Abstract | ||
---|---|---|
In our paper, the problem of very low bit rate segmental speech coding is addressed. The basic units are found automatically in the training database using temporal decomposition, vector quantiza- tion and multigrams. They are modelled by HMMs. The coding is based on recognition and synthesis. In single speaker tests, we obtained intelligible and naturally sounding speech at mean rate of 211.2 b/s. In the end, future extensions of our scheme (diphone- like synthesis and speaker adaptation) as well as possible use of automatically derived units in recognition are discussed. |
Year | DOI | Venue |
---|---|---|
1998 | 10.1109/ICASSP.1998.675337 | Acoustics, Speech and Signal Processing, 1998. Proceedings of the 1998 IEEE International Conference |
Keywords | Field | DocType |
hidden Markov models,speech coding,speech intelligibility,speech recognition,speech synthesis,vector quantisation,vocoders,211.2 bit/s,HMM,VLBR coding,automatically derived units,diphone-like synthesis,intelligible speech,multigrams,naturally sounding speech,phonetic approach,segmental speech coding,segmental vocoder,single speaker tests,speaker adaptation,speech recognition,speech synthesis,temporal decomposition,training database,vector quantization,very low bit rate coding | Speech processing,Speech synthesis,Vector sum excited linear prediction,Speech coding,Pattern recognition,Voice activity detection,Computer science,Speech recognition,Vector quantization,Speaker recognition,Artificial intelligence,Linear predictive coding | Conference |
Volume | ISSN | ISBN |
2 | 1520-6149 | 0-7803-4428-6 |
Citations | PageRank | References |
20 | 1.30 | 8 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jan Cernocký | 1 | 1273 | 135.94 |
Geneviève Baudoin | 2 | 138 | 19.08 |
Gérard Chollet | 3 | 725 | 129.74 |