Article(id=1327017833927651718, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1327017642604475226, articleNumber=null, orderNo=null, doi=10.7544/issn1000-1239.202440329, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=research-article, receivedDate=1716220800000, receivedDateStr=2024-05-21, revisedDate=1743004800000, revisedDateStr=2025-03-27, acceptedDate=1743696000000, acceptedDateStr=2025-04-04, onlineDate=1794315306406, onlineDateStr=2026-11-10, pubDate=1761926400000, pubDateStr=2025-11-01, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1794315306406, onlineIssueDateStr=2026-11-10, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1794315306406, creator=13701087609, updateTime=1794315306406, updator=13701087609, issue=Issue{id=1327017642604475226, tenantId=1146029695717560320, journalId=1324759914146529478, year='2025', volume='62', issue='11', pageStart='2651', pageEnd='2887', issueExtLink='null', onlineDate='null', pubDate='1761926400000', pubDateStr='2025-11-01', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1794315260792, creator='13701087609', updateTime=1794376986370, updator='13701087609', preIssue=null, nextIssue=null, articleTotal=null, ext={EN=IssueExt(id=1327276538975859352, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1327017642604475226, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1327276538980053657, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1327017642604475226, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null, downloadFileDto=null}, startPage=2725, endPage=2736, ext={EN=ArticleExt(id=1327017834108006791, articleId=1327017833927651718, tenantId=1146029695717560320, journalId=1324759914146529478, language=EN, title=Neural Speech Codec Method Based on Mel Spectrogram and Squeeze-Excitation-Weighted Quantization, columnId=1325489891789910943, journalTitle=Journal of Computer Research and Development, columnName=Artificial Intelligence, runingTitle=null, highlight=null, articleAbstract=

At present, end-to-end speech neural codecs, represented by SoundStream, have demonstrated outstanding performance in reconstructed speech quality. However, these methods require extensive convolutional computations, leading to lengthy encoding times. To address this issue, we introduce a neural speech codec method based on Mel spectrogram and squeeze-excitation-weighted quantization. This method aims to maintain high speech perceptual quality while reducing computational costs and increasing operational speed, thereby minimizing latency. Specifically, we utilize Mel spectrogram features as input, capitalize on the temporal compression properties during Mel spectrogram extraction, and combine a lower-layer convolutional encoder to simplify the computation process. Additionally, inspired by squeezed excitation network concepts, we extract excitation weights for each dimension of the output features from the encoder’s final layer. These weights are used as the weighting coefficients for each dimension of the compressed features when calculating codebook distances in the quantizer, thus enabling the learning of correlations between features and enhancing the performance of quantization. Experimental results on the LibriTTS and VCTK datasets indicate that this method significantly enhances the computational speed of the encoder and improves the reconstructed speech quality at lower bit rates (≤3 Kbps). For instance, at a bitrate of 1.5 Kbps, the real-time factor (RTF) of encoding computations can increase by up to 4.6 times. Regarding perceptual quality, at a bitrate of 0.75 Kbps, objective metrics such as short-time objective intelligibility (STOI) and virtual speech quality objective listener (VISQOL) show an average improvement of 8.72% compared with the baseline. Additionally, ablation studies not only demonstrate that the optimization effect of compressed excitation weight methods is inversely correlated with bitrate, but also reveal that, compared with the periodic activation function Snake, the Relu activation function can significantly speed up processing while maintaining comparable speech perceptual quality.

, authors=Junzuo Zhou1, Jiangyan Yi1, *, Jianhua Tao2, 3, Yong Ren1, Tao Wang1, authorsList=Junzuo Zhou, Jiangyan Yi, Jianhua Tao, Yong Ren, Tao Wang, authorCompany=null, correspAuthors=Jiangyan Yi, authorNote=

born in 2000. Master candidate. His main research interest includes speech synthesis

born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

born in 1998. PhD candidate. His main research interest includes speech synthesis

born in 1996. PhD. His main research interest includes speech synthesis

, correspAuthorsNote=null, copyrightStatement=Copyright © 2025 Journal of Computer Research and Development. All rights reserved., copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, fund=null), CN=ArticleExt(id=1327017835857031580, articleId=1327017833927651718, tenantId=1146029695717560320, journalId=1324759914146529478, language=CN, title=基于梅尔谱与压缩激励加权量化的语音神经编解码方法, columnId=1325489892104483745, journalTitle=计算机研究与发展, columnName=人工智能, runingTitle=null, highlight=null, articleAbstract=

目前,以 SoundStream 等为代表的端到端语音神经编解码器在重建语音感知质量方面展现了优异性能. 然而,这些方法需要大量的卷积计算,从而导致较长的编码时间. 为缓解上述问题,提出基于梅尔谱和压缩激励加权量化的神经语音编解码方法. 该方法旨在保持较高语音感知质量的同时降低计算代价,加快运行速度,从而减少时延. 具体而言,采用梅尔谱特征作为输入,借助梅尔谱提取过程中时域压缩的性质,并结合低层卷积编码器以简化运算过程. 此外,借鉴压缩激励网络思想,提取了编码器最后一层输出特征各维度的激励权重,将其作为量化器中计算码本距离时压缩特征各维度的权重系数,由此学习特征间的相关性并优化量化性能. 在 LibriTTS 和 VCTK 数据集上的实验结果表明,该方法显著提升了编码器计算速度,且能在较低比特率时(≤3 Kbps)提升重建语音质量. 以比特率1.5 Kbps时为例,编码计算实时率(real-time factor,RTF)最多可提升4.6倍. 对于感知质量,以0.75 Kbps为例,短时客观可懂度(short-time objective intelligibility,STOI)、虚拟语音质量客观评估(virtual speech quality objective listener,VISQOL)等客观指标相较基线平均可提升8.72%. 此外,消融实验不仅表明压缩激励权重方法的优化效果与比特率呈反相关,而且发现Relu激活函数相较周期性质激活函数Snake而言,在语音感知质量相当的情况下,能大量加快运行速度.

, authors=周俊佐1, 易江燕1, *, 陶建华2, 3, 任勇1, 汪涛1, authorsList=周俊佐, 易江燕, 陶建华, 任勇, 汪涛, authorCompany=null, correspAuthors=易江燕, authorNote=

2000年生. 硕士研究生. 主要研究方向为语音合成

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

1998年生. 博士研究生. 主要研究方向为语音合成

1996年生. 博士. 主要研究方向为语音合成

, correspAuthorsNote=
易江燕()
, copyrightStatement=版权所有 © 2025 计算机研究与发展编辑部, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=mB23/zBYcHuPnCWcIU1wgA==, magXml=yn+kDatDWc5kNpG1zmi+Gw==, pdfUrl=null, pdf=kfxqQ/b0d3v77wwt+ResFw==, pdfFileSize=1908245, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=gEsCgL3pmKGc/MFnAKDSwQ==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=jL7qLJQWAh4/1Vst8TSBmQ==, mapNumber=null, fund=null)}, authors=[Author(id=1327017836339376551, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=zhoujunzuo2023@ia.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017836469399977, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836339376551, language=EN, stringName=Junzuo Zhou, firstName=Junzuo, middleName=null, lastName=Zhou, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 2000. Master candidate. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 2000. Master candidate. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017836532314538, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836339376551, language=CN, stringName=周俊佐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"G6Cc6ZXkwbYOhylFlTXI1g==","content":"

2000年生. 硕士研究生. 主要研究方向为语音合成

"}, bioImg=G6Cc6ZXkwbYOhylFlTXI1g==, bioContent=

2000年生. 硕士研究生. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017836612006316, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=jiangyan.yi@nlpr.ia.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1327017836695892398, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836612006316, language=EN, stringName=Jiangyan Yi, firstName=Jiangyan, middleName=null, lastName=Yi, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

"}, bioImg=null, bioContent=

born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017836788167087, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836612006316, language=CN, stringName=易江燕, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1中国科学院自动化研究所 北京 100190, bio={"img":"Y6IZUg19571lrS3SG6x3Ew==","content":"

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

"}, bioImg=Y6IZUg19571lrS3SG6x3Ew==, bioContent=

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017836872053169, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017836989493684, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836872053169, language=EN, stringName=Jianhua Tao, firstName=Jianhua, middleName=null, lastName=Tao, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, address=2Department of Automation, Tsinghua University, Beijing 100084
3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084, bio={"content":"

born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

"}, bioImg=null, bioContent=

born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837073379766, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836872053169, language=CN, stringName=陶建华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, address=2清华大学自动化系 北京 100084
3北京信息科学与技术国家研究中心(清华大学) 北京 100084, bio={"img":"ar6mn702nAnH7GRimHQ9BQ==","content":"

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

"}, bioImg=ar6mn702nAnH7GRimHQ9BQ==, bioContent=

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836159021472, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=2, ext=[AuthorCompanyExt(id=1327017836167410081, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Department of Automation, Tsinghua University, Beijing 100084), AuthorCompanyExt(id=1327017836175798690, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2清华大学自动化系 北京 100084)]), AuthorCompany(id=1327017836242907555, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=3, ext=[AuthorCompanyExt(id=1327017836251296164, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084), AuthorCompanyExt(id=1327017836263879077, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3北京信息科学与技术国家研究中心(清华大学) 北京 100084)])]), Author(id=1327017837161460152, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017837245346234, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837161460152, language=EN, stringName=Yong Ren, firstName=Yong, middleName=null, lastName=Ren, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1998. PhD candidate. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 1998. PhD candidate. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837329232315, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837161460152, language=CN, stringName=任勇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"+HumlNn1XAhshqsdXs32aw==","content":"

1998年生. 博士研究生. 主要研究方向为语音合成

"}, bioImg=+HumlNn1XAhshqsdXs32aw==, bioContent=

1998年生. 博士研究生. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017837396341181, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017837484421567, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837396341181, language=EN, stringName=Tao Wang, firstName=Tao, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1996. PhD. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 1996. PhD. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837551530432, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837396341181, language=CN, stringName=汪涛, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"WqHuMxLwQwmSAuPYeAPuHQ==","content":"

