{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/learning-filterbanks-from-raw-speech-for","title":"Learning Filterbanks from Raw Speech for Phone Recognition","arxiv_id":"1711.01161","date":"2017-11-03","proceeding":null,"authors":["Neil Zeghidour","Nicolas Usunier","Iasonas Kokkinos","Thomas Schatz","Gabriel Synnaeve","Emmanuel Dupoux"],"abstract":"We train a bank of complex filters that operates on the raw waveform and is\nfed into a convolutional neural network for end-to-end phone recognition. These\ntime-domain filterbanks (TD-filterbanks) are initialized as an approximation of\nmel-filterbanks, and then fine-tuned jointly with the remaining convolutional\narchitecture. We perform phone recognition experiments on TIMIT and show that\nfor several architectures, models trained on TD-filterbanks consistently\noutperform their counterparts trained on comparable mel-filterbanks. We get our\nbest performance by learning all front-end steps, from pre-emphasis up to\naveraging. Finally, we observe that the filters at convergence have an\nasymmetric impulse response, and that some of them remain almost analytic.","url_abs":"http://arxiv.org/abs/1711.01161v2","url_pdf":"http://arxiv.org/pdf/1711.01161v2.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"learning-filterbanks-from-raw-speech-for","repo_url":"https://github.com/facebookresearch/tdfbanks","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"ok","spdx":"NOASSERTION"}},{"paper_slug":"learning-filterbanks-from-raw-speech-for","repo_url":"https://github.com/google-research/leaf-audio","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"tf","reach":{"status":"unanswered"}}],"tasks":[],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1711.01161","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}