JPS5870295A - 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 capable of electronically switching regeneration frequencies.

Generally, when analyzing audio, the sound quality deteriorates as the analysis frequency (sampling frequency) is lowered. vice versa,
As the number of sampling points decreases, the data storage capacity decreases. As described above, the relationship between the sound quality of voice and the data capacity when chimpling is generally proportional, but the proportional relationship between individual voices is not necessarily the same. For example, even if voice A has good sound quality at a sampling frequency of 8KHz, voice B has the same sampling frequency of 8K) (the deterioration in sound quality is noticeable at z, and conversely, in order to obtain the same quality as voice A, it is necessary to There are many phenomena that require high frequencies.

Therefore, when outputting audio, if the reproduction frequency is set to a single frequency, it is necessary to set the reproduction frequency to a high one in order to obtain good sound quality for all audio, and in this case, the problem is that the data capacity increases. There is.

Conversely, when the data capacity is limited, there is a problem in that the sound quality of individual voices varies.

The present invention has been made in view of the above-mentioned conventional points.In particular, in speech synthesis, individual speech units (here, "speech unit" is any degree determined at the time of speech analysis, and in some cases, phoneme units, word units) etc.), the optimum reproduction frequency information (for example, two types, 8 KHz and 16 KHz, and the data is represented by 1 bit, and if the data is expressed as 16 KHz)
+ - "O" as digital information indicating 8KHz playback) is stored in the memory in digital format, and by performing waveform synthesis according to the above playback frequency information, the playback frequency is specified for each voice and the sound quality is made uniform. It is an object of the present invention to provide a speech synthesis device that can reduce data capacity.

Hereinafter, one embodiment of the device of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing one embodiment of the present invention, and FIG. 2 is a block diagram showing another embodiment.

In FIG. 1, 1 is a memory for storing data to be output, for example, a ROM can be used, and this ROM stores data IA indicating the distinction of reproduction frequencies and compressed audio data 2. There is.

In the figure, the case of one voice is shown for simplicity, but N
If there are voices, the set of data IA and IB is ROMI.
There will be N groups within. The audio data IB is subjected to the inverse compression process in the decoder 2 and becomes waveform data. At this time, with this device, the reproduction frequency is 8KHz, 16KHz.
The case where either KHz is adopted is shown, but D/, converter 4
Since all of the data is taken in as 16 KHz data, if the data IA indicates 8 KHz reproduction, the decoder 3 performs processing to repeat the same data twice and sends it to the % converter 4. When indicating 16KHz reproduction, the data from the decoder 2 is sent directly to the converter 4. The data converted into an analog signal by the converter 4 is sent to an output circuit 5 including an amplifier and a filter system. ) is entered. In this embodiment, since there are two types of reproduction frequencies, 8KHz and 16KHz, data I indicating the distinction of reproduction frequencies in the ROMI is
A is represented by an l pit (i.e., 111 I+ indicates 1'6KHz + "O" indicates 8KHz playback).
, the playback frequency is 8KHz.

In the case of three or four types such as 12 KHz and 16 KHz, this can be handled by setting the data IA in the ROM to 2 bits. Furthermore, by increasing the number of pits in the data IA, it is possible to support a larger number of playback frequencies, but this results in limitations such as an increase in data capacity and complicated hardware.

In order to solve the above problem, the method shown in FIG. 2 can be adopted. In other words, in the case of Fig. 1, the drive clock frequency of the converter 4 is fixed at 16KHz, so if the reproduction frequency is 8KHz, the decoder 3 outputs 16KH.
Although the data is converted into z data, it is possible to correspond to different frequencies by separately controlling the clock frequency of the converter 4. Specifically, in FIG. 2, when the data IA indicates a playback frequency of 8KHz, the clock frequency of the clock pulse generator 7 is set to 8KHz, and when the playback frequency is 16KHz, the clock frequency is set to 16KHz.
In this case, the decoder 3 (FIG. 1) can be made unnecessary. In the above embodiment, the playback frequency is 8 KHz taking into consideration hardware constraints.

Although switching between two types of reproduction frequencies of 16 KHz is performed, it goes without saying that this can be implemented in various ways.

As explained above, according to the present invention, desired playback frequency information for each audio unit is stored in a memory in digital format, and waveform synthesis is performed according to the playback frequency information, thereby achieving uniformity of sound quality level and data storage. The amount can be reduced.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the speech synthesis device according to the present invention, and FIG. 2 is a block diagram showing another embodiment.1 In the figure, 1: memory, 2.3: decoder, 4: Converter, 5: Output circuit, 6.7: Clock frequency generator. Agent Patent Attorney Aihiko Fukushi

Claims (1)

[Claims] 1. In speech synthesis, desired playback frequency information is stored in a memory in digital format for each audio unit, and the playback frequency is switched by performing waveform synthesis based on the playback frequency information stored in the memory. A speech synthesis device characterized by: