A hybrid scheme for encoding audio signal using hidden Markov models of waveforms

摘要

This paper reports on recent results related to audiophonic signals encodingusing time-scale and time-frequency transform. More precisely, non-linear,structured approximations for tonal and transient components using local cosineand wavelet bases will be described, yielding expansions of audio signals inthe form tonal + transient + residual. We describe a general formulationinvolving hidden Markov models, together with corresponding rate estimates.Estimators for the balance transient/tonal are also discussed.