JPH0756598B2 - Speech synthesis method of speech synthesizer - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a voice synthesizing system of a voice synthesizing apparatus for synthesizing an arbitrary vocabulary by combining phonological chains.

[Background of the Invention]

The conventional method of speech rate control of analysis and synthesis method is
Both vowel parts and consonant parts of synthetic speech varied in duration at a constant rate. However, when humans change the speaking rate, the rate of change in duration of vowel parts is usually larger than the rate of change in duration of consonant parts, which is not suitable for human speaking speed. It was

[Object of the Invention]

An object of the present invention is to make the speech rate variable in a speech synthesizer that synthesizes an arbitrary vocabulary.

[Outline of Invention]

In a voice synthesizer for synthesizing arbitrary vocabulary, the user must be comfortable enough to get used to the synthesized voice, and be familiar with it. I often want. Therefore, the present invention
In a speech synthesizer for synthesizing an arbitrary vocabulary by combining vowel phoneme chains, the utterance speed of synthetic speech is changed by changing the phoneme duration of consonant parts and the phoneme duration of vowel parts at different rates. It is a thing.

Example of Invention

An embodiment of the present invention will be described below with reference to the drawings. The speech synthesizer includes an input unit for receiving an input code, a chord phoneme conversion unit for converting the code into a phoneme symbol corresponding to a phoneme chain, a rule unit for reading a rule parameter corresponding to the phoneme symbol from a rule parameter file, and a phoneme. The phonological chain corresponding to the symbol is read from the phonological chain file 8, and the phonological unit 4 for performing the combining process, the voice synthesizing unit 5 for synthesizing the voice from the acoustic parameters, the voice output unit 6 for outputting the synthesized voice, and the whole of them are controlled It is composed of the control unit 9.

In the present embodiment, it is assumed that "YKHAMA" is input as the input code. The input code is converted into a phoneme code string corresponding to (Y, K, HA, MA) and a phoneme chain in the code phoneme conversion unit 2. Next, rule part 3
Then, the rule parameter corresponding to the phonological symbol is read from the rule parameter file 7, and the rule parameter includes the consonant duration sequence (TC ₁ , TC ₂ , TC ₃ , TC ₄ ) and the vowel duration sequence (TV ₁ , TV ₂ , TV ₃ , TV ₄ ) are included. Phonological unit 4
Then, the phoneme chain corresponding to the phoneme symbol is read from the phoneme chain file 8 and the consonant duration sequence (TC ₁ , TC ₂ , TC ₃ , TC ₄ )
Depending on the value of the vowel duration sequence (TV ₁ , TV ₂ , TV ₃ , TV ₄ ), the acoustic parameters of the consonant part and vowel part of each phoneme chain are cut or extended, and each phoneme chain is combined. The voice synthesizer 5 synthesizes voices from the combined acoustic parameters, and the synthesized voices are output from the voice output unit 6.

According to the present invention, in the above process, using the speech rate control coefficients a ₁ and a ₂ , TCi = a ₁ · TCi (i = 1, ..., 4) TVi = a ₂ · TVi (i = 1, ... , 4), the consonant duration and vowel duration are changed, and the utterance speed of synthetic speech is changed. Here, a ₁ and a ₂ are constants determined by the set speech rate.

〔The invention's effect〕

According to the present invention, since the speech rate of the synthesized speech of the speech synthesizer can be changed, the user can obtain the synthesized speech with the speech rate that is easy to use.

[Brief description of drawings]

FIG. 1 shows an embodiment of the present invention and is a block diagram of a speech synthesizer. DESCRIPTION OF SYMBOLS 1 ... Input part, 2 ... Chord phoneme conversion part 3 ... Rule part, 4 ... Phoneme connection part 5 ... Speech synthesis part, 6 ... Speech output part 7 ... Rule parameter file 8 ... Phoneme chain file 9 ... Control part

Claims (1)

[Claims] 1. An input unit for receiving an input code, a phoneme conversion unit for converting the code input from the input unit into phoneme symbols, and a phoneme after processing the phoneme symbols into vowel / consonant chains according to a predetermined rule. In a voice synthesizing device for synthesizing an arbitrary vocabulary corresponding to an input code, the voice synthesizing unit includes a plurality of voice synthesizing units for synthesizing a plurality of voice synthesizing units, a voice output unit for outputting the synthesized voice, and a control unit. A vowel duration and a rule parameter file for accumulating a plurality of consonant durations independent of the vowel duration, and a rule section for selecting and holding any vowel duration and consonant duration from the rule parameter file A phoneme chain file accumulating a plurality of phoneme chains, and a phoneme chain corresponding to the phoneme symbol from the phoneme chain file, and a mother of the selected phoneme chain. And a consonant duration that changes the consonant duration according to the vowel duration and the consonant duration held in the rule portion, and a phonological concatenation portion that concatenates a phoneme chain, and an arbitrary vocabulary corresponding to an input code. A speech synthesizing method for a speech synthesizing device, which is a speech synthesizing unit having a variable speech rate of synthetic speech, which independently changes the vowel part consonant phoneme duration and the consonant part phoneme duration when synthesizing.