JPH06332489A - Generating method of accent component basic table for voice synthesizer - Google Patents

Abstract

PURPOSE:To obtain easily understandable synthesized voice by controlling accents in multiple steps. CONSTITUTION:A basic accent pattern generating and processing section 6 is given the data, to which accents are added by a Japanese processing section 2, against the inputted text by a text inputting section 1. The data are read from a basic accent pattern table 7, processed by the section 6, corrected and processed by a correction processing section 8, added with the output of a phrase pattern computing section 3A and voice-synthesized by a voice synthesizer section 4.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention realizes a pitch pattern closer to human utterance by providing a plurality of pitch target values per phoneme in controlling the pitch of synthetic speech and controlling the accent width in multiple stages. The present invention relates to a method of creating an accent component basic table of a voice synthesizing device which is referred to when a basic pattern of an accent component is generated in a regular voice synthesizing device.

[0002]

2. Description of the Related Art As shown in FIG. 3, a speech synthesizing apparatus using a rule synthesizing method performs a morphological analysis by a Japanese processing unit 2 on a text input from a text input unit 1 to find a position of a pose and a phrase of a word or phrase. The phonetic conversion and accent addition are performed by cutting and referring to a dictionary. For example, when the input text is "Today's weather is nice", the Japanese processing result is shown in the table below.

[0003]

[Table 1]

For the text resulting from this processing, in the intonation control unit 3, the phrase pattern calculation unit 3A obtains a phrase component (in pauses) obtained from the number of mora (consonant + minimum unit of sound represented by vowels) included in the text. The pitch of the sound when a person is caught in a breath is calculated), and the accent component (the pitch of the sound of each word) is calculated by the accent pattern calculation unit 3B, and the respective components are overlaid as shown in FIG. A combined intonation control pattern is calculated. The phrase component gradually decreases from a high pitch at the beginning when a person speaks to a low pitch due to a decrease in subglottic pressure. As the accent component, one pitch target value is given for each analysis unit and linear interpolation is performed between them as described above, and three pitch target values are given for each analysis unit and linear interpolation is performed between them. .

In the latter accent component calculation, target values P _{1 to} P ₃ are given to the categories shown in Tables 2 and 3 below for each CV unit (consonant + vowel) and V unit (vowel), and the category indicated by * is added. The pitch is linearly interpolated from the preceding and following data.

[0006]

[Table 2]

[0007]

[Table 3]

The voice synthesis section 4 adjusts the pitch of each syllable according to the intonation control pattern given by the intonation processing section 3, and uses the source wave pattern and the parameters of the articulation filter associated with each syllable as a response output of the articulation filter. Get synthetic speech.

As described above, the conventional intonation control system adds intonation to the synthesized voice by synthesizing the phrase component according to the number of moras contained in the input text and the accent component for each mora. Of these, the accent component has a plurality of target pitches per mora, and the pitch change during that time is determined by interpolation processing.

[0010]

Since the pitch of the accent component is simply determined from the height of the accent,
It was easy to get a uniform pitch change and mechanical synthetic voice. In addition, since the connection between syllables and syllables and syllables and syllables is not considered at all, the change in accents before and after the mora may become unsmooth, resulting in unnatural synthetic speech as syllables or sentences as a whole. There was a problem.

The present invention has been made in view of the above circumstances. Accents are controlled in multiple stages rather than in two stages of high and low, and in fact, a pitch pattern when a person utters is reflected in a synthesized voice to facilitate listening. An object of the present invention is to provide a method of creating an accent component basic table of a speech synthesizer which can produce synthetic speech.

[0012]

In order to achieve the above-mentioned object, the present invention obtains an intonation control pattern of a phrase component from the number of mora in the phrase of the input text for Japanese-processed input text, An accent control pattern of the accent component is obtained by correcting and interpolating the pitch target value determined for each mora in the phrase of the input text according to the accent environment, and the intonation control pattern of the input text is obtained by superimposing the phrase component and the accent component. ,
A basic accent pattern table that stores a table of basic accent patterns for each mora accent environment is obtained, and a basic accent pattern corresponding to the input mora accent environment is obtained from the table,
A basic accent pattern generation processing unit for correcting the pitch of the mora is obtained, and the processing result of this processing unit is corrected by the number of mora and the mora position subjected to Japanese processing to obtain an accent component intonation pattern. In the rule-based speech synthesis method, the basic accent pattern table groups the environments in which various mora are uttered, registers the value of the basic component of the accent for each environment, and then In addition, it was made to store the normalized values of 3 consonant parts and 4 vowel parts for each mora environment in the accent environment considering the preceding mora, preceding mora, subsequent mora, and subsequent mora. It is characterized by.

[0013]

[Function] The basic accent pattern table is created as described above, and when the accent basic component is determined, the table is referred to. Therefore, the pitch pattern when a person actually utters is reflected in the synthesized voice. It becomes possible to obtain a synthetic voice with high naturalness.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram to which an embodiment of the present invention is applied.
1 is different from FIG. 3 in that an accent pattern generation processing unit 5 is provided in place of the accent pattern calculation unit 3B.

The accent pattern generation processing unit 5 is provided with the number of mora in the delimiter and the accent pattern data (see Table 1) of the mora as the Japanese processing result of the input text. The accent pattern generation processing unit 5 sends this data to the basic accent pattern generation processing unit 6,
Processing is performed by the basic accent pattern table 7 and the correction processing unit 8, which will be described in detail later.

