JPS58108597A - Synthetic voice generator - 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 synthetic sound waveform generation device that generates synthetic sounds such as voices and musical tones.

When generating synthesized sounds such as voices and musical tones by digital signal processing, a certain sampling period (Ts') is used.
Each sample value is obtained by digital calculation, and this is digital-to-analog converted (hereinafter referred to as DA conversion) to generate a synthesized sound.

In a well-known example, a digital synthesis filter having a transfer characteristic that simulates the spectral envelope of the target synthesized sound has a sample value that repeats the same sample value for every number of samples (denoted as N/). Input as the sound source waveform,
An unsynthesized sound has been generated by using the output of the digital synthesis filter as a sample value sequence of the target synthetic sound. In this case, the pitch period (referred to as Tp') of the synthesized sound is expressed by the following equation (1), and conventionally, the pitch synchronization Tp was changed by changing the number of samples g.

Tp'=N'・Ts' (1) According to this conventional example, a constant sampling period (Ts')
Only a synthesized sound with a pitch period (Tp) times the number of samples (N') can be generated.

In other examples, the pitch period (Tp') of the target synthesized sound
The number of samples (N') corresponding to the time length (i.e., '・TI)', which is an integer (k') of
Synthetic sounds were generated by storing a sequence of sample values (hereinafter referred to as elemental waveforms) and using the elemental waveforms repeatedly.In this case, the pitch synchronization of the synthesized sounds (
TI) is expressed by the following equation (2), and conventionally, the pitch-period (Tp') can be changed by using the number of samples N' or by using a different element waveform for an integer.

I Tp/= (kz)・Ts' (2) According to this example, a certain sampling synchronization (an integer multiple or an integer of a certain sampling synchronization with a pitch period (TI) having a time length that is an integer ratio with Ts) Only pitch-synchronized synthetic sounds with comparable time lengths can be generated, and it is difficult to generate synthetic sounds with arbitrary pitch periods.In the second example above, if ′ is set to a large number in N′%, Although a synthesized sound with a pitch period close to the desired pitch period can be obtained to some extent, increasing the number of samples N' increases the storage capacity for storing sample values, which is not a good idea.

If the desired pitch phase cannot be accurately generated, it becomes a problem when synthesizing songs, musical tones, etc., since it is not possible to generate melodies with accurate pitches. For example, for a certain number of samples, set sampling synchronization so that the sound of Tedo can be generated accurately.
The sounds of 7 and shi cannot be produced accurately because they are not in integer ratio with the sound of do.

SUMMARY OF THE INVENTION An object of the present invention is to provide a synthetic sound generation device that is capable of generating a synthetic sound having a relatively accurate pitch period.

According to the present invention, there is provided a means for generating variable sampling period data, a means for generating a sample value sequence having variable sample values, and a means for controlling the sampling period data and the number of samples to obtain a desired pitch ring synchronization ( Ts)
and the number of samples (N) or its integer (k
By controlling the pitch period so that 1/2 of the pitch period becomes the desired pitch period, a synthesized sound with a relatively accurate pitch period can be generated. This temporal formula holds true.

Tp = k-Ts (3) This formula (
3) is the same as equation (2) in form 1 or equation (1) if k is set to 1, but in equations (1) and (2), sampling synchronization T, / is just a constant, Equation (3) differs in that the sampling synchronization Ts is a variable that can be controlled simultaneously with the number of samples N.

Next, the present invention will be explained in detail using the drawings.

First, data indicating the pitch period of the target synthesized sound is input from the pitch period data input terminal l to the time control data generation circuit 3.
The 6-hour control data generation circuit 3 inputs sampling synchronization data (formula (3) ) corresponding to Ts of ), the number of samples (equation (3)
(corresponding to N) are output respectively. The sampling clock signal generation circuit 4 generates a sampling clock signal of a sampling period according to the data from the time control data generation circuit 3, and transmits the sampling clock signal transmission line 5.
Via the sound source waveform generation circuit 6, the synthesis filter? , D.A.
It is sent to the conversion circuit 8. The sound source waveform generation circuit 6 sequentially and sequentially generates sample values of several sound source waveforms according to the data from the time control data generation circuit 3 every time a clock signal is input from the sampling clock signal generation circuit 4. Send it cyclically to the synthesis filter.

On the other hand, from the synthesis filter control data input terminal 2, data for controlling the transfer characteristics of the synthesis filter is input to the synthesis filter. The synthesis filter 7 filters the sound source waveform input from the sound source waveform generation circuit 6 and outputs it to the DA conversion circuit 8 in synchronization with the sampling clock sent from the sampling clock signal generation circuit 4 according to the control data from the terminal 2. The DA conversion circuit 8 converts the sample value of the waveform outputted from the synthesis filter 7 into digital/analog in synchronization with the sampling clock sent from the sampling clock signal generation circuit 4 and outputs it to the synthesized sound output terminal 9. .

Note that although this example applies the present invention to a type of synthesized sound generation device that uses a synthesis filter, the present invention can also be applied to a type of synthesized sound generation device that uses segmental waveforms. It is clear that this can be done.

[Brief explanation of the drawing]

The figure is a block diagram showing an embodiment of the present invention. In the figure,

Claims (1)

[Claims] In a synthetic sound generation device that generates and outputs a synthetic sound waveform of voice, musical tone, etc. every sampling period using digital signal processing, means for generating variable sampling period data and a sample value sequence with a variable number of samples are provided. 1. A synthetic sound generation device comprising: generating means; and means for simultaneously controlling the sampling period data and the number of samples to generate a synthetic sound waveform having a desired pitch period.