 is instructed to switch so that the signal on the noise generator 55 side is output. Further, the decoded signal power detector 58 that detects the power of the decoded signal output from the transmission code decoder 53 detects the power amount in the last section of the voiced section and notifies the noise power controller 59 of it.

The noise generator 55, which receives the instruction from the voice / silence indicator 54, outputs a noise signal and also a noise power controller.
59 indicates that the power of this noise signal is almost equal to the power notified from the decoded signal power detector 58 in the early part of the silent section, and gradually decreases to a constant amount of power as time passes. Take control.

As a result, the decoded speech output to the terminal 57 through the switch 56 has the waveform shown in FIG. 3 (e), and the change in power per unit time during the transition from the voiced section to the silent section is small (( b) of f), the continuity of the signal during the transition is maintained.

As described above, in the voice decoding device of the embodiment,
By adaptively controlling the output power of the signal from the noise generator 55, the continuity of the signal between the voiced section and the silence section is maintained, and the deterioration of the quality of the decoded speech is prevented.

[0022]

As is clear from the above description of the embodiments, the speech decoding apparatus of the present invention can reduce the consumption of transmission power even when the transmitting side omits the transmission of the speech signal in the silent section. By outputting a noise signal while maintaining continuity with the power of the decoded signal in the voiced section, it is possible to prevent deterioration of the quality of the decoded voice, and to make a natural call between the transmitting side and the receiving side. Can be done.

Further, by gradually reducing the output of this noise signal, it is possible to avoid a situation in which a signal of more power than necessary is transmitted from the terminal 57 in the silent section.

[Brief description of drawings]

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

FIG. 2 is a block diagram showing a configuration of a transmission side (a) and a reception side (b) of the digital communication device,

FIG. 3 is an explanatory diagram illustrating waveforms and powers of input / output signals,

FIG. 4 is a block diagram showing a configuration of a conventional speech decoding device.

[Explanation of symbols]

 1 Input speech encoder 2 Signal assembler 3 Voice / silence determiner 4 Signal decomposer 5 Speech decoding device 6, 7, 8, 9, 51, 52, 57 Terminal 53 Transmission code decoder 54 Voice / silence instruction 55 Noise generator 56 Switch 58 Decoded signal power detector 59 Noise power controller

Claims (1)

[Claims] 1. A transmission code decoding means for decoding a transmission code transmitted in a voiced section and outputting a decoded signal, a noise generation means for outputting a noise signal in a silent section, and the transmission code decoding means or noise generation. In a voice decoding device having a changeover switch for selecting and transmitting one of the outputs of the means, the power of the noise signal output in a silent section is controlled by being adapted to the power of the decoded signal in a voiced section. A speech decoding apparatus comprising a noise power control means.