The basic accent pattern generation processing unit 6 is
Processing for reading the basic accent pattern corresponding to the accent environment of the input data from the table 7 is performed. This basic accent pattern table 7 is determined by taking into consideration the types of sounds before and after the mora, accent patterns, etc. when determining the value of the accent component of a mora in a sentence when performing rule composition. This is determined by referring to the values of accent components for determining the accent components for various environments in which mora appears, respectively.

Then, when creating the table, the environments in which various mora are uttered are grouped, and the value of the basic component of the accent is registered for each environment. After that, the accent environment considering the later-described 5 moras including the relevant moras is obtained, and seven values (3 consonant parts and 4 vowel parts) representing the basic components of the accent are normalized and registered for each mora environment. The basic accent pattern table 7 is created.

The following five terms are adopted as factors to be considered when grouping.

1. Was it uttered in the first break? After that? (2 ways) 2. Accent environment with a total of 5 mora before and after (shown in Table 4. In Table 4, 10 patterns are not shown, but there are actually 98 patterns)

[0020]

[Table 4]

With respect to the five preceding mora, the preceding mora, the relevant mora, the subsequent mora, and the subsequent mora shown in Table 4, "L" is LOW, "H" is HIGH, and "C".
Represents the beginning of a sentence or phrase, and “)” represents the end of a sentence or phrase. In Table 4, for example, "O" in "AOI"
Mora is the middle of the accent type 2, so "(LH
L) ”, and the mora“ A ”becomes the environment“ ((LHL) ”.

3. Type of preceding mora (as long or other?) 4. Number of genres of the mora (10 types) CV mora: voiced consonants, consonants, long sounds CV mora: voiced consonants, non-consonants, long sounds CV mora: voiced consonants, consonants, non-long sounds CV mora: voiced consonants, non-consonants, non-long sounds CV Mora: unvoiced consonant, consonant, long CV Mora: unvoiced consonant, non-consonant, long CV Mora: unvoiced consonant, consonant, non-long CV Mora: unvoiced consonant, non-consonant, non-long V Mora: long V Mora: non-long 5. Types of subsequent moras (5 types) Consonants (consonants when CV mora) Voiced consonants (consonants when CV mora) Unvoiced consonants (consonants when CV mora) Vowels (V mora) End of sentence / Phrase end Japanese Patent Application No. 4-43 for each environment grouped as described above.
The databases are classified based on the accent component database storing the accent component pattern shown in No. 145. From the database thus classified, a typical accent pattern is selected for each group, and seven values (three consonant parts and four vowel parts) representing the basic components of this accent are normalized, and the accent basic tables are respectively set. Register to go. The above operation is performed for all groups existing in the database, and the structure of the accent basic table obtained is shown in FIG. 2, the data of the consonant part is No. 1-No. No. 3 as the data of the 3 consonant parts. 1-No. Up to 4, as comments, a indicates whether it is the first division, b indicates the accent environment, c indicates the relevant mora, d indicates the preceding mora, and e indicates the subsequent mora.

By constructing the accent basic table grouped as shown in FIG. 2, the management of the database becomes easy and the time when the database is used can be shortened.

[0024]

As described above, according to the present invention,
After grouping the mora environments in the accent basic table, the values of the accent basic components are registered for each environment, and the seven normalized values of the basic components are registered for each mora environment. Since it has been created, referring to the table when generating the accent component in the pitch pattern of the synthesized voice at the time of synthesis makes it possible to obtain a synthesized voice that is very easy to hear. Further, when the accent basic table is created as described above, it becomes possible to reflect the pitch pattern when a person actually utters in the synthesized voice.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of the present invention.

FIG. 2 is an explanatory diagram of an accent basic table.

FIG. 3 is a block diagram of a conventional speech synthesis apparatus using a rule synthesis method.

FIG. 4 is a diagram showing an intonation processing mode.

[Explanation of symbols]

 1 ... Text input unit 2 ... Japanese processing unit 3A ... Phrase pattern calculation unit 4 ... Speech synthesis unit 5 ... Accent pattern generation processing unit 6 ... Basic accent pattern generation processing unit 7 ... Basic accent pattern table 8 ... Correction processing unit

Claims (1)

[Claims] 1. For input text processed in Japanese,
An intonation control pattern of the phrase component is obtained from the number of mora in the phrase of the input text, and an accent control pattern of the accent component is obtained by correcting the pitch target value determined for each mora in the phrase of the input text according to the accent environment and interrogation processing. Then, an intonation control pattern of the input text is obtained by superimposing the phrase component and the accent component, a basic accent pattern table for storing a table of basic accent patterns for each accent environment of the mora is obtained, and the accent environment of the mora is input. The basic accent pattern corresponding to is obtained from the table, and a basic accent pattern generation processing unit for correcting the pitch of the mora is obtained, and the processing result of this processing unit is calculated according to the number of mora in the phrase cut and the mora position processed in Japanese. Correct the accent component intonation In the speech synthesis method of the rule synthesizing method including a correction processing unit for obtaining a tone, the basic accent pattern table groups the environments in which various mora are uttered, and then the value of the basic component of the accent for each environment. After registering, the accented environment considering the preceding mora, the preceding mora, the subsequent mora, and the subsequent mora including the relevant mora, and the normalized values of 3 consonant parts and 4 vowel parts for each mora environment. A method of creating an accent component basic table for a speech synthesizer, characterized in that the above is stored and created.