LP-TRAP: Linear predictive temporal patterns
Autoregressive modeling is applied for approximating the temporal evolution of spectral density in critical-band-sized subbands of a segment of speech signal. The generalized autocorrelation linear predictive technique allows for a compromise between fitting the peaks and the troughs of the Hilbert envelope of the signal in the sub-band. The cosine transform coefficients of the approximated sub-band envelopes, computed recursively from the all-pole polynomials, are used as inputs to a TRAP-based speech recognition system and are shown to improve recognition accuracy.
- icslp04-lptraps.pdf application/x-pdf 172 KB Download File
Also Published In
- INTERSPEECH 2004 - ICSLIP: Proceedings, 8th International Conference on Spoken Language Processing: October 4 - 8, 2004, International Convention Center Jeju, Jeju Island, Korea
- Sunjin Printing Co.
More About This Work
- Academic Units
- Electrical Engineering
- Published Here
- June 28, 2012