JPH0614628B2 - Echo prevention method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION Field of the Invention The present invention, in a voice analysis and synthesis system, controls data output from a voice analysis unit based on sound source intensity information included in input data of a voice synthesis unit, The present invention relates to an echo generation blocking method for blocking the generation of echo signals.

[Background of the Invention]

A conventional method for preventing the generation of an echo signal in a voice analysis and synthesis system is to provide an echo suppressor or an echo canceller, for example.

FIG. 1 is an explanatory view of an example of a conventional echo generation prevention system, in which 1 is a telephone, 2 is a hybrid transformer, 3 is an echo suppressor (or echo canceller), 4 is a voice analysis unit, and 5 is a voice synthesis unit. , 6 are transmission lines. The voice analysis unit 4 and the voice synthesis unit 5 are collectively referred to as a voice analysis / synthesis system.

In such a conventional method, (a) an echo suppressor and an echo canceller, which are expensive, are necessary. In the case of (b) the echo suppressor, it is usually possible to control whether or not to disconnect the line according to the input level value. Since it is performed, malfunctions are likely to occur, and in the case of an echo canceller, it is necessary to perform canceller promptly, so it is technically very difficult to perform this accurately.

[Object of the Invention]

SUMMARY OF THE INVENTION An object of the present invention is to eliminate the above-mentioned drawbacks of the prior art and to provide an echo generation preventing system for completely, reliably and economically preventing the generation of an echo signal in a voice analysis / synthesis system.

[Outline of Invention]

The echo generation prevention method according to the present invention analyzes voice intensity information of incoming data to a voice synthesis unit of a voice analysis / synthesis system, and when the analysis result corresponds to a silent state, the voice analysis / synthesis unit The data from the voice analysis unit of the system is sent as it is, and if it does not correspond to the silent state, the data from the voice analysis unit is converted so that the sound source intensity corresponds to the silent state and sent. By controlling so that the generation of an echo signal in the transmission data is prevented.

Example of Invention

 Embodiments of the present invention will be described below with reference to the drawings.

FIG. 2 is a block diagram of an embodiment of a voice analysis / synthesis system to which the echo generation prevention system according to the present invention is applied, and FIG. 3 is a processing flowchart thereof.

Here, 11 is a telephone, 12 is a hybrid transformer,
13 is a low-pass filter related to the voice analysis unit, and 14 is the same A
D converter, 15 is the same voice analyzer, 16 is a low-pass filter related to the voice synthesizer, 17 is the same DA converter, 18 is the same voice synthesizer, 19 is a control unit, 20 is a digital interface unit, 21 is a transmission line. Is. The digital interface unit 20 has functions such as parallel conversion, serial conversion, and digital switch.

First, in the case of transmission, the analog signal input from the telephone 11 passes through the hybrid transformer 12, the aliasing component is removed by the low-pass filter 13, the AD converter 14 converts the analog signal into a digital signal, and the voice analyzer 15 converts the voice signal. Parameters are extracted. The result is read by the control unit 19 via the bus and formed into a predetermined frame format, and then the digital interface unit 2
0 is transmitted to the transmission circuit 21 in a serial format.

On the other hand, in the case of reception, the data input from the transmission line 21 to the digital interface unit 20 is converted into a parallel format there, then read by the control unit 19 and given to the voice synthesizer 18 via the bus. The digital voice signal reproduced by the voice synthesizer 18 is DA
The analog signal is converted into an analog signal format by the converter 17, becomes an analog signal shaped by the low-pass filter 16, and is sent to the telephone 11 through the hybrid transformer 12.

In the above operation, the control unit 19 determines the sound source intensity included in the reception data read from the digital interface unit 20, and controls the transmission intensity of the transmission data so that the reception data always has priority. That is, when the received information is voiced indicating the presence of voice, the analysis result received from the voice analyzer 15 is format-converted, and when the transmission data is transferred to the digital interface unit, the sound source intensity is changed to Not to show the presence of (silently)
Let

As a result, the received voice signal is transmitted to the hybrid transformer 12
It wraps around from the low pass filter 13, AD converter 1
4. Even when the signal flows to the voice analyzer 15, the control unit 19 changes the sound source intensity information so as to represent the state of no sound (silence), so that the echo signal is not output from the digital interface unit 20 to the transmission line 21. Absent.

Specifically, for example, the voice analyzer 15 is a DSP (digital signal processor), the control unit 19 is a microcomputer with a built-in memory, and the voice synthesizer 18 is a dedicated LSI.
The voice analyzer 15 extracts the parameters representing the characteristics of the voice, the fundamental period, the voiced unvoiced information, the sound source intensity, the transfer characteristics, etc., and transfers them to the control unit 19. In addition to reading these data, the control unit 19 also reads the received data from the partner telephone, stores the data in the built-in memory, and outputs the data to the voice synthesizer 18. Assuming that there is no voice when the sound source intensity = 0, the sound source intensity is read from the incoming data stored in the memory, and if it is 0 (including a value that can be considered practically silent), voice analysis is performed. The data transferred from the device 15 is output to the digital interface unit 20 only by converting it into a predetermined format.
Of the data transferred from the voice analyzer 15, the data related to the sound source intensity is changed to 0, subjected to a predetermined format conversion, and then output to the digital interface unit 20.

〔The invention's effect〕

As described above in detail, according to the present invention, it is possible to prevent the generation of an echo signal economically and completely and reliably without using an expensive device such as an echo canceller or an echo suppressor. In addition, a remarkable effect can be obtained in improving call quality and economic efficiency of telephone lines.

[Brief description of drawings]

FIG. 1 is an explanatory view of an example of a conventional echo generation prevention system,
FIG. 2 is a block diagram of an embodiment of a voice analysis / synthesis system to which the echo generation prevention system according to the present invention is applied, and FIG. 3 is a processing flowchart thereof. 11 ... Telephone, 12 ... Hybrid transformer, 13
...... Low-pass filter, 14 ... AD converter, 15 ...
Speech analyzer, 16 ... Low-pass filter, 17 ... DA
Converter, 18 ... Voice synthesizer, 19 ... Control unit, 20 ...
… Digital interface, 21… Transmission line.

Claims (1)

[Claims] 1. A voice analysis / synthesis system for digitizing analog voice information, transmitting the digital data, and synthesizing analog voice information from the received digital data, wherein the voice analysis / synthesis system comprises a digital voice analyzer and a digital voice. It comprises a synthesizer and a control unit and a storage unit common to them, stores the received digital data in the storage unit, and the control unit determines the presence of the received data based on the voice intensity information of the stored digital data. A sound state and a silence state are discriminated. As a result of the discrimination, when the received digital data corresponds to the silence state, the digitized digital data is sent as it is from the voice analysis / synthesis system. If the discrimination result corresponds to the voiced state, the sound source strength of the above digital data to be transmitted. Echogenic blocking method characterized by sending by converting such that corresponds to the silent state, to prevent the occurrence of the echo signal in the transmission data generated by the received data has.