{"id":"https://openalex.org/W4408351949","doi":"https://doi.org/10.1109/icassp49660.2025.10889794","title":"Speaking from Coarse to Fine: Improving Neural Codec Language Model via Multi-Scale Speech Coding and Generation","display_name":"Speaking from Coarse to Fine: Improving Neural Codec Language Model via Multi-Scale Speech Coding and Generation","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408351949","doi":"https://doi.org/10.1109/icassp49660.2025.10889794"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10889794","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10889794","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5072447806","display_name":"Haohan Guo","orcid":"https://orcid.org/0000-0002-3393-9984"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Haohan Guo","raw_affiliation_strings":["The Chinese University of Hong Kong,Hong Kong SAR, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,Hong Kong SAR, China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101692260","display_name":"Fenglong Xie","orcid":null},"institutions":[{"id":"https://openalex.org/I862669128","display_name":"Xiaomi (China)","ror":"https://ror.org/029f7bn57","country_code":"CN","type":"company","lineage":["https://openalex.org/I862669128"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Fenglong Xie","raw_affiliation_strings":["Xiaohongshu Inc.,Shanghai,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Xiaohongshu Inc.,Shanghai,China","institution_ids":["https://openalex.org/I862669128"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5028178332","display_name":"Dongchao Yang","orcid":"https://orcid.org/0000-0003-0879-4047"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Dongchao Yang","raw_affiliation_strings":["The Chinese University of Hong Kong,Hong Kong SAR, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,Hong Kong SAR, China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5025119324","display_name":"Xixin Wu","orcid":"https://orcid.org/0000-0001-9543-1572"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Xixin Wu","raw_affiliation_strings":["The Chinese University of Hong Kong,Hong Kong SAR, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,Hong Kong SAR, China","institution_ids":["https://openalex.org/I177725633"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5019458385","display_name":"Helen Meng","orcid":"https://orcid.org/0000-0002-4427-3532"},"institutions":[{"id":"https://openalex.org/I177725633","display_name":"Chinese University of Hong Kong","ror":"https://ror.org/00t33hh48","country_code":"HK","type":"education","lineage":["https://openalex.org/I177725633"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Helen Meng","raw_affiliation_strings":["The Chinese University of Hong Kong,Hong Kong SAR, China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"The Chinese University of Hong Kong,Hong Kong SAR, China","institution_ids":["https://openalex.org/I177725633"]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":3.346,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":{"value":0.90301048,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":{"min":95,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9650999903678894,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10201","display_name":"Speech Recognition and Synthesis","score":0.9650999903678894,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.7996232509613037},{"id":"https://openalex.org/keywords/speech-recognition","display_name":"Speech recognition","score":0.6368523836135864},{"id":"https://openalex.org/keywords/codec","display_name":"Codec","score":0.6091368794441223},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.5930512547492981},{"id":"https://openalex.org/keywords/codec2","display_name":"Codec2","score":0.5646288394927979},{"id":"https://openalex.org/keywords/adaptive-multi-rate-audio-codec","display_name":"Adaptive Multi-Rate audio codec","score":0.5471760630607605},{"id":"https://openalex.org/keywords/coding","display_name":"Coding (social sciences)","score":0.5433725118637085},{"id":"https://openalex.org/keywords/speech-coding","display_name":"Speech coding","score":0.5420052409172058},{"id":"https://openalex.org/keywords/scale","display_name":"Scale (ratio)","score":0.42061564326286316},{"id":"https://openalex.org/keywords/voice-activity-detection","display_name":"Voice activity detection","score":0.3593042194843292},{"id":"https://openalex.org/keywords/speech-processing","display_name":"Speech processing","score":0.29780787229537964},{"id":"https://openalex.org/keywords/linear-predictive-coding","display_name":"Linear predictive coding","score":0.2745380401611328},{"id":"https://openalex.org/keywords/telecommunications","display_name":"Telecommunications","score":0.07868623733520508}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7996232509613037},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.6368523836135864},{"id":"https://openalex.org/C161765866","wikidata":"https://www.wikidata.org/wiki/Q184748","display_name":"Codec","level":2,"score":0.6091368794441223},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.5930512547492981},{"id":"https://openalex.org/C75217168","wikidata":"https://www.wikidata.org/wiki/Q1105653","display_name":"Codec2","level":4,"score":0.5646288394927979},{"id":"https://openalex.org/C177067256","wikidata":"https://www.wikidata.org/wiki/Q4676210","display_name":"Adaptive