M
M. Hunt
Researcher at National Research Council
Publications - 8
Citations - 267
M. Hunt is an academic researcher from National Research Council. The author has contributed to research in topics: Formant & Voice activity detection. The author has an hindex of 8, co-authored 8 publications receiving 263 citations.
Papers
More filters
Proceedings ArticleDOI
A comparison of several acoustic representations for speech recognition with degraded and undegraded speech
M. Hunt,C. Lefebvre +1 more
TL;DR: Several acoustic representations have been compared in speaker-dependent and independent connected and isolated-word recognition tests with undegraded speech and with speech degraded by adding white noise and by applying a 6-dB/octave spectral tilt.
Proceedings ArticleDOI
Speaker dependent and independent speech recognition experiments with an auditory model
M. Hunt,C. Lefebvre +1 more
TL;DR: The performance of an auditory model has been compared with a conventional filterbank mel-cepstrum representation in speaker-dependent and speaker-independent spoken digit recognition tests and the model performed better than the conventional representation with degraded speech.
Proceedings ArticleDOI
Speech recognition using an auditory model with pitch-synchronous analysis
M. Hunt,C. Lefebvre +1 more
TL;DR: It is shown that instants of glottal excitation can be derived from the model even with noisy speech, and the problem of interaction with harmonics of F0 can be solved.
Proceedings ArticleDOI
Speech recognition using a cochlear model
M. Hunt,C. Lefebvre +1 more
TL;DR: A computational model of the peripheral auditory system consisting of a bank of digital filters followed by compression and half-wave rectification stages and by a set of generalized synchrony detectors that respond to coherence in the signal at the center frequency of the channel is described.
Proceedings ArticleDOI
Evaluating the performance of connected-word speech recognition systems
TL;DR: Simulations show that the commonly used dynamic programming word-sequence matching algorithm has serious shortcomings as an evaluation method at low performance levels, though it is generally reliable at high performance levels and a method using word end-point information provides precise, detailed performance analyses.