Name
Abstract | ||
|---|---|---|
A new paradigm for efficient coding of the fixed-codebook excitation in CELP coders at low-bit rates is presented. In this scheme, the non-zero components of the fixed-codebook excitation are localized to a set of windows whose locations are dependent on the pitch frequency and on the energy contour of a modified residual signal. Highly efficient coding is thus achieved by allocating most of the fixed excitation bits to capture the essential excitation events. The paradigm is validated by computer simulation of a variable-rate speech codec. The performance of the codec is evaluated by informal subjective tests comparing it with TIA standard variable rate speech codecs. The results confirm that the proposed scheme can be used to reproduce speech at average bit rates from 2.3 to 3.4 kbps with very high quality and intelligibility. |
Year | Venue | Keywords |
|---|---|---|
2000 | EUSIPCO | speech,speech coding |
Field | DocType | ISBN |
Vector sum excited linear prediction,Code-excited linear prediction,Speech coding,Computer science,Adaptive Multi-Rate audio codec,Speech recognition,Full Rate,Multi-Band Excitation,Harmonic Vector Excitation Coding,Codec2 | Conference | 978-952-1504-43-3 |
Citations | PageRank | References |
0 | 0.34 | 4 |
Authors | ||
8 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Vladimir Cuperman | 1 | 56 | 11.32 |
| allen gersho | 2 | 3 | 1.19 |
| jan a m van der linden | 3 | 0 | 0.34 |
| Ajit V. Rao | 4 | 197 | 20.50 |
| tungchiang yang | 5 | 0 | 0.34 |
| Sassan Ahmadi | 6 | 61 | 4.70 |
| ryan heidari | 7 | 0 | 0.34 |
| Fenghua Liu | 8 | 0 | 0.68 |