JPH028899A - Voice decoder - Google Patents

Abstract

PURPOSE:To reduce a feeling of physical disorder at the time of inserting a random noise signal when a voice is absent during a conversation by adjusting the level of the random noise signal with power information. CONSTITUTION:A random noise generator8 generates the random noise signal 9 and an amplifier attenuator 13 supplies the power information 7 to calculate a level-adjusted random noise signal 14. Namely, when the power information 7 is large, the random noise signal 14 is large and when the power information 7 is small, the random noise signal 14 is small. Then when the power information 7 is large, a selector 10 selects a decoded signal 6 as the output signal 11 and when the power information 7 is small, the random noise signal 14 is selected. Consequently, the feeling of physical disorder is reduced when the random noise signal is inserted in the no-voice state in the conversation.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a speech decoder, particularly to a speech decoder in a speech analysis and synthesis system.

[Conventional technology]

Conventionally, this type of audio decoder compares the value of the power information of the input signal with a preset value, and if it is thick, outputs the decoded audio signal, and if it is small, outputs a certain level of random noise. Was.

A block diagram of the conventional audio decoder mentioned above is shown in FIG.

Input signal 2 to input terminal 1, and demultiplexer 5
The signal is divided into encoded signal 4 and power information 7.

A random noise generator 8 outputs a random noise signal 9 at a constant level. - On the other hand, the decoder 5 outputs a decoded signal 6. Next, power information 7
Using , either the decoded signal 6 or the random noise signal 9 is selected based on the R profit/loss 10, and this is set as the output signal 11.

[Problem to be solved by the invention]

Here, 9 is a random noise signal of a constant level, so when switching from the decoded signal 6 to the random noise signal 9, and from the random noise signal 9 to the decoded signal 6, a random noise signal of a constant level suddenly occurs. 9 is output, or the decoded signal 6 is suddenly output from the random noise signal 9 at a certain level.
There is a drawback that a strange feeling is felt when the power information 7 is small, that is, when there is no call, and when the power information 7 is large, that is, when switching during a call.

The present invention aims to eliminate these drawbacks of the conventional ones, and provides a voice decoder that reduces the sense of strangeness when inserting a random noise signal when there is no call during a conversation.

[Means to solve the problem]

According to the present invention, means for receiving an input signal and outputting a decoded signal, means for generating a random noise signal, and means for amplifying and attenuating the random noise signal according to the magnitude of power information of the input signal; An audio decoder characterized in that it has means for selecting whether to output a decoded signal or a random noise signal depending on the magnitude of the power information.

〔Example〕

Next, the present invention will be explained with reference to the drawings.

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

Input signal 2 to input terminal 1 and demultiplexer 6
The signal is divided into encoded signal 4 and power information 7. Next, the encoded signal 4 is converted into a decoded signal 6 by a decoder 5. A random noise signal 9 is generated in a random noise generator 8, power information 7 is given to an amplification/attenuation device 13, and a level-adjusted random noise signal 14 is calculated. Specifically, if the power information 7 is thick, the random noise signal 14 is also large.

If the power information 7 is small, the random noise signal 14 is also calculated to be small. Furthermore, the decoded signal 6 and the random noise signal 14 are used as the power information 7 by the selector 10.
, and output it to the output terminal 12 as the output signal 11. Specifically, for the output signal 11, if the power information 7 is large, the decoded signal 6 is selected, and if the power information 7 is small, the random noise signal 14 is selected.

In addition, as for the realization hardware, the random noise generator 8 and the amplification attenuator 13 are implemented by a microprocessor (ex, μPD7720AO, μPD7720AO, μPD
77230R, etc.).

〔Effect of the invention〕

As described above, the present invention has the effect of reducing the sense of strangeness when inserting a random noise signal during a silent conversation by adjusting the magnitude of the random noise signal using power information.

noise signal.

[Brief explanation of the drawing]

FIG. 1 shows a configuration diagram of an embodiment of the present invention, and FIG. 2 shows a configuration diagram of a conventional one.

Claims (1)

[Claims] 1. means for receiving an input signal and outputting a decoded signal; means for generating a random noise signal; means for amplifying and attenuating the random noise signal according to the magnitude of the power information of the input signal; 1. A speech decoder comprising means for selecting whether to output a decoded signal or a random noise signal depending on the magnitude.