Title:
【発明の名称】コード駆動線形予測音声符号化方式
Document Type and Number:
Japanese Patent JP3092519
Kind Code:
B2
Abstract:
PROBLEM TO BE SOLVED: To compress the information volume by not generating codes in a steady-state section at the time of sounding, and to provide a code excited linear predictive speech encoding system for improving the encoding efficiency. SOLUTION: In a frame A judged as sounding and when a steady-state judging unit 8 compares the power and the pitch of the frame A with those of a steady-state section start frame and the powers and the pitches are the same respectively, the frame A is judged as a steady-state and no encoding information is transmitted by an encoder. A decoder, when encoded information is transmitted to it, stores synthesized speech in a memory, and when encoded information is not transmitted to it, it produces a synthesized speech by shifting the synthesized speech stored in the memory by a pitch.
More Like This:
Inventors:
Koichi Shibagaki
Application Number:
JP19563296A
Publication Date:
September 25, 2000
Filing Date:
July 05, 1996
Export Citation:
Assignee:
NEC
International Classes:
G10L19/00; G10L19/038; G10L19/04; G10L19/08; G10L19/125; G10L19/16; H03M7/30; (IPC1-7): G10L19/04; H03M7/30
Domestic Patent References:
| JP63254499A | ||||
| JP2146100A | ||||
| JP1257999A |
Attorney, Agent or Firm:
Asamichi Kato