Multi-Rate audio codec","level":4,"score":0.5471760630607605},{"id":"https://openalex.org/C179518139","wikidata":"https://www.wikidata.org/wiki/Q5140297","display_name":"Coding (social sciences)","level":2,"score":0.5433725118637085},{"id":"https://openalex.org/C13895895","wikidata":"https://www.wikidata.org/wiki/Q3270773","display_name":"Speech coding","level":2,"score":0.5420052409172058},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.42061564326286316},{"id":"https://openalex.org/C204201278","wikidata":"https://www.wikidata.org/wiki/Q1332614","display_name":"Voice activity detection","level":3,"score":0.3593042194843292},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.29780787229537964},{"id":"https://openalex.org/C59883199","wikidata":"https://www.wikidata.org/wiki/Q1826438","display_name":"Linear predictive coding","level":3,"score":0.2745380401611328},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.07868623733520508},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.0},{"id":"https://openalex.org/C62520636","wikidata":"https://www.wikidata.org/wiki/Q944","display_name":"Quantum mechanics","level":1,"score":0.0},{"id":"https://openalex.org/C121332964","wikidata":"https://www.wikidata.org/wiki/Q413","display_name":"Physics","level":0,"score":0.0},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10889794","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10889794","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[{"display_name":"Quality Education","score":0.6000000238418579,"id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[{"id":"https://openalex.org/F4320326427","display_name":"Innovation and Technology Fund","ror":null}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":32,"referenced_works":["https://openalex.org/W2119913432","https://openalex.org/W2124509324","https://openalex.org/W2752796333","https://openalex.org/W3024869864","https://openalex.org/W4296068817","https://openalex.org/W4307323391","https://openalex.org/W4367721746","https://openalex.org/W4391833199","https://openalex.org/W4402111239","https://openalex.org/W4402111455","https://openalex.org/W4402111470","https://openalex.org/W4406461319","https://openalex.org/W4406495593","https://openalex.org/W4408355207","https://openalex.org/W6757817989","https://openalex.org/W6762931180","https://openalex.org/W6778883912","https://openalex.org/W6838843145","https://openalex.org/W6848735303","https://openalex.org/W6850334629","https://openalex.org/W6852421699","https://openalex.org/W6852581948","https://openalex.org/W6853096648","https://openalex.org/W6853515095","https://openalex.org/W6854866820","https://openalex.org/W6857409037","https://openalex.org/W6861507043","https://openalex.org/W6869301554","https://openalex.org/W6869585042","https://openalex.org/W6870085509","https://openalex.org/W6870225445","https://openalex.org/W6936129901"],"related_works":["https://openalex.org/W2374470469","https://openalex.org/W2159899134","https://openalex.org/W2899318368","https://openalex.org/W2363660481","https://openalex.org/W2352448339","https://openalex.org/W2370747337","https://openalex.org/W2129913000","https://openalex.org/W137020845","https://openalex.org/W2787189739","https://openalex.org/W1542588102"],"abstract_inverted_index":{"The":[0,116,137],"neural":[1,60,157],"codec":[2,158],"language":[3,159],"model":[4,160],"(CLM)":[5],"has":[6],"demonstrated":[7],"remarkable":[8],"performance":[9],"in":[10,88,113,126,153],"text-to-speech":[11],"(TTS)":[12],"synthesis.":[13],"However,":[14],"troubled":[15],"by":[16],"\"recency":[17],"bias\",":[18],"CLM":[19],"lacks":[20],"sufficient":[21],"attention":[22],"to":[23,52,63],"coarse-grained":[24],"information":[25],"at":[26],"a":[27,42,58,67,150,155],"higher":[28],"temporal":[29],"scale,":[30],"often":[31],"producing":[32],"unnatural":[33],"or":[34],"even":[35],"unintelligible":[36],"speech.":[37],"This":[38],"work":[39],"proposes":[40],"CoFi-Speech,":[41],"coarse-to-fine":[43,138],"CLM-TTS":[44],"approach,":[45,145],"employing":[46],"multi-scale":[47,59,68,119,128,139],"speech":[48,65,130,135],"coding":[49,120],"and":[50,95,110],"generation":[51,94],"address":[53],"this":[54,86],"issue.":[55],"We":[56],"train":[57],"codec,":[61],"CoFi-Codec,":[62],"encode":[64],"into":[66],"discrete":[69,129],"representation,":[70],"comprising":[71],"multiple":[72],"token":[73],"sequences":[74],"with":[75],"different":[76],"time":[77],"resolutions.":[78],"Then,":[79],"we":[80],"propose":[81],"CoFi-LM":[82],"that":[83],"can":[84],"generate":[85],"representation":[87],"two":[89],"modes:":[90],"the":[91,96,122,143],"single-LM-based":[92],"chain-of-scale":[93],"multiple-LM-based":[97],"stack-of-scale":[98,144],"generation.":[99],"In":[100],"experiments,":[101],"CoFi-Speech":[102],"significantly":[103],"outperforms":[104],"single-scale":[105],"baseline":[106],"systems":[107],"on":[108],"naturalness":[109],"speaker":[111],"similarity":[112],"zero-shot":[114],"TTS.":[115,162],"analysis":[117],"of":[118,124],"demonstrates":[121],"effectiveness":[123],"CoFi-Codec":[125],"learning":[127],"representations":[131],"while":[132],"keeping":[133],"high-quality":[134,156],"reconstruction.":[136],"generation,":[140],"especially":[141],"for":[142,161],"is":[146],"also":[147],"validated":[148],"as":[149],"crucial":[151],"approach":[152],"pursuing":[154]},"counts_by_year":[{"year":2025,"cited_by_count":2}],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}