1996年生. 博士. 主要研究方向为语音合成

"}, bioImg=WqHuMxLwQwmSAuPYeAPuHQ==, bioContent=

1996年生. 博士. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])])], keywords=[Keyword(id=1327017837673165249, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, orderNo=1, keyword=speech codec), Keyword(id=1327017837736079810, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, orderNo=2, keyword=Mel-spectrogram), Keyword(id=1327017837803188675, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, orderNo=3, keyword=squeeze-and-excitation networks), Keyword(id=1327017837866103236, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, orderNo=4, keyword=residual vector quantization), Keyword(id=1327017837941600709, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, orderNo=5, keyword=generative adversarial network), Keyword(id=1327017838012903878, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, orderNo=1, keyword=语音编解码), Keyword(id=1327017838075818439, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, orderNo=2, keyword=梅尔谱图), Keyword(id=1327017838180676040, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, orderNo=3, keyword=压缩激励网络), Keyword(id=1327017838277145033, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, orderNo=4, keyword=残差矢量量化), Keyword(id=1327017838369419722, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, orderNo=5, keyword=生成对抗网络)], refs=[Reference(id=1327017840890196454, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=1, rfOrder=0, authorNames=null, journalName=null, refType=null, unstructuredReference=De Andrade J F, De Campos M L R, Apolinario J A. Speech privacy for modern mobile communication systems[C]//Proc of the 33rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2008: 1777−1780, articleTitle=null, refAbstract=null), Reference(id=1327017840957305319, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=2, rfOrder=1, authorNames=null, journalName=null, refType=null, unstructuredReference=Haneche H, Ouahabi A, Boudraa B. Compressed sensing-speech coding scheme for mobile communications[J]. Circuits, Systems, and Signal Processing, 2021, 40(10): 5106−5126, articleTitle=null, refAbstract=null), Reference(id=1327017841024414184, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=3, rfOrder=2, authorNames=null, journalName=null, refType=null, unstructuredReference=Budagavi M, Gibson J D. Speech coding in mobile radio communications[J]. Proceedings of the IEEE, 1998, 86(7): 1402−1412, articleTitle=null, refAbstract=null), Reference(id=1327017841095717353, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=4, rfOrder=3, authorNames=null, journalName=null, refType=null, unstructuredReference=Bessette B, Salami R, Lefebvre R, et al. The adaptive multirate wideband speech codec (AMR-WB)[J]. IEEE Transactions on Speech and Audio Processing, 2002, 10(8): 620−636, articleTitle=null, refAbstract=null), Reference(id=1327017841158631914, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=5, rfOrder=4, authorNames=null, journalName=null, refType=null, unstructuredReference=Cox R V, Kroon P. Low bit-rate speech coders for multimedia communication[J]. IEEE Communications Magazine, 1996, 34(12): 34−41, articleTitle=null, refAbstract=null), Reference(id=1327017841221546475, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=6, rfOrder=5, authorNames=null, journalName=null, refType=null, unstructuredReference=Huang Yongfeng, Liu Chenghao, Tang Shanyu, et al. Steganography integration into a low-bit rate speech codec[J]. IEEE Transactions on Information Forensics and Security, 2012, 7(6): 1865−1875, articleTitle=null, refAbstract=null), Reference(id=1327017841297043948, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=7, rfOrder=6, authorNames=null, journalName=null, refType=null, unstructuredReference=Valin J M, Terriberry T B, Montgomery C, et al. A high-quality speech and audio codec with less than 10-ms delay[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2009, 18(1): 58−67, articleTitle=null, refAbstract=null), Reference(id=1327017841376735725, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=8, rfOrder=7, authorNames=null, journalName=null, refType=null, unstructuredReference=Hicsonmez S, Sencar H T, Avcibas I. Audio codec identification from coded and transcoded audios[J]. Digital Signal Processing, 2013, 23(5): 1720−1730, articleTitle=null, refAbstract=null), Reference(id=1327017841448038894, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=9, rfOrder=8, authorNames=null, journalName=null, refType=null, unstructuredReference=Dietz M, Multrus M, Eksler V, et al. Overview of the EVS codec architecture[C]//Proc of the 40th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2015: 5698−5702., articleTitle=null, refAbstract=null), Reference(id=1327017841510953455, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=10, rfOrder=9, authorNames=null, journalName=null, refType=null, unstructuredReference=Valin J M, Vos K, Terriberry T. Definition of the opus audio codec[EB/OL]. 2012[2024-12-26]. https://datatracker.ietf.org/doc/html/rfc6716, articleTitle=null, refAbstract=null), Reference(id=1327017841578062320, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=11, rfOrder=10, authorNames=null, journalName=null, refType=null, unstructuredReference=Zeghidour N, Luebs A, Omran A, et al. SoundStream: An end-to-end neural audio codec[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2021, 30: 495−507, articleTitle=null, refAbstract=null), Reference(id=1327017841640976881, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=12, rfOrder=11, authorNames=null, journalName=null, refType=null, unstructuredReference=Biswas A, Jia D. Audio codec enhancement with generative adversarial networks[C]//Proc of the 45th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2020: 356−360, articleTitle=null, refAbstract=null), Reference(id=1327017841729057266, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=13, rfOrder=12, authorNames=null, journalName=null, refType=null, unstructuredReference=Stimberg F, Narest A, Bazzica A, et al. WaveNetEQ—Packet loss concealment with waveRNN[C]//Proc of the 54th Asilomar Conf on Signals, Systems, and Computers. Piscataway, NJ: IEEE, 2020: 672−676, articleTitle=null, refAbstract=null), Reference(id=1327017841787777523, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=14, rfOrder=13, authorNames=null, journalName=null, refType=null, unstructuredReference=Xiao Wei, Liu Wenzhe, Wang Meng, et al. Multi-mode neural speech coding based on deep generative networks[C]//Proc of the 24th Annual Conf of the Int Speech Communication Association. Grenoble, France: ISCA, 2023: 819−823, articleTitle=null, refAbstract=null), Reference(id=1327017841850692084, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=15, rfOrder=14, authorNames=null, journalName=null, refType=null, unstructuredReference=Wu Y-C, Gebru I D, Marković D, et al. Audiodec: An open-source streaming high-fidelity neural audio codec[C]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023: 1−5, articleTitle=null, refAbstract=null), Reference(id=1327017841917800949, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=16, rfOrder=15, authorNames=null, journalName=null, refType=null, unstructuredReference=Jiang Xue, Peng Xiulian, Zhang Yuan, et al. Disentangled feature learning for real-time neural speech coding[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10094723, articleTitle=null, refAbstract=null), Reference(id=1327017841984909814, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=17, rfOrder=16, authorNames=null, journalName=null, refType=null, unstructuredReference=Petermann D, Jang I, Kim M. Native multi-band audio coding within hyper-autoencoded reconstruction propagation networks[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10094593, articleTitle=null, refAbstract=null), Reference(id=1327017842043630071, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=18, rfOrder=17, authorNames=null, journalName=null, refType=null, unstructuredReference=Lim H, Lee J, Kim B H, et al. End-to-end neural audio coding in the MDCT domain[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10096243, articleTitle=null, refAbstract=null), Reference(id=1327017842106544632, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=19, rfOrder=18, authorNames=null, journalName=null, refType=null, unstructuredReference=Kleijn W B, Lim F S C, Luebs A, et al. Wavenet based low rate speech coding[C]//Proc of the 43rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2018: 676−680, articleTitle=null, refAbstract=null), Reference(id=1327017842161070585, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=20, rfOrder=19, authorNames=null, journalName=null, refType=null, unstructuredReference=Jang Inseon, Yang Haici, Lim W, et al. Personalized neural speech codec[C]//Proc of the 49th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2024: 991−995, articleTitle=null, refAbstract=null), Reference(id=1327017842232373754, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=21, rfOrder=20, authorNames=null, journalName=null, refType=null, unstructuredReference=Du Zhihao, Zhang Shiliang, Hu Kai, et al. Funcodec: A fundamental, reproducible and integrable open-source toolkit for neural speech codec[C]//Proc of the 49th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2024: 591−595, articleTitle=null, refAbstract=null), Reference(id=1327017842303676923, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=22, rfOrder=21, authorNames=null, journalName=null, refType=null, unstructuredReference=Gârbacea C, Oord A, Li Yazhe, et al. Low bit-rate speech coding with VQ-VAE and a WaveNet decoder[C]//Proc of the 44th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2019: 735−739, articleTitle=null, refAbstract=null), Reference(id=1327017842395951612, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=23, rfOrder=22, authorNames=null, journalName=null, refType=null, unstructuredReference=Kleijn W B, Storus A, Chinen M, et al. Generative speech coding with predictive variance regularization[C]//Proc of the 45th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2021: 6478−6482, articleTitle=null, refAbstract=null), Reference(id=1327017842484031997, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=24, rfOrder=23, authorNames=null, journalName=null, refType=null, unstructuredReference=Oord A, Dieleman S, Zen Heiga, et al. WaveNet: A generative model for raw audio[J]. arXiv preprint, arXiv: 1609.03499, 2016, articleTitle=null, refAbstract=null), Reference(id=1327017842572112382, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=25, rfOrder=24, authorNames=null, journalName=null, refType=null, unstructuredReference=Kankanahalli S. End-to-end optimized speech coding with deep neural networks[C]//Proc of the 43rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2018: 2521−2525, articleTitle=null, refAbstract=null), Reference(id=1327017842635026943, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=26, rfOrder=25, authorNames=null, journalName=null, refType=null, unstructuredReference=Oord A, Vinyals O. Neural discrete representation learning[C/OL]//Proc of the 31st Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2017[2025-02-05]. https://dl.acm.org/doi/10.5555/3295222.3295378, articleTitle=null, refAbstract=null), Reference(id=1327017842697941504, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=27, rfOrder=26, authorNames=null, journalName=null, refType=null, unstructuredReference=Défossez A, Copet J, Synnaeve G, et al. High fidelity neural audio compression[J]. arXiv preprint, arXiv: 2210.13438, 2022, articleTitle=null, refAbstract=null), Reference(id=1327017842756661761, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=28, rfOrder=27, authorNames=null, journalName=null, refType=null, unstructuredReference=Ratnarajah A, Zhang Shixiong, Yu Dong. M3-AUDIODEC: Multi-channel multi-speaker multi-spatial audio codec[J]. arXiv preprint, arXiv: 2309.07416, 2023, articleTitle=null, refAbstract=null), Reference(id=1327017842832159234, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=29, rfOrder=28, authorNames=null, journalName=null, refType=null, unstructuredReference=Yang Dongchao, Liu Songxiang, Huang Rongjie, et al. Hifi-codec: Group-residual vector quantization for high fidelity audio codec[J]. arXiv preprint, arXiv: 2305.02765, 2023, articleTitle=null, refAbstract=null), Reference(id=1327017842903462403, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=30, rfOrder=29, authorNames=null, journalName=null, refType=null, unstructuredReference=O’shaughnessy D. Speech Communications: Human and Machine[M]. Piscataway, NJ: IEEE, 1999, articleTitle=null, refAbstract=null), Reference(id=1327017842970571268, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=31, rfOrder=30, authorNames=null, journalName=null, refType=null, unstructuredReference=Davis S, Mermelstein P. Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences[J]. IEEE Transactions on Acoustics, Speech, and Signal processing, 1980, 28(4): 357−366, articleTitle=null, refAbstract=null), Reference(id=1327017843029291525, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=32, rfOrder=31, authorNames=null, journalName=null, refType=null, unstructuredReference=Hasanabadi M R. MFCC-GAN codec: A new AI-based audio coding[J]. arXiv preprint, arXiv: 2310.14300, 2023, articleTitle=null, refAbstract=null), Reference(id=1327017843096400390, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=33, rfOrder=32, authorNames=null, journalName=null, refType=null, unstructuredReference=Hu Jie, Shen Li, Sun Gang. Squeeze-and-excitation networks[C]//Proc of the 31st IEEE Conf on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2018: 7132−7141, articleTitle=null, refAbstract=null), Reference(id=1327017843171897863, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=34, rfOrder=33, authorNames=null, journalName=null, refType=null, unstructuredReference=Zen Heiga, Dang V, Clark R, et al. LibriTTS: A corpus derived from librispeech for text-to-speech[J]. arXiv preprint, arXiv: 1904.02882, 2019, articleTitle=null, refAbstract=null), Reference(id=1327017843251589640, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=35, rfOrder=34, authorNames=null, journalName=null, refType=null, unstructuredReference=Liu Zhaoyu, Mak B. Cross-lingual multi-speaker text-to-speech synthesis for voice cloning without using parallel corpus for unseen speakers[J]. arXiv preprint, arXiv: 1911.11601, 2019, articleTitle=null, refAbstract=null), Reference(id=1327017843318698505, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=36, rfOrder=35, authorNames=null, journalName=null, refType=null, unstructuredReference=Taal C H, Hendriks R C, Heusdens R, et al. A short-time objective intelligibility measure for time-frequency weighted noisy speech[C]//Proc of the 35th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2010: 4214−4217, articleTitle=null, refAbstract=null), Reference(id=1327017843406778890, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=37, rfOrder=36, authorNames=null, journalName=null, refType=null, unstructuredReference=Hines A, Skoglund J, Kokaram A, et al. VISQOL v3: An open source production ready objective speech and audio mestric[J]. arXiv preprint, arXiv: 2004.09584, 2020, articleTitle=null, refAbstract=null), Reference(id=1327017843465499147, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=38, rfOrder=37, authorNames=null, journalName=null, refType=null, unstructuredReference=Kumar R, Seetharaman P, Luebs A, et al. High-fidelity audio compression with improved rvqgan[C/OL]//Proc of the 38th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2024[2025-02-05]. https://openreview.net/forum?id=qjnl1QUnFA, articleTitle=null, refAbstract=null), Reference(id=1327017843524219404, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=39, rfOrder=38, authorNames=null, journalName=null, refType=null, unstructuredReference=Lee S, Ping W, Ginsburg B, et al. BigVGAN: A universal neural vocoder with large-scale training[J]. arXiv preprint. arXiv: 2206.04658, 2022, articleTitle=null, refAbstract=null), Reference(id=1327017843595522573, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=40, rfOrder=39, authorNames=null, journalName=null, refType=null, unstructuredReference=Dietz M, Multrus M, Eksler V, et al. Overview of the EVS codec architecture[C]//Proc of the 40th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2015: 5698–5702, articleTitle=null, refAbstract=null), Reference(id=1327017843654242830, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=41, rfOrder=40, authorNames=null, journalName=null, refType=null, unstructuredReference=高玮玮, 单明陶, 宋楠, 等. 嵌入SENet的改进YOLOv4眼底图像微动脉瘤自动检测算法[J]. 生物医学工程学杂志, 2022, 39(4): 713−720, articleTitle=null, refAbstract=null), Reference(id=1327017843725545999, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=41, rfOrder=41, authorNames=null, journalName=null, refType=null, unstructuredReference=Gao Weiwei, Shan Mingtao, Song Nan, et al. Detection of microaneurysms in fundus images based on improved YOLOv4 with SENet embedded[J]. Journal of Biomedical Engineering, 2022, 39(4): 713−720 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1327017843784266256, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=42, rfOrder=42, authorNames=null, journalName=null, refType=null, unstructuredReference=Chen Qiang, Liu Li, Han Rui, et al. Image identification method on high speed railway contact network based on YOLO v3 and SENet[C]//Proc of the 38th Chinese Control Conf (CCC). Piscataway, NJ: IEEE, 2019: 8772−8777, articleTitle=null, refAbstract=null), Reference(id=1327017843838792209, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=43, rfOrder=43, authorNames=null, journalName=null, refType=null, unstructuredReference=王成龙, 易江燕, 陶建华, 等. 基于全局-时频注意力网络的语音伪造检测[J]. 计算机研究与发展, 2021, 58(7): 1466−1475, articleTitle=null, refAbstract=null), Reference(id=1327017843897512466, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=43, rfOrder=44, authorNames=null, journalName=null, refType=null, unstructuredReference=Wang Chenglong, Yi Jiangyan, Tao Jianhua, et al. Global and temporal-frequency attention based network in audio deepfake detection[J]. Journal of Computer Research and Development, 2021, 58(7): 1466−1475 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1327017843968815635, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=44, rfOrder=45, authorNames=null, journalName=null, refType=null, unstructuredReference=He Kaiming, Zhang Xiangyu, Ren Shaoqing, et al. Deep residual learning for image recognition[C]//Proc of the 29th IEEE Conf on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2016: 770−778, articleTitle=null, refAbstract=null), Reference(id=1327017844031730196, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=45, rfOrder=46, authorNames=null, journalName=null, refType=null, unstructuredReference=Kong J, Kim J, Bae J. Hifi-GAN: Generative adversarial networks for efficient and high fidelity speech synthesis[C]//Proc of the 34th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2020: 17022−17033, articleTitle=null, refAbstract=null), Reference(id=1327017844094644757, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=46, rfOrder=47, authorNames=null, journalName=null, refType=null, unstructuredReference=Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[C/OL]//Proc of the 28th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2014[2025-02-05]. https://www.researchgate.net/publication/263012109_Generative_Adversarial_Networks, articleTitle=null, refAbstract=null), Reference(id=1327017844153365014, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=47, rfOrder=48, authorNames=null, journalName=null, refType=null, unstructuredReference=Jassim W A, Skoglund J, Chinen M, et al. WARP-Q: Quality prediction for generative neural speech codecs[C]//Proc of the 46th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2021: 401−405, articleTitle=null, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)]), AuthorCompany(id=1327017836159021472, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=2, ext=[AuthorCompanyExt(id=1327017836167410081, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Department of Automation, Tsinghua University, Beijing 100084), AuthorCompanyExt(id=1327017836175798690, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2清华大学自动化系 北京 100084)]), AuthorCompany(id=1327017836242907555, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=3, ext=[AuthorCompanyExt(id=1327017836251296164, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084), AuthorCompanyExt(id=1327017836263879077, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3北京信息科学与技术国家研究中心(清华大学) 北京 100084)])], figs=[ArticleFig(id=1327017838524608972, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Figure 1, caption=The structure of residual vector quantization, figureFileSmall=uNYIICEW/I/Wo5Uk6YtXgw==, figureFileBig=gEsCgL3pmKGc/MFnAKDSwQ==, tableContent=null), ArticleFig(id=1327017838591717837, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=图1, caption=残差矢量量化结构, figureFileSmall=uNYIICEW/I/Wo5Uk6YtXgw==, figureFileBig=gEsCgL3pmKGc/MFnAKDSwQ==, tableContent=null), ArticleFig(id=1327017838868541902, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Figure 2, caption=Structure diagram of neural speech codec method based on Mel spectrogram and squeeze-excitation-weighted quantization, figureFileSmall=JnSTGBpv9KMI3mO070Xr4w==, figureFileBig=ZXwCoEE0vmu4EjM4XX+fHg==, tableContent=null), ArticleFig(id=1327017839061479887, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=图2, caption=基于梅尔谱与压缩激励加权量化的语音神经编解码方法结构图, figureFileSmall=JnSTGBpv9KMI3mO070Xr4w==, figureFileBig=ZXwCoEE0vmu4EjM4XX+fHg==, tableContent=null), ArticleFig(id=1327017839241834960, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Figure 3, caption=Comparison of Mel spectrogram, figureFileSmall=4dMjaax+lQpcxE8ojK+bjw==, figureFileBig=nPvfTfmnuitCpZYRh2dIJQ==, tableContent=null), ArticleFig(id=1327017839313138129, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=图3, caption=梅尔谱图对比, figureFileSmall=4dMjaax+lQpcxE8ojK+bjw==, figureFileBig=nPvfTfmnuitCpZYRh2dIJQ==, tableContent=null), ArticleFig(id=1327017839371858386, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Figure 4, caption=Ablation experiments on squeeze-excitation-weighted mechanism (Input characteristic is Mel spectrogram ), figureFileSmall=5EcSUU3VutjQ5TGs0MLECg==, figureFileBig=zTh2vuwkdWvhVXtXg2lVfg==, tableContent=null), ArticleFig(id=1327017839447355859, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=图4, caption=关于压缩激励加权机制的消融实验(输入特征为梅尔谱图), figureFileSmall=5EcSUU3VutjQ5TGs0MLECg==, figureFileBig=zTh2vuwkdWvhVXtXg2lVfg==, tableContent=null), ArticleFig(id=1327017839514464724, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Figure 5, caption=Ablation experiments on squeeze-excitation-weighted mechanism (Input characteristic is waveform), figureFileSmall=nk5AD8POZbk7ma9R4xqZcQ==, figureFileBig=DZsxAaz5wOAz97Vf0RpTzA==, tableContent=null), ArticleFig(id=1327017839589962197, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=图5, caption=关于压缩激励加权机制的消融实验(输入特征为波形), figureFileSmall=nk5AD8POZbk7ma9R4xqZcQ==, figureFileBig=DZsxAaz5wOAz97Vf0RpTzA==, tableContent=null), ArticleFig(id=1327017839652876758, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 1, caption=Network Architecture of Convolutional Unit, figureFileSmall=null, figureFileBig=null, tableContent=
卷积单元卷积核数/步长/空洞数
Conv 14/2/1
Conv 211/1/1
Conv 311/1/1
Conv 411/1/3
Conv 511/1/1
Conv 611/1/5
Conv 711/1/1
Group Norm
Conv 87/1/1
Conv 97/1/1
Conv 107/1/3
Conv 117/1/1
Conv 127/1/5
Conv 137/1/1
Group Norm
Conv 143/1/1
Conv 153/1/1
Conv 163/1/3
Conv 173/1/1
Conv 183/1/5
Conv 193/1/1
Group Norm
), ArticleFig(id=1327017839736762839, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表1, caption=

卷积单元网络架构

, figureFileSmall=null, figureFileBig=null, tableContent=
卷积单元卷积核数/步长/空洞数
Conv 14/2/1
Conv 211/1/1
Conv 311/1/1
Conv 411/1/3
Conv 511/1/1
Conv 611/1/5
Conv 711/1/1
Group Norm
Conv 87/1/1
Conv 97/1/1
Conv 107/1/3
Conv 117/1/1
Conv 127/1/5
Conv 137/1/1
Group Norm
Conv 143/1/1
Conv 153/1/1
Conv 163/1/3
Conv 173/1/1
Conv 183/1/5
Conv 193/1/1
Group Norm
), ArticleFig(id=1327017839820648920, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 2, caption=Metrics Evaluated on the VCTK Dataset for Each Neural Codec, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/
Kbps
码本
数
模型$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑STOI ↑WARP-Q ↓VISQOL ↑MOS ↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
0.751SoundStream18.8170.7032.9882.1341.80
Encodec19.1880.7362.9172.2262.28
Hifi-codec5.5060.6343.0242.0281.88
本文模型36.6560.7472.6042.5362.40
1.52SoundStream15.4380.7362.8602.3152.52
Encodec18.9230.7772.6942.4923.20
Hifi-codec6.5070.7512.7092.5202.84
本文模型36.4470.8012.3712.9583.62
34SoundStream15.7600.7572.7982.4072.36
Encodec18.5500.792.6252.5843.16
Hifi-codec6.5490.8392.1253.3113.52
本文模型35.6320.8461.9973.3343.88
4.56SoundStream20.0360.7902.7112.6513.16
Encodec22.1190.8092.5352.7583.40
Hifi-codec6.5620.8621.9123.5623.62
本文模型22.3380.8591.7983.4923.94
68SoundStream20.1450.7992.5022.6943.50
Encodec21.4230.8181.8522.7993.44
Hifi-codec5.6730.8781.8523.6653.92
本文模型23.4930.8651.7763.5523.98
), ArticleFig(id=1327017839917117913, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表2, caption=

各神经编解码器在VCTK数据集上的评价指标

, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/
Kbps
码本
数
模型$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑STOI ↑WARP-Q ↓VISQOL ↑MOS ↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
0.751SoundStream18.8170.7032.9882.1341.80
Encodec19.1880.7362.9172.2262.28
Hifi-codec5.5060.6343.0242.0281.88
本文模型36.6560.7472.6042.5362.40
1.52SoundStream15.4380.7362.8602.3152.52
Encodec18.9230.7772.6942.4923.20
Hifi-codec6.5070.7512.7092.5202.84
本文模型36.4470.8012.3712.9583.62
34SoundStream15.7600.7572.7982.4072.36
Encodec18.5500.792.6252.5843.16
Hifi-codec6.5490.8392.1253.3113.52
本文模型35.6320.8461.9973.3343.88
4.56SoundStream20.0360.7902.7112.6513.16
Encodec22.1190.8092.5352.7583.40
Hifi-codec6.5620.8621.9123.5623.62
本文模型22.3380.8591.7983.4923.94
68SoundStream20.1450.7992.5022.6943.50
Encodec21.4230.8181.8522.7993.44
Hifi-codec5.6730.8781.8523.6653.92
本文模型23.4930.8651.7763.5523.98
), ArticleFig(id=1327017839996809690, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 3, caption=Evaluation Comparison of Our Method and the General Codec Opus, figureFileSmall=null, figureFileBig=null, tableContent=
模型STOI ↑WARP-Q ↓VISQOL ↑MOS ↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
Opus-10ms0.7032.6102.4763.32
Opus-5.0ms0.8582.3833.0373.58
Opus-2.5ms0.9561.9763.4753.86
本文模型0.8651.7763.5523.98
), ArticleFig(id=1327017840072307163, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表3, caption=

本文方法与通用编解码器Opus的评测对比

, figureFileSmall=null, figureFileBig=null, tableContent=
模型STOI ↑WARP-Q ↓VISQOL ↑MOS ↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
Opus-10ms0.7032.6102.4763.32
Opus-5.0ms0.8582.3833.0373.58
Opus-2.5ms0.9561.9763.4753.86
本文模型0.8651.7763.5523.98
), ArticleFig(id=1327017840143610332, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 4, caption=Experimental Configuration of Different Compression Rates, figureFileSmall=null, figureFileBig=null, tableContent=
压缩率编码器下采样倍数解码器上采样倍数解码器各层卷积核
18025, 4, 4, 210, 8, 8, 4
32028, 5, 3, 216, 11, 7, 4
640210, 8, 4, 220, 16, 8, 4
), ArticleFig(id=1327017840219107805, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表4, caption=

不同压缩率的实验配置

, figureFileSmall=null, figureFileBig=null, tableContent=
压缩率编码器下采样倍数解码器上采样倍数解码器各层卷积核
18025, 4, 4, 210, 8, 8, 4
32028, 5, 3, 216, 11, 7, 4
640210, 8, 4, 220, 16, 8, 4
), ArticleFig(id=1327017840282022366, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 5, caption=Ablation Experiments on Different Compression Ratios and Codebook Count, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps压缩率码本数STOI ↑WARP-Q ↓VISQOL↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.

3
18020.8432.0283.314
32040.8461.9973.334
64080.8541.8523.427

6
18040.8981.6083.786
32080.8651.7763.552
640160.8351.8483.387
), ArticleFig(id=1327017840349131231, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表5, caption=

对于不同压缩率和码本数量的消融实验

, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps压缩率码本数STOI ↑WARP-Q ↓VISQOL↑
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.

3
18020.8432.0283.314
32040.8461.9973.334
64080.8541.8523.427

6
18040.8981.6083.786
32080.8651.7763.552
640160.8351.8483.387
), ArticleFig(id=1327017840420434400, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 6, caption=Ablation Experiments on Squeeze-Excitation-Weighted Mechanism (Input Characteristic is Mel Spectrogram), figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Mel-Input+SE-weight(本文)Mel-Input
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
1.52$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.44742.515
STOI ↑0.8010.795
WARP-Q ↓2.3712.318
VISQOL ↑2.9582.926
3.04$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑35.63244.498
STOI ↑0.8460.837
WARP-Q ↓1.9972.035
VISQOL ↑3.3343.310
6.08$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑23.49332.345
STOI ↑0.8650.866
WARP-Q ↓1.7761.733
VISQOL ↑3.5523.499
), ArticleFig(id=1327017840487543265, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表6, caption=

压缩激励加权机制的消融实验(输入特征为梅尔谱图)

, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Mel-Input+SE-weight(本文)Mel-Input
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
1.52$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.44742.515
STOI ↑0.8010.795
WARP-Q ↓2.3712.318
VISQOL ↑2.9582.926
3.04$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑35.63244.498
STOI ↑0.8460.837
WARP-Q ↓1.9972.035
VISQOL ↑3.3343.310
6.08$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑23.49332.345
STOI ↑0.8650.866
WARP-Q ↓1.7761.733
VISQOL ↑3.5523.499
), ArticleFig(id=1327017840554652130, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 7, caption=Ablation Experiments on Squeeze-Excitation-Weighted Mechanism (Input Characteristic is Waveform), figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Wave-Input+SE-weightWave-Input
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
1.52$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.5276.507
STOI ↑0.7530.751
WARP-Q ↓2.5262.709
VISQOL ↑2.6302.520
3.04$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.7026.549
STOI ↑0.8370.839
WARP-Q ↓2.1472.125
VISQOL ↑3.3293.311
6.08$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.3695.673
STOI ↑0.8710.878
WARP-Q ↓2.0201.852
VISQOL ↑3.4773.665
), ArticleFig(id=1327017840642732515, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表7, caption=

压缩激励加权机制的消融实验(输入特征为波形)

, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Wave-Input+SE-weightWave-Input
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
1.52$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.5276.507
STOI ↑0.7530.751
WARP-Q ↓2.5262.709
VISQOL ↑2.6302.520
3.04$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.7026.549
STOI ↑0.8370.839
WARP-Q ↓2.1472.125
VISQOL ↑3.3293.311
6.08$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑5.3695.673
STOI ↑0.8710.878
WARP-Q ↓2.0201.852
VISQOL ↑3.4773.665
), ArticleFig(id=1327017840722424292, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=EN, label=Table 8, caption=Ablation Experiments on Activation Function, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Relu(本文)Snake
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
0.751$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.65611.590
STOI ↑0.7470.746
WARP-Q ↓2.6042.568
VISQOL ↑2.5362.538
1.502$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.44711.784
STOI ↑0.8010.810
WARP-Q ↓2.3712.224
VISQOL ↑2.9583.073
3.004$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑35.63210.821
STOI ↑0.8460.851
WARP-Q ↓1.9971.989
VISQOL ↑3.3343.234
4.506$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑22.3385.915
STOI ↑0.8590.892
WARP-Q ↓1.7981.563
VISQOL ↑3.4923.638
6.008$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑23.4934.967
STOI ↑0.8650.900
WARP-Q ↓1.7761.448
VISQOL ↑3.5523.658
), ArticleFig(id=1327017840797921765, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, language=CN, label=表8, caption=

关于激活函数的消融实验

, figureFileSmall=null, figureFileBig=null, tableContent=
比特率/Kbps码本数指标Relu(本文)Snake
注:黑体数值表示最优结果,“↑”表示数值越大越好,“↓”表示数值越小越好.
0.751$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.65611.590
STOI ↑0.7470.746
WARP-Q ↓2.6042.568
VISQOL ↑2.5362.538
1.502$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑36.44711.784
STOI ↑0.8010.810
WARP-Q ↓2.3712.224
VISQOL ↑2.9583.073
3.004$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑35.63210.821
STOI ↑0.8460.851
WARP-Q ↓1.9971.989
VISQOL ↑3.3343.234
4.506$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑22.3385.915
STOI ↑0.8590.892
WARP-Q ↓1.7981.563
VISQOL ↑3.4923.638
6.008$ {\mathrm{R}\mathrm{T}\mathrm{F}}_{\mathrm{e}\mathrm{n}\mathrm{c}} $↑23.4934.967
STOI ↑0.8650.900
WARP-Q ↓1.7761.448
VISQOL ↑3.5523.658
)], attaches=null, journal=Journal(id=1324759606947315908, delFlag=0, nameCn=计算机研究与发展, nameEn=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, issn=1000-1239, eissn=null, cn=11-1777/TP, coden=null, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=k2l8zgqwstuxxevFX2qxVA==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Computer Research and Development, journalRemark=null, publicationField=null, createdTime=1793776903141, updatedTime=1793850698612, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=J, firstLetterEn=J, subjectCode=Electronic Information, subjectName=null, subjectCodeEn=Electronic Information, subjectNameEn=null, picCn=k2l8zgqwstuxxevFX2qxVA==, picEn=r6k0Ry+RD+c/VXj5QarmDw==, jcr=null, cjcr=null, exts=[JournalExt(id=1325069127710286688, language=CN, name=计算机研究与发展, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698640, updatedTime=1793850698640, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1325069127777395553, language=EN, name=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698656, updatedTime=1793850698656, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1324759914146529478, websiteList=[Website(id=1325070300139892771, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN, language=CN, createTime=1793850978169, createBy=18614031015, updateTime=1793851040259, updateBy=18614031015, name=计算机研究与发展-中文, tplId=1146099689490845704, title=计算机研究与发展, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070893751349360, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=articleTextType, value=kx, createTime=1793851119697, updateTime=1793851119697, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893730377837, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=banner, value=null, createTime=1793851119692, updateTime=1793851119692, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893772320883, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=grayFlag, value=0, createTime=1793851119702, updateTime=1793851119702, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893721989228, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851119690, updateTime=1793851119690, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893784903797, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=minRunFlag, value=0, createTime=1793851119705, updateTime=1793851119705, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893742960751, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic, createTime=1793851119695, updateTime=1793851119695, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893780709492, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=silenceFlag, value=0, createTime=1793851119704, updateTime=1793851119704, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893738766446, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1793851119694, updateTime=1793851119694, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893755543665, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeColor, value=null, createTime=1793851119698, updateTime=1793851119698, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893763932274, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeStyle, value=null, createTime=1793851119700, updateTime=1793851119700, creator=18614031015, updator=18614031015)]), Website(id=1325070300580294698, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN, language=EN, createTime=1793850978274, createBy=18614031015, updateTime=1793851036638, updateBy=18614031015, name=计算机研究与发展-英文, tplId=1146101810881728533, title=Journal of Computer Research and Development, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070925288321148, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=articleTextType, value=kx, createTime=1793851127216, updateTime=1793851127216, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925263155321, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=banner, value=null, createTime=1793851127210, updateTime=1793851127210, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925309292671, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=grayFlag, value=0, createTime=1793851127221, updateTime=1793851127221, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925254766712, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851127208, updateTime=1793851127208, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925326069889, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=minRunFlag, value=0, createTime=1793851127225, updateTime=1793851127225, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925279932539, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic, createTime=1793851127214, updateTime=1793851127214, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925317681280, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=silenceFlag, value=0, createTime=1793851127223, updateTime=1793851127223, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925271543930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1793851127212, updateTime=1793851127212, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925292515453, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeColor, value=null, createTime=1793851127217, updateTime=1793851127217, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925300904062, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeStyle, value=null, createTime=1793851127219, updateTime=1793851127219, creator=18614031015, updator=18614031015)])], journalTitle=计算机研究与发展, weixinUrl=null, journalUrl=https://crad.ict.ac.cn/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Computer Research and Development, journalPhotoCn=k2l8zgqwstuxxevFX2qxVA==, journalPhotoEn=r6k0Ry+RD+c/VXj5QarmDw==, journalFirstLetter=J, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false, interPubPlatform=, interPubPlatformUrl=null), detailUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202440329, detailUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/10.7544/issn1000-1239.202440329, pdfUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/PDF/10.7544/issn1000-1239.202440329, pdfUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/PDF/10.7544/issn1000-1239.202440329, aliStartDate=0, aliEndDate=0, collectionFlag=false, citedCount=null, citedUrl=null, previewStatus=0, delFlag=0, hasFullText=1, orderTime=1761926400000, fullTextJson=null, articleText=null, reference=null)
收藏切换
基于梅尔谱与压缩激励加权量化的语音神经编解码方法
周俊佐 1 , 易江燕 1, * , 陶建华 2, 3 , 任勇 1 , 汪涛 1
计算机研究与发展 | 人工智能 2025,62(11): 2725-2736
收起
收藏切换
计算机研究与发展 |人工智能 2025 , 62 (11) : 2725 -2736
基于梅尔谱与压缩激励加权量化的语音神经编解码方法
全屏
[Author(id=1327017836339376551, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=zhoujunzuo2023@ia.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017836469399977, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836339376551, language=EN, stringName=Junzuo Zhou, firstName=Junzuo, middleName=null, lastName=Zhou, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 2000. Master candidate. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 2000. Master candidate. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017836532314538, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836339376551, language=CN, stringName=周俊佐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"G6Cc6ZXkwbYOhylFlTXI1g==","content":"

2000年生. 硕士研究生. 主要研究方向为语音合成

"}, bioImg=G6Cc6ZXkwbYOhylFlTXI1g==, bioContent=

2000年生. 硕士研究生. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017836612006316, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=jiangyan.yi@nlpr.ia.ac.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1327017836695892398, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836612006316, language=EN, stringName=Jiangyan Yi, firstName=Jiangyan, middleName=null, lastName=Yi, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

"}, bioImg=null, bioContent=

born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017836788167087, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836612006316, language=CN, stringName=易江燕, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, *, address=1中国科学院自动化研究所 北京 100190, bio={"img":"Y6IZUg19571lrS3SG6x3Ew==","content":"

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

"}, bioImg=Y6IZUg19571lrS3SG6x3Ew==, bioContent=

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017836872053169, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017836989493684, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836872053169, language=EN, stringName=Jianhua Tao, firstName=Jianhua, middleName=null, lastName=Tao, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, address=2Department of Automation, Tsinghua University, Beijing 100084
3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084, bio={"content":"

born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

"}, bioImg=null, bioContent=

born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837073379766, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017836872053169, language=CN, stringName=陶建华, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, 3, address=2清华大学自动化系 北京 100084
3北京信息科学与技术国家研究中心(清华大学) 北京 100084, bio={"img":"ar6mn702nAnH7GRimHQ9BQ==","content":"

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

"}, bioImg=ar6mn702nAnH7GRimHQ9BQ==, bioContent=

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836159021472, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=2, ext=[AuthorCompanyExt(id=1327017836167410081, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2Department of Automation, Tsinghua University, Beijing 100084), AuthorCompanyExt(id=1327017836175798690, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836159021472, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2清华大学自动化系 北京 100084)]), AuthorCompany(id=1327017836242907555, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=3, ext=[AuthorCompanyExt(id=1327017836251296164, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084), AuthorCompanyExt(id=1327017836263879077, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836242907555, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3北京信息科学与技术国家研究中心(清华大学) 北京 100084)])]), Author(id=1327017837161460152, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017837245346234, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837161460152, language=EN, stringName=Yong Ren, firstName=Yong, middleName=null, lastName=Ren, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1998. PhD candidate. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 1998. PhD candidate. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837329232315, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837161460152, language=CN, stringName=任勇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"+HumlNn1XAhshqsdXs32aw==","content":"

1998年生. 博士研究生. 主要研究方向为语音合成

"}, bioImg=+HumlNn1XAhshqsdXs32aw==, bioContent=

1998年生. 博士研究生. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])]), Author(id=1327017837396341181, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1327017837484421567, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837396341181, language=EN, stringName=Tao Wang, firstName=Tao, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190, bio={"content":"

born in 1996. PhD. His main research interest includes speech synthesis

"}, bioImg=null, bioContent=

born in 1996. PhD. His main research interest includes speech synthesis

, aboutCorrespAuthor=null), CN=AuthorExt(id=1327017837551530432, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, authorId=1327017837396341181, language=CN, stringName=汪涛, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1中国科学院自动化研究所 北京 100190, bio={"img":"WqHuMxLwQwmSAuPYeAPuHQ==","content":"

1996年生. 博士. 主要研究方向为语音合成

"}, bioImg=WqHuMxLwQwmSAuPYeAPuHQ==, bioContent=

1996年生. 博士. 主要研究方向为语音合成

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1327017836075135389, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, xref=1, ext=[AuthorCompanyExt(id=1327017836083523998, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1Institute of Automation, Chinese Academy of Sciences, Beijing 100190), AuthorCompanyExt(id=1327017836096106911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1327017833927651718, companyId=1327017836075135389, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1中国科学院自动化研究所 北京 100190)])])]
周俊佐1 , 易江燕1, * , 陶建华2, 3, 任勇1, 汪涛1
作者信息
  • 1中国科学院自动化研究所 北京 100190
  • 2清华大学自动化系 北京 100084
  • 3北京信息科学与技术国家研究中心(清华大学) 北京 100084
通讯作者:
作者简介:

2000年生. 硕士研究生. 主要研究方向为语音合成

1984年生. 博士,硕士生导师. 主要研究方向为语音信息处理、语音生成与鉴别、持续学习

1972年生. 博士,博士生导师. 主要研究方向为智能信息融合与处理、语音处理、情感计算、大数据分析

1998年生. 博士研究生. 主要研究方向为语音合成

1996年生. 博士. 主要研究方向为语音合成

Neural Speech Codec Method Based on Mel Spectrogram and Squeeze-Excitation-Weighted Quantization
Junzuo Zhou1 , Jiangyan Yi1, * , Jianhua Tao2, 3, Yong Ren1, Tao Wang1
Affiliations
  • 1Institute of Automation, Chinese Academy of Sciences, Beijing 100190
  • 2Department of Automation, Tsinghua University, Beijing 100084
  • 3Beijing National Research Center for Information Science and Technology (Tsinghua University), Beijing 100084
  • born in 2000. Master candidate. His main research interest includes speech synthesis

    born in 1984. PhD, master supervisor. Her main research interests include speech information processing, speech generation and identification, and continuous learning

    born in 1972. PhD, PhD supervisor. His main research interests include intelligent information fusion and processing, speech processing, affective computing, and big data analysis

    born in 1998. PhD candidate. His main research interest includes speech synthesis

    born in 1996. PhD. His main research interest includes speech synthesis

出版时间: 2025-11-01 doi: 10.7544/issn1000-1239.202440329
文章导航
收藏切换

目前,以 SoundStream 等为代表的端到端语音神经编解码器在重建语音感知质量方面展现了优异性能. 然而,这些方法需要大量的卷积计算,从而导致较长的编码时间. 为缓解上述问题,提出基于梅尔谱和压缩激励加权量化的神经语音编解码方法. 该方法旨在保持较高语音感知质量的同时降低计算代价,加快运行速度,从而减少时延. 具体而言,采用梅尔谱特征作为输入,借助梅尔谱提取过程中时域压缩的性质,并结合低层卷积编码器以简化运算过程. 此外,借鉴压缩激励网络思想,提取了编码器最后一层输出特征各维度的激励权重,将其作为量化器中计算码本距离时压缩特征各维度的权重系数,由此学习特征间的相关性并优化量化性能. 在 LibriTTS 和 VCTK 数据集上的实验结果表明,该方法显著提升了编码器计算速度,且能在较低比特率时(≤3 Kbps)提升重建语音质量. 以比特率1.5 Kbps时为例,编码计算实时率(real-time factor,RTF)最多可提升4.6倍. 对于感知质量,以0.75 Kbps为例,短时客观可懂度(short-time objective intelligibility,STOI)、虚拟语音质量客观评估(virtual speech quality objective listener,VISQOL)等客观指标相较基线平均可提升8.72%. 此外,消融实验不仅表明压缩激励权重方法的优化效果与比特率呈反相关,而且发现Relu激活函数相较周期性质激活函数Snake而言,在语音感知质量相当的情况下,能大量加快运行速度.

语音编解码  /  梅尔谱图  /  压缩激励网络  /  残差矢量量化  /  生成对抗网络

At present, end-to-end speech neural codecs, represented by SoundStream, have demonstrated outstanding performance in reconstructed speech quality. However, these methods require extensive convolutional computations, leading to lengthy encoding times. To address this issue, we introduce a neural speech codec method based on Mel spectrogram and squeeze-excitation-weighted quantization. This method aims to maintain high speech perceptual quality while reducing computational costs and increasing operational speed, thereby minimizing latency. Specifically, we utilize Mel spectrogram features as input, capitalize on the temporal compression properties during Mel spectrogram extraction, and combine a lower-layer convolutional encoder to simplify the computation process. Additionally, inspired by squeezed excitation network concepts, we extract excitation weights for each dimension of the output features from the encoder’s final layer. These weights are used as the weighting coefficients for each dimension of the compressed features when calculating codebook distances in the quantizer, thus enabling the learning of correlations between features and enhancing the performance of quantization. Experimental results on the LibriTTS and VCTK datasets indicate that this method significantly enhances the computational speed of the encoder and improves the reconstructed speech quality at lower bit rates (≤3 Kbps). For instance, at a bitrate of 1.5 Kbps, the real-time factor (RTF) of encoding computations can increase by up to 4.6 times. Regarding perceptual quality, at a bitrate of 0.75 Kbps, objective metrics such as short-time objective intelligibility (STOI) and virtual speech quality objective listener (VISQOL) show an average improvement of 8.72% compared with the baseline. Additionally, ablation studies not only demonstrate that the optimization effect of compressed excitation weight methods is inversely correlated with bitrate, but also reveal that, compared with the periodic activation function Snake, the Relu activation function can significantly speed up processing while maintaining comparable speech perceptual quality.

speech codec  /  Mel-spectrogram  /  squeeze-and-excitation networks  /  residual vector quantization  /  generative adversarial network
周俊佐, 易江燕, 陶建华, 任勇, 汪涛. 基于梅尔谱与压缩激励加权量化的语音神经编解码方法. 计算机研究与发展, 2025 , 62 (11) : 2725 -2736 . DOI: 10.7544/issn1000-1239.202440329
Junzuo Zhou, Jiangyan Yi, Jianhua Tao, Yong Ren, Tao Wang. Neural Speech Codec Method Based on Mel Spectrogram and Squeeze-Excitation-Weighted Quantization[J]. Journal of Computer Research and Development, 2025 , 62 (11) : 2725 -2736 . DOI: 10.7544/issn1000-1239.202440329
语音编解码是移动通信、互联网通信等众多领域的重要技术之一[1-5]. 语音信号在传输过程中经历了发送端的信号处理和特征提取,随后通过数据压缩传输至接收端,最终接收端通过解码器将恢复的特征解码重建成语音波形,这构成了典型的语音编解码系统. 其包括编码器、量化器和解码器3个模块. 传统的语音编解码器大多基于数字信号处理方法,针对不同适用条件结合一些专家知识精心设计和选择构建,例如利用心理声学和语音合成等领域的知识来提高编码效率等[6-10]. 然而,这些方法不仅适用性受到限制,其生成的语音质量也有限[11]. 以神经网络为代表的机器学习方法最初仅被应用于语音降噪等编解码的后处理阶段[12-13]. 随着“数据驱动”模式下深度学习技术的进步,这些方法从辅助优化逐渐转变为编解码器本身的核心组件之一,不仅便于设计,而且展现了出色的性能,在不同网络带宽条件下均能解码得到较高质量的重建语音[14-21].
由于解码过程与语音合成领域的声码器同属于波形的生成过程,因此,早期的工作尝试直接用神经声码器模型实现语音解码[22-23]. 2023年,Petermann等人[17]使用 WaveNet[24]作为语音生成模型,利用其自回归生成能力,显著提高了解码语音的质量. 2018年,Kankanahalli[25]实现了神经语音编解码系统的端到端优化,该系统无需手动特征工程,全面优化宽带语音编码管道中的各个步骤(包括压缩、量化和解压缩),大幅提升了系统的适应性. 2019年,Gârbacea等人[22]同时采用基于矢量量化变分自动编码 VQ-VAE[26]和 WaveNet 解码器的神经网络架构进行语音编解码. VQ-VAE通过将编码特征离散化以完成数据压缩过程,提升了量化器的性能. 2021年,Kleijn等人[23]推出了 Lyra,对经过KLT变换(Karhunen–Loève transform,KLT)的梅尔谱进行矢量量化,并采用WaveGRU作为解码器,实现了高效的波形恢复. Zeghidour等人[11]开发的 SoundStream 利用生成对抗网络(generative adversarial network,GAN)的模式进行对抗性训练,并引入残差矢量量化的技巧,使单个模型能够处理不同的比特率,从而适用于各种网络带宽. 此外,这还是一种对波形采用全卷积编解码器的端到端编解码系统. 2022年,Défossez等人[27]开发的Encodec在SoundStream基础上更进一步,引入了轻量级的语言模型、熵编码等技术进行优化,且可以通过分别处理左右声道来压缩立体声音频. 而Ratnarajah等人[28]则专为高效压缩多声道语音提出M3-AUDIODEC,展示了神经编解码器在多声道音频编码上的显著提升. 2023年,Yang等人[29]在SoundStream的基础上提出Hifi-codec,采用分组残差矢量量化方法,进一步提升了重建语音质量. 上述语音神经编解码方法普遍采用卷积编码器直接从波形中学习特征,获得了很好的重建语音质量,但卷积编码器通过调节卷积单元中下采样倍数完成时域帧的逐步压缩,在提取优秀潜在特征的同时,以一定的卷积计算量为代价.
梅尔谱特征作为声学领域的经典手工特征,符合人耳听觉的感知特性[30-31]. 尽管过去的一些编解码器方法[23]曾利用梅尔谱图作为编码器输入,但其缺乏量化方法优化[32],相应的编解码设计使得相较SoundStream等波形方法重建音频质量有所不足.
为了在保持重建语音质量的同时降低语音编码计算开销,本文提出了基于梅尔谱与压缩激励加权量化的语音神经编解码方法. 由于梅尔谱的提取过程伴随着时域帧的逐步压缩,因此,本文对梅尔谱运用卷积编码器进行特征提取,采用的卷积层数更少;降低计算规模,从而减少时延,以平衡更加多样的用户需求和更为稳定的用户体验. 对于量化器,由于卷积编码器各个输出通道信息量具有差异[33],该不均匀性将影响矢量量化过程中各个通道维度的重要程度. 因此,本文借鉴了压缩激励网络(squeeze-and-excitation networks,SENet)思想[26],提取编码器最后一层各维度的激励权重,使其作为量化器中计算码本距离时各维度的权重系数. 即通过编码器的自适应学习捕捉特征之间的相互依赖性,减少冗余信息[26],从而确定对量化器更重要的通道,提升量化性能.
本文在LibriTTS[34]和VCTK[35]数据集上进行实验,结果表明使用梅尔谱图作为输入特征,结合低层卷积编码器,且采用压缩激励的方法利用梅尔谱图经过低层编码器输出特征信息量的不均匀性,不仅普遍可以降低时延,并且能在较低比特率环境中($ \le 3\;\mathrm{K}\mathrm{b}\mathrm{p}\mathrm{s} $)提升感知质量. 以感知质量较好的基线Hifi-codec[29]为基准,比特率为1.5 Kbps时,本文方法编码计算的实时率(real time factor,RTF)最多可提升至4.6倍. 此外,较低比特率的情况下,本文方法的感知质量超越所有基准模型. 特别是在0.75 Kbps时,与最佳的Encodec[27]相比,短时客观可懂度(short-time objective intelligibility,STOI)[36]和虚拟语音质量客观评估(virtual speech quality objective listener,VISQOL)[37]的平均提升率为8.72%.
本文探究了不同比特率下各压缩率与码本数目组合的消融实验. 发现比特率提高时,增加压缩率不利于模型的训练效果和性能,应当同时兼顾码本数目进行参数选择. 此外,通过对不同比特率和不同输入特征进行压缩激励权重的消融实验,本文探索了压缩激励权重方法的性质. 其优化效果与比特率呈反相关,且相较波形特征,更适用于梅尔谱编解码器. 最后,本文还对神经解码器网络中的激活函数进行了消融实验,通过比较Relu激活函数和具有周期特性的Snake激活函数[38-39],研究结果表明:在保持语音质量相当的情况下,Relu激活函数能显著提高运行速度. 特别是在低比特率环境下($ \le 3\;\mathrm{K}\mathrm{b}\mathrm{p}\mathrm{s} $),Relu函数在保持语音质量几乎不变的同时,实时率(RTF)均能提升2倍以上,因此更适合实际需求.
本节从梅尔谱图、压缩激励网络、残差矢量量化3个方面介绍相关工作.
梅尔谱图作为一种语音处理中常用的前端特征,其原理根植于人耳对频率感知的非线性特性,特别是对低频信号更为敏感. 为了模拟这种听觉特性,人们引入了梅尔标度,这种非线性对数变换针对频率标度进行转换,将语音语谱图的频率维度应用梅尔标度即为梅尔谱图[30,40]. 该转换过程涉及语谱图与多个梅尔滤波器相乘,而语音的语谱图则源自对语音序列进行短时傅里叶变换并提取幅度谱. 在短时傅里叶变换中,帧移操作对时域信号进行压缩,因此梅尔谱图的时域帧数远远少于原始波形数据点,为后续的编解码过程提供了便利.
众所周知,卷积网络能够通过融合各层局部感受野中的空间和信道信息来构造信息特征. 压缩激励网络(SENet)[26]的本质是注意力机制在卷积网络领域中的应用. 作为一个简化的结构, 它可以插入卷积网络中,并对通道关系之间的相互依赖性加以关注,此前在视觉等众多领域其适用性都得到了验证[41-42],这种子结构也适用于语音场景[43]. 具体而言,通过对卷积的每个输出通道预测一个常数权重,并对该通道加权. 这种注意力机制以轻量级的计算代价让模型更偏向信号的最具信息量的部分.
在语音编解码器的量化环节中,编码器提取的特征需要被压缩,矢量量化是常用的方法之一[22]. 通过建立一张码本,将连续的特征空间转化为离散的token. 采用的码本数目越多,解码器恢复的语音质量就越高,但更多的码本也会消耗传输时的网络带宽. 因此,对于神经编解码器,不同的网络带宽环境需要专门训练各自数目的码本,这大大增加了使用和训练的成本. 对此,人们设计了残差矢量量化 (residual vector quantization,RVQ)[11]结构,如图1所示. 该结构通过将上一量化层$ {q}_{i}(\cdot ) $离散化后与其输入标准值$ {y}_{i} $的残差作为下一量化层$ {q}_{i+1}(\cdot ) $的输入,最终累加各量化层离散化后的token作为输入连续特征$ x $的重建值$ {x'} $,即
$ {y}_{1}=x\text{,} $
$ {y}_{i+1}={q}_{i+1}\left({y}_{i}-{q}_{i}\left({y}_{i}\right)\right)\text{,} $
$ {x'}=\sum _{i=1}^{N}{y}_{i}\text{,} $
其中量化层$ {q}_{i}\left(\cdot \right) $是在码本空间中选择距离最近的条目,由于传输时仅记录条目的序号,从而实现信息压缩的效果. 上述过程,残差堆积层数$ N $越多,离散化精度也越高. 此外,残差结构具有相当的灵活性. 可以1次训练多层码本,而推理时根据带宽限制只选用前几层($ n\le N $)进行累积重建,大大扩宽了应用范围.
为了加强码本空间的表示能力,可以采用对同一量化层码本拆分成多组的形式,即分组矢量残差量化(group residual vector quantization,GRVQ)方法[29],进一步深化了量化能力.
本文工作中将使用压缩激励机制对量化部分进行优化.
本文采取的系统整体框架如图2所示,生成部分分为编码器、量化器和解码器3个组件. 此外,还包括鉴别器部分. 语音波形提取梅尔谱后,输入卷积编码器. 卷积编码器遵循与 Hifi-codec 相似的结构[29],但采用较少的卷积模块数目,即由首尾的1维卷积层和中间的卷积单元组成. 每个卷积单元由3个残差单元[44]和1个下采样层组成[11]. 解码器则由首尾的1维卷积层和中间4个卷积单元组成[11].
量化器接受编码器的输出特征和压缩激励权重. 在编码器压缩率给定的情况下,灵活调节码本数目以实现不同比特率下的调节. 在比特率相对较高时(需要4个及以上的码本数目时),采取 Hifi-codec 的策略,即GRVQ,为了训练的稳定,组数设置为2. 而码本数目低于4个时,则取消分组,退化为普通残差量化.
对于鉴别器,本实验参考声码器 Hifi-GAN的策略,采用对语音波形直接检验的多尺度鉴别器(multi-scale discriminator,MSD)与多周期鉴别器(multi-period discriminator,MPD)[45]. 前者是通过核大小为4的步幅平均池化(strided average pooling)层对波形序列进行2倍和4倍的下采样之后的语音序列进行操作. 后者是将1维的原始语音序列$ T $按照固定周期$ p $抽样为长为$ p $、宽为$ T/p $的2维数据,然后对重塑后的数据应用2维卷积. 鉴别器最终输出判定情况与中间层检测时的特征图,以用于后续生成对抗损失和特征损失计算.
编解码系统中压缩率是从波形采样点到进入量化器前时间帧的下采样倍数[27],由于本文采用梅尔谱特征,因此帧移(hop_size)本身具有压缩的作用. 编码器卷积单元中下采样层卷积核的采样步幅(stride)则承担了剩余的压缩效果. 解码器卷积单元中所有的上采样倍数均来自卷积核的采样步幅. 因此需要满足:
$ r=hop\_size\times \prod _{i}{S}_{\mathrm{e}\mathrm{n}\mathrm{c},i}=\prod _{i}{S}_{\mathrm{d}\mathrm{e}\mathrm{c},i,} $
其中$ r $是压缩率,$ {S}_{\mathrm{e}\mathrm{n}\mathrm{c},i} $和 $ {S}_{\mathrm{d}\mathrm{e}\mathrm{c},i} $分别指编码器和解码器第$ i $个卷积单元中下采样层卷积核的采样步幅.
式(4)的成立使得编码器端所需的卷积单元数量相较于直接从波形中进行下采样时减少. 尽管计算梅尔谱图的过程包含短时傅里叶以及梅尔滤波器相乘的运算,这也会带来一定的时间消耗,但其计算量低于具有相同压缩程度的卷积单元所进行的卷积运算,后续的实验证明了这一点.
对于量化过程中的压缩激励权重. 本文希望可以通过动态地分配通道维度的权重来提高网络表达能力,具体而言,我们对卷积编码器的最后一层进行全局平均池化,全局空间信息被聚合到一个信道描述符中,即压缩过程. 继而通过先后2个全连接及中间的Relu层,接连Sigmoid函数,该步骤能学习被挤压通道之间的非线性交互,即激励过程[26]. 最终得到与编码器输出通道数 $ {C}_{\mathrm{e}\mathrm{n}\mathrm{c}} $ 相同的1维向量作为各通道评价分数($ {S core}_{1},{S core}_{2},… ,{S core}_{{C}_{\mathrm{e}\mathrm{n}\mathrm{c}}} $). 如图2所示,作为通道注意力权重,该分数并非与编码器最后1层相乘,而是保留进入量化器模块. 尽管GRVQ中分组的操作使得量化时实际的码本嵌入宽度与待查询的编码器输出通道数$ {C}_{\mathrm{e}\mathrm{n}\mathrm{c}} $并不相等. 但是每轮残差后,各个分组码本会合并,合并后的总码本嵌入宽度 $ W= {C}_{\mathrm{e}\mathrm{n}\mathrm{c}} $. 某帧($ {x}_{1},{x}_{2},… {,x}_{{C}_{\mathrm{e}\mathrm{n}\mathrm{c}}} $)查询码本时,需要计算该帧与码本中各条目($ {c}_{1},{c}_{2},… , c_{W} $)的距离并选取最近的条目,我们定义该距离为
$ d=\sum _{i=1}^{W}{{S core}_{i}\times \left({x}_{i}-{c}_{i}\right)}^{2}. $
本文实验整体框架采用GAN结构,分为鉴别器损失和生成器损失2部分[46].
对于包含多个子鉴别器的鉴别器组,$ K $是鉴别器数目, $ {D}_{i} $代表多周期鉴别器(MPD)或多尺度鉴别器(MSD)的第$ i $个子鉴别器,$ x $ 为原始波形样本,$ \hat{x} $ 即生成器输出的波形,定义鉴别器的对抗损失:
$ {L}_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{D}}}=\frac{1}{K}\sum _{i=1}^{K}{\left[\right({D}_{i}\left(x\right)-1)}^{2}+D{\left(\hat{x}\right)}^{2}]. $
同时,生成器的对抗损失:
$ {L}_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{G}}}=\frac{1}{K}\sum _{i=1}^{K}{\left({D}_{i}\left(\hat{x}\right)-1\right)}^{2}. $
定义生成部分的重建波形的梅尔谱图$ \hat{m} $与原始样本梅尔谱图$ m $之间的L1距离为重建损失(reconstruction loss)$ {L}_{\mathrm{r}\mathrm{e}\mathrm{c}} $:
$ {L}_{\mathrm{r}\mathrm{e}\mathrm{c}}=\|m-\hat{m}\|. $
此外,对于鉴别器$ {D}_{i} $的第 $ l $ 个中间层,$ w $为其输入的中间特征,$ \hat{w} $为该层输出,我们定义总体的E特征损失(feature loss)$ {L}_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}} $[27]:
$ {L}_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}}=\frac{1}{KL}\sum _{i=1}^{K}\sum _{l=1}^{L}\frac{\|{D}_{i}^{l}\left(w\right)-{D}_{i}^{l}\left(\hat{w}\right)\|}{mean\left(\|{D}_{K}^{l}\left(w\right)\|\right)}. $
量化过程中,第$ n $组第$ i $个量化器 $ {q}_{n,i} $,对于其待查帧 $ {z}_{n,i} $,定义承诺损失(commitment loss)$ {L}_{\mathrm{c}} $[27]:
$ {L}_{{\mathrm{c}}}=\sum _{n,i}{\|{z}_{n,i}-{q}_{n,i}({z}_{n,i})\|}_{2}^{2}. $
综上所述,整体损失定义为
$ {{L}_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}={\lambda }_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}}L}_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}}+{\lambda }_{\mathrm{c}}{L}_{\mathrm{c}}+{\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{G}}}{L}_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{G}}}\\ +{\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{D}}}{L}_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{D}}}+{\lambda }_{\mathrm{r}\mathrm{e}\mathrm{c}}{L}_{\mathrm{r}\mathrm{e}\mathrm{c}}\text{,} $
其中$ {\lambda }_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}} $,$ {\lambda }_{\mathrm{c}} $,$ {\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{G}}} $,$ {\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{D}}} $,$ {\lambda }_{\mathrm{r}\mathrm{e}\mathrm{c}} $是超参数.
本实验采用的训练数据集是LibriTTS英文多说话人数据集[34],该数据集由24 kHz的2 456位说话人组成,总计持续时长为585 h. 我们的评估数据集源自VCTK[22],该数据集与LibriTTS一样同属于英文多说话人数据集. 它包含110位英语多说话人录制的48 kHz语音. 我们从中随机抽取100条语音,并将其降采样为24 kHz.
由于神经编解码器的压缩率、码本数量和其实现的比特率之间相互制约,确定其中2个参数后,剩余的参数也随之确定. 码本数量和压缩率均将影响模型的训练过程. 为了调节不同码本数目进行各比特率下的性能比较,主实验中将压缩率$ r $统一设置为320,梅尔谱图频域维度为80,hop_size设为160,对于压缩率的影响,后续亦将设计消融实验探究.
此外,编码器卷积单元下采样倍数设置为2. 解码器卷积单元上采样倍数逐个设置为8, 5, 4, 2. 以编码器唯一的下采样卷积单元为例,其网络参数如表1所示. Conv 1进行下采样,此外,Conv 2~3,Conv 4~5,…,Conv 18~19等相邻2层卷积分别组成残差连接结构. 训练时,批大小取16 , 学习率为 0.000 2. 超参数$ {\lambda }_{\mathrm{f}\mathrm{e}\mathrm{a}\mathrm{t}} $,$ {\lambda }_{\mathrm{c}} $,$ {\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{G}}} $,$ {\lambda }_{\mathrm{a}\mathrm{d}{\mathrm{v}}_{\mathrm{D}}} $,$ {\lambda }_{\mathrm{r}\mathrm{e}\mathrm{c}} $ 设置为1,10,1,1,45. 编码器输出特征512维. 单个码本的索引范围为[0, 1023]. 因此我们的模型和基线进行了比特率分别为0.75 Kbps,1.5 Kbps,3 Kbps,4.5 Kbps,6 Kbps的实验,分别对应1,2,4,6,8个数目的码本. 所有实验均在单卡2080Ti上训练了25个epoch,测试时统一采用CPU运行(Intel® Xeon® CPU E5-2640 v4 @ 2.40 GHz). 本实验采取Hifi-codec[20],Encodec[19],SoundStream[18]模型作为基线,其中Encodec和SoundStream 采用Yang等人[29]在Hifi-codec工作中重新实现的代码1. 除了上述神经编解码器,本实验也与当前广泛应用的Opus进行了比较.
本实验客观指标采用RTF衡量时延,其定义为输入语音的时长与编码所需时间的比率[11]. 因此,该指标与编码器的运行速度正相关. 评测中,无论是本文方法还是基线,编码时间均包含了从波形输入到生成编码索引的整个过程,这也包括了波形到梅尔谱图的转换时间.
本实验采用STOI[36]、VISQOL[37]以及WARP-Q [47]评价语音的感知质量. 其中WARP-Q专门被提出应用于神经语音编解码器的评测指标,本文采用其原始评分值进行比较. 本实验主观指标采用MOS评分,该数据由5位志愿者对100条随机语音从1~5打分完成评测.
本文的所有测试实验和消融实验都在VCTK数据集上进行. 采用训练之外的同类别数据集进行测试的实验模式不仅符合实际的应用场景,也对模型的泛化能力要求更高. 最终,各神经编解码器在VCTK上的结果如表2所示,可见本文方法在时间指标,即编码器端的RTF相对所有同类神经编解码器在所测的任意比特率下都具有优势. 以基线里平均感知质量较好的Hifi-codec作为基准,比特率为1.5 Kbps时编码实时率提升幅度最大,最多可提升4.6倍. 而比特率为3 Kbps时提升幅度较小,依然达到了3.15倍. 对于语音感知质量指标STOI,WARP-Q,VISQOL,在比特率较低(0.75 Kbps, 1.5 Kbps, 3.0 Kbps)时表现良好,能超过所有基线模型,但比特率较高时感知质量却有所不足. 以极低比特率0.75 Kbps时为例,相较最好的Encodec客观感知质量平均提升8.72%. 但比特率为6 Kbps时,相较Hifi-codec,STOI,VISQOL指标已经低于基线,3种客观指标平均降低了0.46%. 主观MOS的测定结果也进一步支持了该结论. 从残差矢量量化的原理分析,所用的码本越多,残差层数就越深,但是除去第1层接收的是编码器输出特征以外,后续接收的是前层量化的残差. 因此,由编码器计算的压缩激励权重所反映的通道之间的相关性和重要性对于后续残差层的效用有所削弱,甚至不再利于量化,后续的消融实验也进一步证实了这点,即对于重建语音感知质量,本文方法更有利于较低比特率,比特率较高时失效.
此外,表3展示了本文方法与目前实际场景中广泛使用的通用编解码方法Opus的对比结果. 由于Opus的最低支持带宽为6 Kbps,并且其帧大小为2.5 ms,5 ms,10 ms等值,因此该比较实验存在一定的局限. 所有编解码器均在6 Kbps的带宽条件下进行测试,神经编解码器采用8码本,压缩率取320倍. 结果显示,Opus在STOI指标上具有一些优势.
本实验中原始语音和重建语音的梅尔谱图对比如图3所示. 以6 Kbps和0.75 Kbps时为例,分别代表了高比特率与低比特率场景. 将图3(b)(c)、图3(d)(e)与图3(a)相比较,本文方法的重建语音梅尔谱图均与原始语音更为贴近,这意味着更相似的听感体验.
为了研究压缩率对所提出编解码器音质的影响,我们在比特率为3 Kbps和6 Kbps的条件下进行了消融实验. 实验中,较高的压缩率需要增加码本的数目. 具体的实验配置如表4所示,实验结果见表5. 在3 Kbps条件下,随着压缩率和码本数的增加,语音质量指标均有所提升. 然而,在6 Kbps条件下,尽管压缩率和码本数增加,音质和语音可懂度却出现下降. 这表明在更高的比特率中,扩大压缩率导致的码本数目增大并不总是对模型的训练和最终性能产生积极的影响. 这一结果强调了在不同比特率下,需要谨慎选择压缩率和码本数.
为了探究压缩激励方法的有效性和适用边界,我们将比特率设置为1.5 Kbps,3 Kbps,6 Kbps 三个档次进行了消融实验. 为了进一步对比,我们将输入特征分别设为梅尔谱图和波形进行消融探究,其中波形实验仿照Hifi-codec模型并恢复了多层卷积编码器结构. 最终结果分别如表6和表7所示,图4和图5分别对表6和表7中音频质量的各项指标进行了可视化,其中WARP-Q越低音频质量越高.
上述图表显示梅尔谱特征下,压缩激励权重(SE-weight)的方法在1.5 Kbps和3 Kbps的比特率下表现出一定的有效性,但在6 Kbps时效果开始减弱;而在波形特征下,该方法仅在1.5 Kbps的极低比特率下具有一定的效果,而在3 Kbps时其效果已显著降低. 同时,表4和表5均显示压缩激励方法会引发轻微的时延损耗,因此实际使用中应结合带宽环境进行取舍. 此外,我们可以认为,相较波形特征而言,梅尔谱特征作为输入(Mel-Input)与压缩激励权重方法更为适配,大多数情形下,这样组合在速度和质量上更具优势. 该现象也说明梅尔卷积编码器各个输出通道信息量之间的差异性会大于波形卷积编码器. 这是符合认知的,因为梅尔谱频率维度之间本身就具有显著差异,人语音的能量更多地集中在低频区域,且人耳对不同频带的感知不同,比如对低频感知更加明显,对高频信息感知较为模糊,这样的差异性将更大地开发低比特率时压缩激励权重方法的潜能.
本节旨在讨论激活函数对梅尔谱编解码器语音感知质量与时延的影响. 实验中将在梅尔谱编解码器中进行Relu激活函数与Snake激活函数的对比. Snake函数作为一种具有周期性特征的激活函数,能够有效地适应语音波形高周期性的性质. 在声码器和波形编解码器任务中,Snake激活函数已被证实能显著提升语音质量[38-39]. 消融实验结果如表8所示,对于梅尔谱编解码器,Relu激活函数能在语音质量与Snake激活函数相当的情况下,运行速度均具有明显提升. 尤其是比特率较低的情况下,以0.75 Kbps和1.5 Kbps为例,Relu激活函数和Snake激活函数在语音质量上几乎没有差距,RTF分别能提高2.16倍和2.09倍,因此Relu激活函数更适合应用场景实际需求.
本文采用了使用梅尔谱做为编码器输入特征,并采用低层卷积编码器,采用压缩激励的方法利用了梅尔谱图经过低层编码器输出特征各通道信息量的不均匀性. 本文在LibriTTS 和VCTK数据集上进行实验,结果表明,该方法在提升编码器端运行速度上具有优势,减少了时延. 此外,在较低比特率的场景中,重建的语音相比波形编解码器基线具有更好的感知质量. 通过消融实验,本文探究了压缩激励权重方法在不同带宽和不同输入特征中的适用情况,进一步确定了压缩激励权重更适合低比特率条件的结论. 此外,本文还对编解码器的激活函数进行了消融探究,采用的Relu激活函数相比周期性Snake激活函数在运行速度上更具突出优势. 在未来,可以根据对梅尔谱图的先验知识,设计更高效的、结合其特点的编解码系统.
作者贡献声明:周俊佐提出了算法思路、完成实验并撰写论文;易江燕提出指导意见并修改论文;陶建华参与实验;任勇实现了方法和实验设计;汪涛负责方法和实验设计,并修改论文.

参考文献 引证文献
排序方式:
1
De Andrade J F, De Campos M L R, Apolinario J A. Speech privacy for modern mobile communication systems[C]//Proc of the 33rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2008: 1777−1780
2
Haneche H, Ouahabi A, Boudraa B. Compressed sensing-speech coding scheme for mobile communications[J]. Circuits, Systems, and Signal Processing, 2021, 40(10): 5106−5126
3
Budagavi M, Gibson J D. Speech coding in mobile radio communications[J]. Proceedings of the IEEE, 1998, 86(7): 1402−1412
4
Bessette B, Salami R, Lefebvre R, et al. The adaptive multirate wideband speech codec (AMR-WB)[J]. IEEE Transactions on Speech and Audio Processing, 2002, 10(8): 620−636
5
Cox R V, Kroon P. Low bit-rate speech coders for multimedia communication[J]. IEEE Communications Magazine, 1996, 34(12): 34−41
6
Huang Yongfeng, Liu Chenghao, Tang Shanyu, et al. Steganography integration into a low-bit rate speech codec[J]. IEEE Transactions on Information Forensics and Security, 2012, 7(6): 1865−1875
7
Valin J M, Terriberry T B, Montgomery C, et al. A high-quality speech and audio codec with less than 10-ms delay[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2009, 18(1): 58−67
8
Hicsonmez S, Sencar H T, Avcibas I. Audio codec identification from coded and transcoded audios[J]. Digital Signal Processing, 2013, 23(5): 1720−1730
9
Dietz M, Multrus M, Eksler V, et al. Overview of the EVS codec architecture[C]//Proc of the 40th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2015: 5698−5702.
10
Valin J M, Vos K, Terriberry T. Definition of the opus audio codec[EB/OL]. 2012[2024-12-26]. https://datatracker.ietf.org/doc/html/rfc6716
11
Zeghidour N, Luebs A, Omran A, et al. SoundStream: An end-to-end neural audio codec[J]. IEEE Transactions on Audio, Speech, and Language Processing, 2021, 30: 495−507
12
Biswas A, Jia D. Audio codec enhancement with generative adversarial networks[C]//Proc of the 45th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2020: 356−360
13
Stimberg F, Narest A, Bazzica A, et al. WaveNetEQ—Packet loss concealment with waveRNN[C]//Proc of the 54th Asilomar Conf on Signals, Systems, and Computers. Piscataway, NJ: IEEE, 2020: 672−676
14
Xiao Wei, Liu Wenzhe, Wang Meng, et al. Multi-mode neural speech coding based on deep generative networks[C]//Proc of the 24th Annual Conf of the Int Speech Communication Association. Grenoble, France: ISCA, 2023: 819−823
15
Wu Y-C, Gebru I D, Marković D, et al. Audiodec: An open-source streaming high-fidelity neural audio codec[C]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023: 1−5
16
Jiang Xue, Peng Xiulian, Zhang Yuan, et al. Disentangled feature learning for real-time neural speech coding[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10094723
17
Petermann D, Jang I, Kim M. Native multi-band audio coding within hyper-autoencoded reconstruction propagation networks[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10094593
18
Lim H, Lee J, Kim B H, et al. End-to-end neural audio coding in the MDCT domain[C/OL]//Proc of the 48th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2023[2025-02-05]. https://ieeexplore.ieee.org/document/10096243
19
Kleijn W B, Lim F S C, Luebs A, et al. Wavenet based low rate speech coding[C]//Proc of the 43rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2018: 676−680
20
Jang Inseon, Yang Haici, Lim W, et al. Personalized neural speech codec[C]//Proc of the 49th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2024: 991−995
21
Du Zhihao, Zhang Shiliang, Hu Kai, et al. Funcodec: A fundamental, reproducible and integrable open-source toolkit for neural speech codec[C]//Proc of the 49th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2024: 591−595
22
Gârbacea C, Oord A, Li Yazhe, et al. Low bit-rate speech coding with VQ-VAE and a WaveNet decoder[C]//Proc of the 44th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2019: 735−739
23
Kleijn W B, Storus A, Chinen M, et al. Generative speech coding with predictive variance regularization[C]//Proc of the 45th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2021: 6478−6482
24
Oord A, Dieleman S, Zen Heiga, et al. WaveNet: A generative model for raw audio[J]. arXiv preprint, arXiv: 1609.03499, 2016
25
Kankanahalli S. End-to-end optimized speech coding with deep neural networks[C]//Proc of the 43rd IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2018: 2521−2525
26
Oord A, Vinyals O. Neural discrete representation learning[C/OL]//Proc of the 31st Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2017[2025-02-05]. https://dl.acm.org/doi/10.5555/3295222.3295378
27
Défossez A, Copet J, Synnaeve G, et al. High fidelity neural audio compression[J]. arXiv preprint, arXiv: 2210.13438, 2022
28
Ratnarajah A, Zhang Shixiong, Yu Dong. M3-AUDIODEC: Multi-channel multi-speaker multi-spatial audio codec[J]. arXiv preprint, arXiv: 2309.07416, 2023
29
Yang Dongchao, Liu Songxiang, Huang Rongjie, et al. Hifi-codec: Group-residual vector quantization for high fidelity audio codec[J]. arXiv preprint, arXiv: 2305.02765, 2023
30
O’shaughnessy D. Speech Communications: Human and Machine[M]. Piscataway, NJ: IEEE, 1999
31
Davis S, Mermelstein P. Comparison of parametric representations for monosyllabic word recognition in continuously spoken sentences[J]. IEEE Transactions on Acoustics, Speech, and Signal processing, 1980, 28(4): 357−366
32
Hasanabadi M R. MFCC-GAN codec: A new AI-based audio coding[J]. arXiv preprint, arXiv: 2310.14300, 2023
33
Hu Jie, Shen Li, Sun Gang. Squeeze-and-excitation networks[C]//Proc of the 31st IEEE Conf on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2018: 7132−7141
34
Zen Heiga, Dang V, Clark R, et al. LibriTTS: A corpus derived from librispeech for text-to-speech[J]. arXiv preprint, arXiv: 1904.02882, 2019
35
Liu Zhaoyu, Mak B. Cross-lingual multi-speaker text-to-speech synthesis for voice cloning without using parallel corpus for unseen speakers[J]. arXiv preprint, arXiv: 1911.11601, 2019
36
Taal C H, Hendriks R C, Heusdens R, et al. A short-time objective intelligibility measure for time-frequency weighted noisy speech[C]//Proc of the 35th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2010: 4214−4217
37
Hines A, Skoglund J, Kokaram A, et al. VISQOL v3: An open source production ready objective speech and audio mestric[J]. arXiv preprint, arXiv: 2004.09584, 2020
38
Kumar R, Seetharaman P, Luebs A, et al. High-fidelity audio compression with improved rvqgan[C/OL]//Proc of the 38th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2024[2025-02-05]. https://openreview.net/forum?id=qjnl1QUnFA
39
Lee S, Ping W, Ginsburg B, et al. BigVGAN: A universal neural vocoder with large-scale training[J]. arXiv preprint. arXiv: 2206.04658, 2022
40
Dietz M, Multrus M, Eksler V, et al. Overview of the EVS codec architecture[C]//Proc of the 40th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2015: 5698–5702
41
高玮玮, 单明陶, 宋楠, 等. 嵌入SENet的改进YOLOv4眼底图像微动脉瘤自动检测算法[J]. 生物医学工程学杂志, 2022, 39(4): 713−720
Gao Weiwei, Shan Mingtao, Song Nan, et al. Detection of microaneurysms in fundus images based on improved YOLOv4 with SENet embedded[J]. Journal of Biomedical Engineering, 2022, 39(4): 713−720 (in Chinese)
42
Chen Qiang, Liu Li, Han Rui, et al. Image identification method on high speed railway contact network based on YOLO v3 and SENet[C]//Proc of the 38th Chinese Control Conf (CCC). Piscataway, NJ: IEEE, 2019: 8772−8777
43
王成龙, 易江燕, 陶建华, 等. 基于全局-时频注意力网络的语音伪造检测[J]. 计算机研究与发展, 2021, 58(7): 1466−1475
Wang Chenglong, Yi Jiangyan, Tao Jianhua, et al. Global and temporal-frequency attention based network in audio deepfake detection[J]. Journal of Computer Research and Development, 2021, 58(7): 1466−1475 (in Chinese)
44
He Kaiming, Zhang Xiangyu, Ren Shaoqing, et al. Deep residual learning for image recognition[C]//Proc of the 29th IEEE Conf on Computer Vision and Pattern Recognition (CVPR). Piscataway, NJ: IEEE, 2016: 770−778
45
Kong J, Kim J, Bae J. Hifi-GAN: Generative adversarial networks for efficient and high fidelity speech synthesis[C]//Proc of the 34th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2020: 17022−17033
46
Goodfellow I, Pouget-Abadie J, Mirza M, et al. Generative adversarial nets[C/OL]//Proc of the 28th Annual Conf on Neural Information Processing Systems (NIPS). Cambridge, MA: MIT, 2014[2025-02-05]. https://www.researchgate.net/publication/263012109_Generative_Adversarial_Networks
47
Jassim W A, Skoglund J, Chinen M, et al. WARP-Q: Quality prediction for generative neural speech codecs[C]//Proc of the 46th IEEE Int Conf on Acoustics, Speech and Signal Processing (ICASSP). Piscataway, NJ: IEEE, 2021: 401−405
2025年第62卷第11期
115
53
0
引用本文
BibTeX
文章信息
doi: 10.7544/issn1000-1239.202440329
  • 接收时间:2024-05-21
  • 首发时间:2026-11-10
  • 出版时间:2025-11-01
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2024-05-21
  • 修回日期:2025-03-27
  • 录用日期:2025-04-04
基金
作者信息
    1中国科学院自动化研究所 北京 100190
    2清华大学自动化系 北京 100084
    3北京信息科学与技术国家研究中心(清华大学) 北京 100084

通讯作者:

参考文献
分享链接
https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202440329
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