PITCH SYNCHRONOUS MULTIBAND (PSMB) CODING OF SPEECH SIGNALS

Citation
H. Yang et al., PITCH SYNCHRONOUS MULTIBAND (PSMB) CODING OF SPEECH SIGNALS, Speech communication, 19(1), 1996, pp. 61-80
Citations number
18
Categorie Soggetti
Communication,"Language & Linguistics
Journal title
ISSN journal
01676393
Volume
19
Issue
1
Year of publication
1996
Pages
61 - 80
Database
ISI
SICI code
0167-6393(1996)19:1<61:PSM(CO>2.0.ZU;2-U
Abstract
A novel speech coding algorithm, named pitch synchronous multi-band (P SMB), is proposed. The new coding algorithm uses the multi-band excita tion (MBE) model to generate a representative pitch-cycle waveform (PC W) for each frame. The representative PCW of a frame is encoded by two out of three codebooks depending upon whether the frame is related or unrelated to the previous frame. When a frame is related to its previ ous frame, the PCW is encoded by a length-converted-excitation (LCE) c odebook and a stochastic codebook. The codevectors of the LCE codebook are derived from the previous PCW. When a frame is unrelated to its p revious frame, it is encoded by a bandlimited single pulse excitation (BSPE) codebook and the stochastic codebook. The new speech coder intr oduces a pitch-period-based coding feature. It overcomes some weakness es existing in the improved MBE (IMBE) speech coder. The PSMB coder op erating at 4 kbps outperforms the Inmarsat 4.15 kbps IMBE coder by a c lear margin. Our listening tests also indicate that it is slightly bet ter than the FS1016 4.8 kbps code excited linear predictive (CELP) cod er in terms of perceptual quality. Fast search algorithms for the thre e codebooks used in PSMB are also developed. The fast algorithms rende r the new speech coder comparable to the FS1016 CELP coder, in terms o f computational complexity.