JPS6053999A - Voice synthesizer - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a speech synthesizer used in a vocoder that inputs PARCOR coefficients as speech spectrum information.

On the transmitting side of the vocoder, the voice to be output is sent as information for choosing between the vocal cords to a speech synthesizer on the receiving side of the vocoder. Furthermore, on the transmitting side of the vocoder, not only the voiced/unvoiced information, which is the voiced/unvoiced discrimination signal mentioned above, but also spectral information, pitch information, etc. are transmitted every frame of about 20 mθ, which is considered to be short enough to accommodate changes in voice. is common. Therefore, in a conventional speech synthesizer, the above information is updated on a frame-by-frame basis, and a pulse generator is used as an excitation waveform for a synthesis filter in a voiced frame, and a noise generator is used as an excitation waveform in an unvoiced frame.

However, in actual speech, unvoiced sounds (e.g. "8")
) and voiced sounds (e.g. "a") do not change clearly and suddenly, but rather gradually compared to the length of the frame, which is one of the causes of unnaturalness in the output sound of the vocoder. ing.

Also, in order to solve the above drawbacks, it would be best to make the voiced/unvoiced judgment result indicate a third state rather than an alternative, but this would reduce the amount of transmitted information, which is a feature of the vocoder. I can't take advantage of it.

An object of the present invention is to provide a speech synthesizer that can obtain intermediate speech that is neither voiced nor unvoiced without increasing the number of bits output from the transmission side of the vocoder.

According to the present invention, in a speech synthesizer that has a pulse generator and a noise generator as an excitation source and a synthesis filter that inputs PAROOR coefficients as speech spectral information, a voiced/unvoiced discrimination signal is input, and voiced and unvoiced sounds are a crossing detector for detecting the crossing portion between the pulse generator, the output of the noise generator, the voiced/unvoiced discrimination signal, the output of the crossing detector, and the PARC! An OR coefficient is input, and a signal obtained by mixing the output of the pulse generator and the output of the noise generator at the mixing ratio controlled by the PAROOR coefficient is input as an excitation waveform to the synthesis filter in the above-mentioned part. A speech synthesizer equipped with a mixer is obtained.

Specifically, in the present invention, the number of past durations of voiced and unvoiced frames of the voiced/unvoiced discrimination signal is used to detect the difference. That is, if the sound has been voiced or unvoiced for n consecutive frames in the past, it is determined that the sound is stationary.

On the other hand, if it does not last for n frames, it is determined that it is the wrong part. Then, when it is determined that there is a difference, the pulse generator and noise generator are operated simultaneously and their outputs are mixed. The mixing ratio at this time is PARO
It is controlled by the parameter 1 in the OR coefficient Ki (i=1. . . , P). In other words, PARC! It is known that the value of 1 for the parameter of the OR coefficient changes clearly between voiced and unvoiced sounds, and the value is distributed between approximately 0.5 and 1 for vowels. fact.

It is known that setting this parameter to 1 can also be used to determine voiced/unvoiced.

In this way, in the transition between voiced and unvoiced sounds, the output of the pulse generator and the output of the noise generator are mixed according to the value of 1 for the parameter, and this mixed signal is used as the excitation waveform of the synthesis filter. By doing so, the two-synthesized speech becomes more natural.

Furthermore, by configuring as described above, there is no need to make any changes to the transmission side of the vocoder, so audio information (spectral information, pitch information, etc.) is not required.

In applications where voiced/unvoiced information, etc.) is temporarily stored in memory and read out sequentially to obtain synthesized speech, it is possible to perform synthesis with improved quality without re-entering or changing the data stored in memory. You can get audio.

The present invention will be described below with reference to nine drawings.

FIG. 1 is a block diagram showing the configuration of a conventional speech synthesizer. The pulse generator 10 has a pulse period of pinch information 2.
00, and the amplitude is controlled by the residual power 300. The amplitude of the noise generator 20 is controlled by the residual power 300.

Pulse signals 11. output from these pulse generators 10 and noise generators 20. The noise signal 21 is sent to a switch 30, and the switch 30 selects either the 1-pulse signal 11 or the noise signal 21 based on the voiced/unvoiced discrimination signal 100.
The output 61 of the switch 3o is sent to the synthesis filter 40 as an excitation waveform. The filter characteristics of the synthesis filter 40 are as follows.

FIG. 2 is a block diagram showing the configuration of an embodiment of the speech synthesizer according to the present invention, and the same symbols as in FIG. 1 indicate the same configuration. The voiced/unvoiced discrimination signal 100 is sent to the crossing detector 50.

A crossing detection signal 51 output from the crossing detector 50 is sent to a mixer 60. In the mixer 60, one of the pulse signal 11 and the noise signal 21 is input as an excitation waveform to the synthesis filter 40 in the same way as in the conventional method except for the crossing part, and the crossing part is controlled by the parameter 1 of the PARCOR coefficient 400. A signal obtained by adding the pulse signal 11 and the noise signal 21 at the mixing ratio set is input to the synthesis filter 40 as an excitation waveform.

