Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet

Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet

There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows....

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Venkatesh, Satvik, Benilov, Arthur, Coleman, Philip, Roskam, Frederic
Format: Artikel
Sprache:eng
Schlagworte:
Computer Science - Learning
Computer Science - Sound
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Venkatesh, Satvik
Benilov, Arthur
Coleman, Philip
Roskam, Frederic
description There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows. In this paper, we investigate the various challenges involved in adapting current demixing models in the literature for this use case. Subsequently, inspired by the Hybrid Demucs architecture, we propose the Hybrid Spectrogram Time-domain Audio Separation Network HS-TasNet, which utilises the advantages of spectral and waveform domains. For a latency of 23 ms, the HS-TasNet obtains an overall signal-to-distortion ratio (SDR) of 4.65 on the MusDB test set, and increases to 5.55 with additional training data. These results demonstrate the potential of efficient demixing for real-time low-latency music applications.
doi_str_mv 10.48550/arxiv.2402.17701
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2402_17701</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2402_17701</sourcerecordid><originalsourceid>FETCH-LOGICAL-a671-b1c622790135ed956f1397c6c31c6ef5b2dfdc36f2dee771095e1633a7ebff953</originalsourceid><addsrcrecordid>eNotz8tOwzAUBFBvWKCWD2CFf8DBj9rGy6oCipQWiWQf3TjXlaW85KRA_p5SWI00I410CLkXPNs8ac0fIX3Hz0xuuMyEtVzcksMHQsvm2CHNhy_Wwoy9X-jhPEVPi-GcPNICR0gwx6Gnl7o_0f1Sp9jQYkQ_p-GUoGMlTEec1-QmQDvh3X-uSPnyXO72LH9_fdttcwbGClYLb6S0jgulsXHaBKGc9cary4BB17IJjVcmyAbRWsGdRmGUAot1CE6rFXn4u716qjHFDtJS_bqqq0v9ADQ4SKY</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet</title><source>arXiv.org</source><creator>Venkatesh, Satvik ; Benilov, Arthur ; Coleman, Philip ; Roskam, Frederic</creator><creatorcontrib>Venkatesh, Satvik ; Benilov, Arthur ; Coleman, Philip ; Roskam, Frederic</creatorcontrib><description>There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows. In this paper, we investigate the various challenges involved in adapting current demixing models in the literature for this use case. Subsequently, inspired by the Hybrid Demucs architecture, we propose the Hybrid Spectrogram Time-domain Audio Separation Network HS-TasNet, which utilises the advantages of spectral and waveform domains. For a latency of 23 ms, the HS-TasNet obtains an overall signal-to-distortion ratio (SDR) of 4.65 on the MusDB test set, and increases to 5.55 with additional training data. These results demonstrate the potential of efficient demixing for real-time low-latency music applications.</description><identifier>DOI: 10.48550/arxiv.2402.17701</identifier><language>eng</language><subject>Computer Science - Learning ; Computer Science - Sound</subject><creationdate>2024-02</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,776,881</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2402.17701$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2402.17701$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Venkatesh, Satvik</creatorcontrib><creatorcontrib>Benilov, Arthur</creatorcontrib><creatorcontrib>Coleman, Philip</creatorcontrib><creatorcontrib>Roskam, Frederic</creatorcontrib><title>Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet</title><description>There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows. In this paper, we investigate the various challenges involved in adapting current demixing models in the literature for this use case. Subsequently, inspired by the Hybrid Demucs architecture, we propose the Hybrid Spectrogram Time-domain Audio Separation Network HS-TasNet, which utilises the advantages of spectral and waveform domains. For a latency of 23 ms, the HS-TasNet obtains an overall signal-to-distortion ratio (SDR) of 4.65 on the MusDB test set, and increases to 5.55 with additional training data. These results demonstrate the potential of efficient demixing for real-time low-latency music applications.</description><subject>Computer Science - Learning</subject><subject>Computer Science - Sound</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotz8tOwzAUBFBvWKCWD2CFf8DBj9rGy6oCipQWiWQf3TjXlaW85KRA_p5SWI00I410CLkXPNs8ac0fIX3Hz0xuuMyEtVzcksMHQsvm2CHNhy_Wwoy9X-jhPEVPi-GcPNICR0gwx6Gnl7o_0f1Sp9jQYkQ_p-GUoGMlTEec1-QmQDvh3X-uSPnyXO72LH9_fdttcwbGClYLb6S0jgulsXHaBKGc9cary4BB17IJjVcmyAbRWsGdRmGUAot1CE6rFXn4u716qjHFDtJS_bqqq0v9ADQ4SKY</recordid><startdate>20240227</startdate><enddate>20240227</enddate><creator>Venkatesh, Satvik</creator><creator>Benilov, Arthur</creator><creator>Coleman, Philip</creator><creator>Roskam, Frederic</creator><scope>AKY</scope><scope>GOX</scope></search><sort><creationdate>20240227</creationdate><title>Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet</title><author>Venkatesh, Satvik ; Benilov, Arthur ; Coleman, Philip ; Roskam, Frederic</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a671-b1c622790135ed956f1397c6c31c6ef5b2dfdc36f2dee771095e1633a7ebff953</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Computer Science - Learning</topic><topic>Computer Science - Sound</topic><toplevel>online_resources</toplevel><creatorcontrib>Venkatesh, Satvik</creatorcontrib><creatorcontrib>Benilov, Arthur</creatorcontrib><creatorcontrib>Coleman, Philip</creatorcontrib><creatorcontrib>Roskam, Frederic</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Venkatesh, Satvik</au><au>Benilov, Arthur</au><au>Coleman, Philip</au><au>Roskam, Frederic</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet</atitle><date>2024-02-27</date><risdate>2024</risdate><abstract>There have been significant advances in deep learning for music demixing in recent years. However, there has been little attention given to how these neural networks can be adapted for real-time low-latency applications, which could be helpful for hearing aids, remixing audio streams and live shows. In this paper, we investigate the various challenges involved in adapting current demixing models in the literature for this use case. Subsequently, inspired by the Hybrid Demucs architecture, we propose the Hybrid Spectrogram Time-domain Audio Separation Network HS-TasNet, which utilises the advantages of spectral and waveform domains. For a latency of 23 ms, the HS-TasNet obtains an overall signal-to-distortion ratio (SDR) of 4.65 on the MusDB test set, and increases to 5.55 with additional training data. These results demonstrate the potential of efficient demixing for real-time low-latency music applications.</abstract><doi>10.48550/arxiv.2402.17701</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2402.17701
ispartof
issn
language eng
recordid cdi_arxiv_primary_2402_17701
source arXiv.org
subjects Computer Science - Learning
Computer Science - Sound
title Real-time Low-latency Music Source Separation using Hybrid Spectrogram-TasNet
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-23T17%3A07%3A00IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Real-time%20Low-latency%20Music%20Source%20Separation%20using%20Hybrid%20Spectrogram-TasNet&rft.au=Venkatesh,%20Satvik&rft.date=2024-02-27&rft_id=info:doi/10.48550/arxiv.2402.17701&rft_dat=%3Carxiv_GOX%3E2402_17701%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true