%Aigaion2 BibTeX export from Idiap Publications %Thursday 21 November 2024 11:47:41 AM @INPROCEEDINGS{vivek-rr-03-32b, author = {Tyagi, Vivek and McCowan, Iain A. and Bourlard, Herv{\'{e}} and Misra, Hemant}, projects = {Idiap}, title = {On Factorizing Spectral Dynamics for Robust Speech Recognition}, booktitle = {Eurospeech}, year = {2003}, note = {IDIAP RR 03-32}, crossref = {vivek-rr-03-32}, abstract = {In this paper, we introduce new dynamic speech features based on the modulation spectrum. These features, termed Mel-cepstrum Modulation Spectrum (MCMS,',','), map the time trajectories of the spectral dynamics into a series of slow and fast moving orthogonal components, providing a more general and discriminative range of dynamic features than traditional delta and acceleration features. The features can be seen as the outputs of an array of band-pass filters spread over the cepstral modulation frequency range of interest. In experiments, it is shown that, as well as providing a slight improvement in clean conditions, these new dynamic features yield a significant increase in speech recognition performance in various noise conditions when compared directly to the standard temporal derivative features and RASTA-PLP features.}, pdf = {https://publications.idiap.ch/attachments/reports/2003/mcms_rr.pdf}, postscript = {ftp://ftp.idiap.ch/pub/reports/2003/mcms_rr.ps.gz}, ipdmembership={speech}, } crossreferenced publications: @TECHREPORT{vivek-rr-03-32, author = {Tyagi, Vivek and McCowan, Iain A. and Bourlard, Herv{\'{e}} and Misra, Hemant}, projects = {Idiap}, title = {On Factorizing Spectral Dynamics for Robust Speech Recognition}, type = {Idiap-RR}, number = {Idiap-RR-32-2003}, year = {2003}, institution = {IDIAP}, note = {in proceedings of Eurospeech 2003}, abstract = {In this paper, we introduce new dynamic speech features based on the modulation spectrum. These features, termed Mel-cepstrum Modulation Spectrum (MCMS,',','), map the time trajectories of the spectral dynamics into a series of slow and fast moving orthogonal components, providing a more general and discriminative range of dynamic features than traditional delta and acceleration features. The features can be seen as the outputs of an array of band-pass filters spread over the cepstral modulation frequency range of interest. In experiments, it is shown that, as well as providing a slight improvement in clean conditions, these new dynamic features yield a significant increase in speech recognition performance in various noise conditions when compared directly to the standard temporal derivative features and RASTA-PLP features.}, pdf = {https://publications.idiap.ch/attachments/reports/2003/mcms_rr.pdf}, postscript = {ftp://ftp.idiap.ch/pub/reports/2003/mcms_rr.ps.gz}, ipdmembership={speech}, }