FIG. 3 is a circuit diagram showing a specific example of the crossing detector 50 shown in FIG. The voiced/unvoiced discrimination signal 100 is input to an n-stage shift register 50a.

The output of each stage of the n-stage shift register 50a is sent to the NOR circuit 50 and the AND circuit 5[IC. In the NOR circuit 50b and the AND circuit 50c,
It is detected whether the outputs of the n-stage shift register 50a are all "0" (unvoiced sound) or all 1" (voiced sound). The outputs of the NOR circuit 50b and the AND circuit 50c are N.
The signal is sent to an OR circuit 50d, where a portion where the output of the n-stage shift register 5Da is not all 0'' or all 1'' is detected, and a crossing detection signal 51 is obtained.

FIG. 4 is a detailed view of the mixer 60 of FIG.

The pulse signal 11 of the pulse generator 10 (FIG. 2) and the noise signal 21 of the noise generator 20 (FIG. 2) are.

Switcher 60 controlled by voiced/unvoiced discrimination signal 100
The signals are sent to multipliers 60c and 60d which are respectively controlled by one and the other input terminals of a and outputs f (0≦f≦1) and (1-f) of a nonlinear circuit 601).

The nonlinear circuit 60a inputs the PARCOR coefficient 400,
'Calculate the mixing ratio by setting the parameter in the PAROOR coefficient 400 to 1. two multipliers 60c;

The outputs of 60d are added together by adder 60e.

The output of the adder 60e is sent to the crossing detection circuit 50 (FIG. 2).
The signal is sent to one input terminal of a switch 60f controlled by the crossing detection signal 51. The output of the switch 60a is input to the other input terminal of the switch 6Of. Therefore, the output of the switch 6Of, that is, the output 61 of the mixer 60,
At the crossing part, the pulse signal 11 and the noise signal 21 are P
A signal mixed at a mixing ratio controlled by the parameters of the ARCOg coefficient 400 appears, and except for the portions outside, either the pulse signal 11 or the noise signal 21 appears as in the conventional case.

FIG. 5 is a diagram showing the characteristics of one output f (control signal of multiplier 60C) of nonlinear circuit 60b in FIG. 4. If the value of 1 for the parameter of PAROOR coefficient 40[1 input to the nonlinear circuit 60b is from 0 to α (for example, α=0.5), continuous values from 0 to 1 are output, and the value of 1 is output for the parameter. If the value is 0 or less, the value of (1-f) is the multiplier 60d.
output as a control signal.

In addition, in the above embodiment, for convenience of explanation, a case was described in which it was realized by hardware.

Needless to say, this can also be achieved using computer software.

As is clear from the above description (=9), the present invention has the effect of providing a speech synthesizer that improves the quality of the pedestal speech in the transition portion of the speech and makes the 1-synthesized speech more natural.

Remaining days below

[Brief explanation of drawings]

FIG. 1 is a block diagram showing the configuration of a conventional speech synthesizer.
FIG. 2 is a block diagram showing the configuration of an embodiment of the speech synthesizer according to the present invention, FIG. 6 is a circuit diagram showing a specific example of the crossing detector shown in FIG. 2, and FIG. 5 is a detailed diagram of the mixer shown in the figure, and FIG. 5 is a diagram showing an example of the characteristics of one output of the nonlinear circuit shown in FIG. 4. 10...Pulse generator, 11...Pulse signal. 20... Noise generator, 21... Noise signal, 60...
・Switching device, 61...Switching device output, 40...Synthesizing filter. 50... Crossover detector, 50a... n-stage shift register, 50b... NOR circuit, 50C... AND
circuit. 5 [], i...NOR circuit, 51... Crossing detection signal. 60...Mixer, 60a...Switcher, 60b...
Nonlinear circuit, 60c, 61M-multiplier, 60e-adder, 60f...switcher, 61...mixer output, 1
00...Voiced/unvoiced discrimination signal, 200...Pitch information, 300...Residual power, 400...PARCO
R coefficient, 500...Synthesized voice output. Figure 1 Figure 4

Claims (1)

[Claims] 1. In a speech synthesizer that has a pulse generator and a noise generator as an excitation source and a synthesis filter that inputs PARCOR coefficients as speech spectrum information, a voiced/unvoiced discrimination signal is input to detect the transition between voiced and unvoiced sounds. A crossing detector for detecting the crossing part, the output of the pulse generator, the output of the noise generator, the voiced/unvoiced discrimination signal, the output of the crossing detector, and the PAROOR coefficient are inputted, and in the crossing part, and a mixer that pulses the output according to a mixing ratio controlled by the PAROOR coefficient.