Learning Filterbanks from Raw Speech for Phoneme Recognition

Abstract : We train a bank of complex filters that operates on the raw waveform and is fed into a convolutional neural network for end-to-end phone recognition. These time-domain filterbanks (TD-filterbanks) are initialized as an approximation of mel-filterbanks, and then fine-tuned jointly with the remaining convolutional architecture. We perform phone recognition experiments on TIMIT and show that for several architectures, models trained on TD-filterbanks consistently outperform their counterparts trained on comparable mel-filterbanks. We get our best performance by learning all front-end steps, from pre-emphasis up to averaging. Finally, we observe that the filters at convergence have an asymmetric impulse response, and that some of them remain almost analytic.
Liste complète des métadonnées

Cited literature [15 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-01888737
Contributor : Emmanuel Dupoux <>
Submitted on : Friday, December 7, 2018 - 2:42:43 PM
Last modification on : Thursday, February 7, 2019 - 4:58:41 PM
Document(s) archivé(s) le : Friday, March 8, 2019 - 2:30:20 PM

File

1711.01161.pdf
Files produced by the author(s)

Identifiers

Citation

Neil Zeghidour, Nicolas Usunier, Iasonas Kokkinos, Thomas Schatz, Gabriel Synnaeve, et al.. Learning Filterbanks from Raw Speech for Phoneme Recognition. ICASSP 2018 - IEEE International Conference on Acoustics, Speech and Signal Processing, Apr 2018, Calgary, Alberta, Canada. ⟨hal-01888737⟩

Share

Metrics

Record views

111

Files downloads

39