JPH0358097A - Determination system for pause insertion position - Google Patents

Abstract

PURPOSE:To determine a pause insertion position by using a numeral corresponding to a combination of parts of speech of a word and the length of the word by storing numerals indicating the easiness of pause insertion by combinations of parts of speech of words before and after word borders. CONSTITUTION:A character string to be converted into a voice is inputted from a character string input terminal 11 and a morpheme analysis part decomposes a sentence into word strings and determines the part of speech and the rendering of each word. The word strings having the parts of speech and the rendering are sent out to a word length calculation part 13 and a pause insertion position determination part 15. A word length calculation part 13 calculates the length of each word constituting the sentence by using the rendering of the word and sends it to the pause insertion position determination part 15. A pause insertion probability storage part 14 is stored with pause insertion probability previously by combinations of parts of speech of words before and after borders. The pause insertion position determination part 15 uses the pause insertion probability and the length of the word to determine the natural insertion position of a pause.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a pause insertion position determination method for determining the position at which a pause is to be inserted in a regular voice gathering for converting a character string into speech.

(Prior art) In regular speech synthesis that converts arbitrary sentences into speech,
It is necessary to insert a pause of an appropriate length at an appropriate position in a sentence, and is important in improving the naturalness of the synthesized speech. In other words, there is a limit to how long a human voice can be uttered in one breath, and the speaker pauses at the boundaries of phrases or words that have weak connections in terms of meaning, either to take a breather or to convey a break in meaning to the listener. put

Therefore, even when producing audible speech, if appropriate pauses are not inserted, the listener will feel that it is unnatural.

Conventionally, the strength of the connection between adjacent bunsetsu has been used to determine the pose insertion position.

In other words, it utilizes the property that the weaker the connection between the preceding and following clauses in terms of meaning, the easier it is to insert a pause.

First, the strength of the connection between bunsetsu and bunsetsu boundaries is expressed by the number of bunsetsu from the preceding clause to the receiving clause, and this measure is called the degree of separation. A large value of separation means that a certain clause is connected to a clause that is further away.
This indicates that the connection with adjacent clauses is weak. Therefore, we believe that there is a high possibility that a pause will be inserted at a bunsetsu boundary with a high degree of separation. Furthermore, humans have a limit to how long they can vocalize in one breath, so the speaker pauses at an appropriate position to catch their breath. Therefore, it is considered that pauses are likely to be inserted when the total number of moras of clauses before and after a clause boundary is large. Based on the above, the product of the degree of separation and the total number of moras is used as an evaluation measure for determining whether to insert a pause, and if this exceeds a certain predetermined threshold, a pause is inserted. shall be taken as a thing.

Regarding the method of determining such a pause insertion position, please refer to Acoustical Society of Japan Speech Research Group Material 878-07 (1978-4).
Examination of intonation rules for F-sentence speech] (Reference 1).

(Problem to be Solved by the Invention) In the conventional method, it was necessary to accurately analyze the dependency relationships between the clauses that make up the sentence. 1- However, it is difficult to always accurately analyze this dependency relationship, and as a result, there is a problem that the selected pose insertion position may become unnatural.

In contrast, an object of the present invention is to provide a pose insertion position determining method that can determine a natural pose insertion position without analyzing such dependency relationships.

(Means for Solving the Problems) The present invention decomposes a sentence represented by an input character string into words that disturb it, and determines whether to insert a pause at the boundary of each decomposed word. In the pause insertion position determination method, a numerical value indicating how easily a pause is inserted at a word boundary is memorized for each combination of parts of speech of words before and after the word boundary. The method is characterized in that it is determined whether or not to insert a pause at the word boundary using a numerical value corresponding to a combination of parts of speech of the word in the word boundary and the lengths of the words before and after the boundary.

(Operation) A pause is likely to be inserted at a word boundary where the meanings of the words before and after the boundary are weakly connected. In the conventional method, the strength of this connection was calculated from the dependency structure of the sentence. On the other hand, when examining the parts of speech of words before and after the boundary where a pause is inserted, it is found that there are parts of speech in which pauses are easily inserted and parts of speech in which it is difficult to insert.

Therefore, by utilizing such properties of parts of speech, it becomes possible to determine a natural pause insertion position without using the dependency structure of sentences.

Therefore, we first checked the parts of speech of the words before and after the boundary where the pause was inserted from the audio of several sentences read aloud by humans (hereinafter referred to as the sentence audio database).
Find out how easily a pause is inserted for each combination of parts of speech. As the ease with which a pause is inserted, for example, the ratio of occurrences in which a certain combination of parts of speech is accompanied by a pause to all occurrences of a combination of parts of speech in the sentence speech database may be used. Below,
This ratio will be called the pause insertion probability. When considering combinations of parts of speech, there are classification methods such as nouns, verbs, and adverbs. In addition, it is also useful to classify parts of speech that are inflected, such as verbs, adjectives, and auxiliary verbs, into conjugated forms such as natural form, conjunctive form, and adjunctive form.

Also, as a method to actually determine the pose insertion position,
For example, a method as shown in (Reference 1) can be considered.

That is, a value calculated from the pause insertion probability at a certain word boundary and the length of words before and after the boundary (for example, the number of moras), such as the product of the pause insertion probability and the number of moras, is used as an evaluation measure. A threshold is set for this evaluation scale, and depending on whether the evaluation scale at each boundary exceeds the threshold, it is determined whether or not to insert a pause at the boundary.

Alternatively, the number of pauses to be inserted may be determined first based on the length of the sentence to be converted into speech, and the pauses may be inserted in descending order of the probability of pause insertion.

By using the above method, it becomes possible to determine a natural pause insertion position by simply using the part of speech of the word obtained through morphological analysis, without using the dependency structure of the sentence to be converted into speech. .

(Embodiment) FIG. 1 is a block diagram showing an embodiment for realizing a pose insertion position determination method according to the present invention.

First, input the character string to be converted into speech into the character string input terminal 1l.
Enter from. The input character string is sent to the morphological analysis unit l
2, the sentence represented by the input character string is broken down into word strings, and the part of speech and pronunciation of each word are determined. This word string with part of speech and pronunciation is sent to the word length calculation section 13 and the pause insertion position determination section 15. When the word string is given, the word length calculation section 13 calculates the length of each word (for example, the number of moras) that makes up the sentence using the pronunciation of the words, and sends it to the pause insertion position determination section 15. .

The pause insertion probability storage unit 14 stores in advance the pause insertion probability for each combination of parts of speech of words before and after the boundary. The pause insertion position determination unit 15 reads out the pause insertion probability stored in the pause insertion probability storage unit 14, which corresponds to the combination of parts of speech of words before and after each boundary, according to the word string with the part of speech. . Using the read pause insertion probability and the length of each word calculated by the word length calculation unit 13, determine the position where the pause is inserted by the method described in the (effect) section, The result is output from the pause insertion position output terminal 16.

(Effects of the Invention) As described above, the present invention determines the pose insertion position without using the results of dependency analysis, which is difficult to analyze correctly. Therefore, it is possible to determine a more accurate and natural pose insertion position than the conventional method. Therefore, the present invention is effective as a pause insertion position determination method in a voice command device or the like that converts an arbitrary sentence given as a character string into voice.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention. In the figure, 12... Morphological analysis unit, 13... Word length calculation unit, 14... Pause insertion probability storage unit, 15...
・This is a pose insertion position determination unit.

Claims (1)

[Claims] In a pause insertion position determination method that decomposes a sentence represented by an input character string into its constituent words and determines whether to insert a pause at the boundary of each decomposed word,
A numerical value indicating how easily a pause is inserted at a word boundary is memorized for each combination of parts of speech of words before and after the word boundary, and a numerical value corresponding to the combination of parts of speech of words before and after each word boundary is stored. and the lengths of words before and after the boundary to determine whether or not to insert a pause at the word boundary.