Title:
SPEECH SYNTHESIZING METHOD
Document Type and Number:
Japanese Patent JPH01185700
Kind Code:
A
Abstract:
PURPOSE:To facilitate a process of transition from a vowel to a consonant by making the timing of a speech waveform signals from 1st and 2nd acoustic tube model parts different and adding them for speech synthesis so that partial outputs of syllables overlap with each other. CONSTITUTION:A syllable processing part 3 generates syllable data which are sectioned by pitch, energy, and continuance from a language processed by a Japanese processing part 1, a syllable data switching part 4 switches the data in syllable data units and supplies them to 1st and 2nd phoneme processing parts 5 and 6 alternately, and 1st and 2nd parameter interpolation parts 7 and 8 interpolate phoneme level data which are processed with syllable parameters by using phoneme parameters. Then 1st and 2nd acoustic pipe model parts 9 and 10 performs control for simulating variation in the sectional area of an acoustic pipe, makes the speech waveform signals from outputted by both model parts 9 and 10 different in timing, and an adder 11 adds them so that they overlap with each other, thus synthesizing a speech. Consequently, tone coupling corresponding to the join of voicing at the time of the transition from a vowel to a consonant is facilitated and the speech becomes natural.
More Like This:
JP3859496 | DEVICE FOR OUTPUTTING EXPRESSION OF WORD |
JPH09230893 | REGULAR SPEECH SYNTHESIS METHOD AND DEVICE THEREFOR |
JPH0359699 | AUDIO RESPONSE DEVICE |
Inventors:
SUDA NORIO
Application Number:
JP924788A
Publication Date:
July 25, 1989
Filing Date:
January 19, 1988
Export Citation:
Assignee:
MEIDENSHA ELECTRIC MFG CO LTD
International Classes:
G10L13/07; (IPC1-7): G10L5/04
Domestic Patent References:
JPS4950803A | 1974-05-17 | |||
JPS61118798A | 1986-06-06 | |||
JP56222898A | ||||
JPS61119798A | 1986-06-06 | |||
JPS56111898A | 1981-09-03 |
Attorney, Agent or Firm:
Fujiya Shiga