JP2513266B2 - Phoneme duration determination device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The phoneme duration determination device of the present invention is particularly applicable to a phoneme duration determination device in a rule speech synthesizer that controls pitch frequency and phoneme duration according to rules. Regarding

[Conventional technology]

When synthesizing voices having different utterance speeds, this type of conventional device controls the utterance speed by expanding and contracting the duration of the standard utterance speed at a fixed rate.
Such a conventional technique is described in detail in "Formant, CV-VC type rule composition" etc. of Material S85-31 (1985-07) of the Acoustical Society of Japan Material.

FIG. 2 is a diagram for explaining such a conventional technique. In order to change the voice speed, a constant expansion / contraction rate α for the entire utterance is taken into consideration without considering the structure of a sentence or the part of speech of a word.
The phoneme duration is expanded and contracted with.

[Problems to be Solved by the Invention]

However, if the entire utterance is uniformly changed as in the above-described conventional apparatus, when the utterance speed changes, the meaning of the sentence, the will of the speaker, and the like are not always transmitted to the listener well.

On the other hand, the present invention provides a phonological duration that makes it possible to generate a synthesized voice having a phonological duration that allows the listener to accurately convey the meaning and content of a sentence even when the utterance speed is changed. It is intended to provide a length determining device.

[Means for solving the problem]

The phoneme duration determination device of the present invention is a unit for inputting a sentence to be converted into speech, a unit for inputting part-of-speech information of words constituting the sentence to be converted into speech, and a unit for inputting a speaking rate. A means for determining the duration length for each word or phrase unit that composes an input sentence according to the input speaking rate and the input part-of-speech information, and the word or phrase from the determined duration time for each word or phrase unit. And a means for determining the duration of each phoneme composing the phrase.

[Action]

When a human reads a certain sentence or conveys information to a listener, the importance of each word or phrase constituting the sentence changes depending on the structure of the sentence and the will of the speaker. Also, words or phrases having important information in one sentence tend to be pronounced clearly to distinguish them from other words or phrases. This tendency is maintained even when the utterance speed changes. Therefore, when the utterance speed is changed, it is considered that the expansion / contraction rate of the duration varies depending on the part of speech of the word, that is, the role of the word in the sentence. Therefore, the phoneme duration is determined using a two-stage method as described below.

First, the speaking rate of the entire sentence and the part-of-speech information of the words that compose the sentence are input. As the part-of-speech information, nouns, verbs, adjectives, etc., or noun phrases, verb phrases, adjective phrases, etc. can be considered. The expansion / contraction rate is calculated for each word or phrase from this part-of-speech information and the utterance speed, and the utterance speed or duration time is determined for each word or phrase.

Next, based on the duration of this word or phrase, the duration of each phoneme that constitutes the word or phrase is determined. For example, a method of determining the duration of each phoneme so as to match the duration of a word or a phrase can be considered. At this time, the factors that determine the duration of each phoneme include the type of phoneme before and after, the presence or absence of an accent nucleus, and the position within a unit called a breath stage called a breath.

By using such a phoneme duration determination method, it is possible to generate a natural phoneme duration even when the utterance speed is changed.

〔Example〕

FIG. 1 is a block diagram showing one embodiment of a phoneme duration determining apparatus according to the present invention.

A character string representing a sentence to be converted into voice and part-of-speech information of words constituting the sentence are input from a character string input terminal 11 and a part-of-speech information input terminal 12, respectively. Further, the speaking speed of the entire sentence is input from the speaking speed input terminal 13. The input of the utterance speed can be performed by a method such as designating in units of speed such as the number of mora per unit time or selecting from several predetermined utterance speeds.

The phrase duration determining unit 14 determines the duration of each word or phrase based on the input utterance speed and the input part-of-speech information. As a method of this determination, for example, a method of obtaining a duration time at a desired utterance speed based on a scaling factor α with respect to the duration time at a standard utterance speed, or the like can be considered.

When obtaining the phoneme duration of the word W whose part of speech is PS (noun, verb, adjective, etc.) at the utterance rate SR, the expansion / contraction rate for the phoneme duration at the standard utterance rate is α
_{Letting W} be α _{W, it} is obtained by a function f of the part of speech PS and the speaking rate SR. That is, α _W = f (PS, SR) In the phoneme duration determination unit 15, based on the input character string and the duration length of the word or the phrase unit determined by the phrase duration determination unit 14, Determine the duration of the individual phonemes that make up a word or phrase. The determined phoneme duration is the phoneme duration output terminal.
It is output from 16.

〔The invention's effect〕

As described above, according to the present invention, it is possible to generate a natural phoneme duration even when the utterance speed is changed, and it is necessary to output synthetic speech with various utterance speeds. This is effective in realizing a phoneme duration determination device used in a response device or the like.

[Brief description of drawings]

FIG. 1 is a block diagram showing an embodiment of a phoneme duration determining device according to the present invention, and FIG. 2 is a diagram for explaining a conventional technique. In the figure, 11 is a character string input terminal, 12 is a part-of-speech information input terminal, 13 is a speech rate input terminal, 14 is a phrase duration determining unit, 15 is a phoneme duration determining unit, and 16 is a phoneme duration output terminal. Is.

Claims (1)

(57) [Claims] 1. A device for determining a phoneme duration for synthesizing voices of various utterance speeds, means for inputting a sentence to be converted into voice, and part-of-speech information of words constituting the sentence to be converted into voice. Means for inputting a speech speed, means for inputting a speech rate, and means for determining a duration length in units of words or phrases constituting the input sentence according to the input speech rate and the input part-of-speech information. And a means for determining the duration of each phoneme that composes the word or phrase from the determined duration of the word or phrase unit.