A Segmental HMM for Speech Waveforms

We present a purely time domain approach to speech processing which identies waveform samples at the boundaries between glottal pulse periods (in voiced speech) or at the boundaries between unvoiced segments. An ecien t algorithm for inferring these boundaries is derived from a simple probabilistic generative model of speech and state of the art results are presented on pitch tracking, voiced/unvoiced detection and timescale modication.