SU1662014A2 - Simulated voice signal generator - Google Patents

Abstract

The invention relates to the field of electrical and radio communications, can be used as a generator in control devices for systems for the transmission and processing of voice information, in particular echo cancellers, and is an improvement of the known device according to the author. N 1571777. To increase the reliability of speech simulation by simulating separately the static parameters of the amplitude distribution of the male and female voices speech tones, units 10 and 11 of the first and second dividers, a pseudorandom sequence generator 1, a clock frequency generator 2, a summation block 3, counting unit 4, switching unit 5, first and second generators 6, 7, first and second half-wave rectifiers 8, 9.

Description

1J

Ik

cl

with

-0

oh oh oh

Yu

The invention relates to the field of electrical and radio communications and can be used as a generator in control devices for systems for the transmission and processing of voice information, in particular, echo-compression devices, and is an improvement of the known device according to the author No. 1571777.

The aim of the invention is to increase the reliability of imitation of a speech signal by simulating separately the statistical parameters of the amplitude distribution of the speech tone of a male and female voice.

The drawing shows a block diagram of the device.

The device contains a pseudorandom sequence generator 1, a clock frequency generator 2, a summation unit 3, a counting unit 4, a switching unit 5, the first and second generators 6, 7, the first and second half-wave rectifiers 8, 9, the first and second blocks

10.11 dividers, in this case the summation unit 3 is connected in series, the first input of which is connected to the output of the generator 1, the input of which is connected to the first output of the generator 2, the switching unit 5, the first generator 6 and the first single-half-rectifier 8, the output of which is connected A second generator 7 is connected in series with the first input of the first block 10, the second input of which is connected to the first output of block 5; the second generator is connected to the second output of block 5, the second rectifier 1 9 whose output is connected to the first input of the second block

11, the second input of which is connected to the second output of block 5, the outputs of the first and second blocks 10; 11 are combined and are the output of the device.

The device allows simulating speech signals of male and female voices in devices for monitoring the characteristics of echo-suppressing devices of the compensation type. Such a characteristic, for example, is the time of convergence (self-tuning) —the dynamic characteristic of an echo canceller.

The device works as follows.

From the first output of the generator 2, the clock frequency pulses (120 or 240 Hz) are fed to the input of the generator 1, from the output of which binary pseudo-random numbers, having a parallel sixteen-bit form, are fed to the first input of block 3. At the same time, the clock pulses from the second output of the generator 2 arrive at the first input of the counting unit 4,

where first up to sixteen clock pulses are counted to obtain an intermediate result of a bitwise summation of a binary sixteen times a pseudo-random number with a previous binary sixteen-bit pseudo-random number in block 3 The intermediate result of the pores of a random sum comes in parallel

0 sixteen bits from the second output of block 3 to the second input of counting block 4, where the recording is performed, and then information is output in parallel form from the first output of counting block 4 to the second

5 input block 3 summation.

The summation process in block 3 continues cyclically sixteen times until sixteen binary sixteen-bit pseudo-random numbers

0 will not be bitwise summed. The result of the summation will be a binary sixteen-bit pseudo-random number having a Gaussian distribution law, which is fed to the input of block 5. The recording of the result of summing from the first output of block 3 to the input of block 5 occurs only when there is a pulse that permits recording and arrives from the second output of the counting unit 4 to the sync input

0 of block 5, block 5 generates control signals that are fed to the inputs of the first or second generator 6, 7, respectively, which, depending on the incoming control signal, change the frequency of the irradiated signal and are fed to the second input of the first or second blocks, respectively, 11, which, depending on the incoming control signal, attenuates the simulated speech signal at the device output in a given range of values.

The choice of the frequency deviation of the first and second generators 6, 7 is carried out in such a way that the total frequency deviation of the frequency equals, respectively, the rms change in the frequency of the tone signal of male and female speech. Signal sinusoidal frequency from the output of the first or second generator 6 or 7

0 is fed to the input of the corresponding first or second half-wave rectifier 8 or 9, in which the final formation of the tone signal of male or female speech occurs. The signal of constant amplitude and having a tone distribution of male and female speech, subject to Gaussian law, enters the first input of the first or second block of 10 or 11 dividers, in which the signal is attenuated in accordance with the Gaussian amplitude distribution of a simulated male or female voice . Thus, the distribution of the frequency and amplitude of the male and female speech signal at the output of the device will obey the Gaussian law.

Claims (1)

Claims of the invention Generated speech signal generator auth.St. No. 1571777, characterized in that, in order to increase the reliability of speech imitation simulation by separately modeling the amplitude statistical parameters prediction of male and female voices, the first and second divider blocks are inserted into it, the first inputs of the first and second divider blocks are connected to the first and second half-wave rectifiers, respectively, the second input of the first divider block is connected to the first output of the switching unit and the input of the first generator, the second input of the second divider unit is connected to the second output of the switching unit and the input of the second generator, and the outputs of the first and second divider units are combined and are output by roystva