Abstract | ||
---|---|---|
A novel nonlinear epoch marking algorithm is proposed for use with voiced speech signals. Epoch detection is useful for speech coding, synthesis and recognition purposes, as it provides both the moment of glottal closure and the in- stantaneous pitch. Our technique functions entirely in sta te space, by operating on a three dimensional reconstruction o f the speech signal which is formed by embedding. By using the fact that one revolution of this reconstructed attracto r is equal to one pitch period, we are able to find points which are pitch synchronous by the use of a Poincare section. Evidently the epoch pulses are pitch synchronous and therefore can be marked. Results using real speech signals are presented to illustrate the performance of the technique. |
Year | Venue | Keywords |
---|---|---|
1998 | EUSIPCO | poincare mapping,speech coding,speech synthesis,poincare maps,epoch detection,epoch marking,nonlinear algorithm,reconstructed attractor,speech recognition,three dimensional reconstruction,voiced speech signals,speech,speech processing,databases |
Field | DocType | ISBN |
Attractor,Speech processing,Speech coding,Voice activity detection,Computer science,Speech recognition,Codec2,State space,Linear predictive coding,Acoustic model | Conference | 978-960-7620-06-4 |
Citations | PageRank | References |
2 | 0.40 | 5 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Iain Mann | 1 | 10 | 1.42 |
Stephen McLaughlin | 2 | 168 | 16.62 |