JPH01186998A - Sentence intonation processing method for voice synthesizing device - Google Patents

Abstract

PURPOSE:To obtain a synthesized voice having an intonation suitable for a Japanese language by distributing a text to subjects and predicates the divide the text into clause units and determining a fundamental intonation and a word accent width in accordance with the number of clauses and positions of clauses with respect to each of subjects and predicates. CONSTITUTION:A sentence analyzer 5b refers to a grammar dictionary 6 and a word dictionary 7 to divide a Japanese-language input text into clause units and distributes respective clauses to subjects and predicates and determines a fundamental intonation set in a table 8 in accordance with number of clauses and clause positions with respect to each of subjects and predicates and determines a word accent width set in a table 9 in accordance with the number of clauses and clause positions with respect to each of subjects and predicates. These fundamental intonations and word accents to which word accent widths are given are superposed to obtain a sentence intonation, thereby obtaining a sentence intonation in accordance with fundamental intonation patterns and accent widths given to word accent positions. Thus, metrical information suitable for the Japanese language is obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application The present invention relates to a speech synthesis device using a rule synthesis method, and particularly to a processing method for sentence intonation.

B1 Summary of the Invention The present invention, in obtaining prosodic information of a text from sentence analysis, divides the text into a main part and a predicate part and divides it into clause units,
By determining the basic intonation and word accent width based on the number of clauses and clause positions for each subject and predicate, it is possible to obtain synthesized speech with appropriate intonation while reducing dictionary capacity and simplifying sentence intonation processing. It is something.

C1 A speech synthesis apparatus based on the conventional technical rule synthesis method has, for example, the configuration shown in FIG. The text analysis unit 1 performs text analysis on a character string of Japanese input text using a dictionary 1a and a text analysis device 1b. In addition to a dictionary for converting the pronunciation of words, the dictionary 1& has a Japanese grammar dictionary for dividing words into clauses and phrases, and also has a dictionary of rules for word accents and basic intonation. The text analysis device 1b converts the input text into a phoneme symbol string of phonemes or syllables by referring to the dictionary 1a, and also generates prosodic information such as word accents and basic intonation.

The speech synthesis rule section 2 is composed of a file 2a and a parameter generation device 2b. The file 2a stores feature parameters of phoneme units, their connection rules, and prosodic information control rules. The parameter generation device 2b generates a control parameter sequence in which feature parameters for phoneme information are linked together with information such as connection time, and also generates a sound source pattern sequence in which pitch, energy, and intonation processing of the sound source is processed based on prosody information.

The speech generation section 3 is composed of a sound source generation device 3a, a speech synthesis digital file 3b, and a speech converter 3c. The sound source generation device 3a generates sound source signals such as pitch and energy according to the sound source pattern sequence. The digital filter 3b changes the parameters of the Percoll coefficient, transfer function, or formant frequency according to the control parameter string, and obtains a synthesized speech data string as a response output to the sound source signal with these parameters. The audio converter 3c is a filter 3b
The output of the converter is converted into an analog signal to obtain an audio waveform, and synthesized audio is output by an electrical pre-conversion means such as a speaker.

In the speech synthesis apparatus as described above, prosody information is created as one sentence intonation by combining basic intonation, word accent, stress, pause, duration, etc. for sentence data. To create this sentence intonation, input the Japanese input text into the dictionary 1.
It is converted into a phonetic sequence with word series segmentation and word boundaries added by segmentation into nine phrases and one sentence, morpheme (the smallest unit that has linguistic meaning) classification, etc. by syntactic analysis with reference to a. For this phonetic string information, basic intonation determined from the number of clauses (mora) in punctuation mark units, accents in word units, pauses in phrase units, stress in words determined from prefixes and suffixes, and phoneme unit The duration time etc. are determined.

D1 Problems to be Solved by the Invention In conventional sentence intonation processing methods, basic intonation and accent are determined using syntactic analysis and words, syllables, and even phoneme units, resulting in a huge number of analysis items and input The number of items to be processed, such as the length of the text and the number of accent types depending on the number of characters, is enormous, creating the problem of requiring complex processing.

In addition, in order to perform the conventional processing method in detail, it is necessary to determine not only the word accent but also the accent width (number of syllables). There were complications.

An object of the present invention is to provide a processing method that can define basic intonation and word accent width without complicating the processing.

61 Means and Effects for Solving the Problems In order to achieve the above object, the present invention divides the Japanese input text into clauses, divides each clause into a main part and a predicate, and divides each clause into a main part and a predicate. Determine the basic intonation set in the table from the number of clauses and clause position, and determine the word accent width set in the table from the number of clauses and clause position for the subject and predicate, respectively, and words given these basic intonation and word accent width. By calculating the sentence intonation by weighting the accent, we can calculate the sentence intonation from the basic intonation pattern of small combinations of the number of clauses and the position of the clause, and the accent width added to the word accent position. Try to match the prosody.

F. Embodiment FIG. 1 is a block diagram of a text analysis section showing an embodiment of the present invention. The dictionary 5a is a Japanese grammar dictionary 6 and a word dictionary 7 (
(accent, duration), an intonation table 8 for the main part and predicate of a sentence, and a word accent width setting table 9.

The sentence analysis device 5b refers to the grammar dictionary 6 and the word dictionary 7 to convert the readings of the words and divides the Japanese text into phrases and phrases, as in the past.

Analyzes sentences such as word accent positions. At this clause break and phrase break, the sentence analysis device 5b finds a clause boundary, and further finds a subject/predicate boundary where the clause is divided into a subject and a predicate based on the difference in particles.

For the subject and predicate of the sentence obtained in this way,
The sentence analysis device 5b determines the basic intonation for each subject and predicate by referring to the intonation table 8 based on the number of clauses and the clause position.

Furthermore, the sentence analysis device 5b determines the accent width of the word included in the subject and predicate based on the number of clauses and the clause position of the subject and predicate by referring to the accent width setting table 9.

Sentence intonation is determined by such a sentence analysis device 5b, that is, it is determined from the main part and predicate of the sentence, the number of phrases included in each, and the phrase position, and the sentence intonation is determined by adding pauses etc. Information is created. Note that the creation of prosody information is performed in the same manner as before.

This example will be explained in detail below.

In the intonation table 8 and the word accent width setting table 9, the basic intonation and word accent width of the pattern are set according to the combination of the number of phrases and the position of phrases for each subject and predicate, as shown in the table below.

Intonation Table Word Accent Setting Table For example, if the basic intonation is the number of clauses in the main part n and the clause position of the main part is in n, the intonation pattern is set to P, and the basic intonation is determined by the pattern P-. . In addition, the word accent width is set to the accent width Bml when the word belongs to a predicate, is included in a clause with the number of clauses m, and is in clause position 1, and the word accent is set to the accent width B at the accent position. It is determined.

Figure 2 shows an example in which the sentence intonation is calculated from the basic intonation and word accent width when the Japanese input text is ``The cherry blossoms at school bloomed beautifully.'' For the basic intonation, a pattern pat in which the number of clauses in the main part is 2 and the clause position is 1, and a pattern Q□ in which the number of clauses in the predicate is 2 and the clause position is 2 are read out from the table, and the basic intonation is obtained by combining both patterns. The word accent has an accent width of A where the main part of the word ``kakkou'' has 2 clauses in the main part and 1 clause position! l, the word ``Sakura'' has 2 clauses and the accent width is A in clause position 2, and the predicate word ``Kirei ni'' has 2 clauses and the accent width is B in clause position 1.
□, the word "saku" has the number of clauses 2 and the accent width B at the clause position 2 is obtained. These patterns and the accent width at the word accent position are superimposed, and the sentence intonation as shown in the figure is determined by filter processing that provides further smoothness.

Therefore, to determine sentence intonation, the basic intonation is determined using a table based on the decomposition of the Japanese input text into the subject and predicate, and the number and position of clauses for each subject and predicate. Similarly, word accents are also accented. The accent width is determined using a table based on the number of clauses of the word for each subject and predicate and the clause position. Therefore, the conventional process of increasing the number of analysis items such as word accent determination in syllable and phoneme units is unnecessary for sentence intonation processing, and the process becomes simple with a small number of basic intonations and word accent widths. In addition, for Japanese sentence intonation processing, which is complex and diverse, we can create sentence intonation based on the basic intonation based on the distinction between subject and predicate, the number of clauses, and the clause position, and the word accent width added to the accent position. It is suitable for prosody and can produce natural synthesized speech.

G1 Effects of the Invention As described above, according to the present invention, a Japanese input text is divided into a subject part and a predicate part in units of clauses, and the basic intonation pattern and word accent width are determined based on the number of clauses and the clause position, respectively. This has the effect of making it possible to obtain prosodic information suitable for Japanese while simplifying sentence intonation processing.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram illustrating intonation waveforms of the embodiment, and FIG. 3 is a block diagram of a speech synthesis apparatus using a rule synthesis method. 5a...Dictionary, 5b...Sentence analysis device, 6...Grammar dictionary, 7...Word dictionary, 8...Intonation table, 9...Word accent setting table. Fig. 1 Block diagram of the embodiment 5i Dictionary 5b Sentence analysis device 6 Grammar dictionary 7 Word dictionary 8 Intonation table 9 Word accent setting table Fig. 2 Implementation Example intonation waveform diagram r-111 main part 11111 entered! “−＼−
--time

Claims (1)

[Claims] (1) Obtain phonological information and prosody information from the Japanese input text, obtain the control parameters of the digital filter corresponding to the phonological information and the sound source pattern of the sound source generator corresponding to the prosody information, and generate the audio signal corresponding to the text. In a speech synthesis device using a rule synthesis method, the Japanese input text is divided into clauses, each clause is divided into a main part and a predicate, and the basics are defined in a table based on the number of clauses and the clause position for each main part and predicate. Determine the intonation, determine the word accent width set in the table from the number of clauses and clause position for each subject and predicate, and superimpose these basic intonations and word accents given the word accent width to obtain the sentence intonation. A sentence intonation processing method for a speech synthesizer, characterized by: