Influence of Loss functions in Noise Reduction using SAGAN

Influence of Loss functions in Noise Reduction using SAGAN | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Influence of Loss functions in Noise Reduction using SAGAN VANI, Anusuya M A, Shraddha Shivakumar, Chayadevi ML This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8178096/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Recent research has demonstrated the feasibility of using generative adversarial networks (GANs) for speech enhancement. While GANs have been applied in this domain, their advantages—particularly in versatile speech enhancement and prominent noise reduction—have not been fully established. A key factor in GAN-based speech enhancement is the loss function, which plays a crucial role in balancing voice quality improvement against noise reduction in speech signals. This study explores the application of the SGAN model for noise reduction, evaluating the performance of various loss functions in conjunction with the discriminator model. To assess performance, the PSNR metric is applied across varying noise levels, ranging from (+/-) 0 to 15 dB, using the Aurora-2 and NOISEX-92 databases. The study presents findings on discriminator accuracy and noise reduction effectiveness, highlighting how different loss functions impact speech quality. Additionally, it examines PSNR performance across multiple types of noise, identifying pink and babel noise as the most significantly reduced, achieving notable accuracy. Various loss functions are compared to provide insights into their effect on speech quality, discriminator accuracy, and loss functions. Acoustics Speech enhancement generative adversarial networks SAGAN Loss functions Speech denoising babble noise pink noise Signal to Noise ratio Full Text Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8178096","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":549028028,"identity":"a21cdeb6-6fde-4e76-b6d4-900839bfb2fd","order_by":0,"name":"VANI","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0ElEQVRIiWNgGAWjYJCCwwwMNsxQtgQDAzNexXAtaSB1jA1Ea2EG64JoIQIYHD978HBh23l2+f7jzx/z7rBIbGDnPYBfy5m8hMMz224zG9zIMWzmPSOR2MDMl4BXi2RDjsFhXpAWCR7GZt42CWMGZh4D/Fr634C0nGMGOuwhcVr4JcC2HGBmOJBgCNIiR4QWoC0855LBfpk5F6iFjZAWNv4c4888ZXbJQIc9+PC2rY6Hn/8Mfi1gwMjGkIwwhLB6EPjDYEecwlEwCkbBKBiRAADrLz0Rf4rtOAAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0002-3735-3104","institution":"JSS Science and Technology University","correspondingAuthor":true,"prefix":"","firstName":"","middleName":"","lastName":"VANI","suffix":""},{"id":549028029,"identity":"763d7f9f-1202-4d98-a5db-e1dfb4feb0d0","order_by":1,"name":"Anusuya M A","email":"","orcid":"","institution":"JSS Science and Technology University","correspondingAuthor":false,"prefix":"","firstName":"Anusuya","middleName":"M","lastName":"A","suffix":""},{"id":549028336,"identity":"16508ced-239f-4d4a-83bc-2230da5094e7","order_by":2,"name":"Shraddha Shivakumar","email":"","orcid":"","institution":"Vidya Vardhaka College of Engineering","correspondingAuthor":false,"prefix":"","firstName":"Shraddha","middleName":"","lastName":"Shivakumar","suffix":""},{"id":549028337,"identity":"ea244915-0c4f-488c-93b5-1f015d4d401e","order_by":3,"name":"Chayadevi ML","email":"","orcid":"","institution":"BNMIT Bangalore","correspondingAuthor":false,"prefix":"","firstName":"Chayadevi","middleName":"","lastName":"ML","suffix":""}],"badges":[],"createdAt":"2025-11-22 06:01:40","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8178096/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8178096/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":96781983,"identity":"8b5c4a06-bf57-464d-a595-f2e815ef776c","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"docx","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":364699,"visible":true,"origin":"","legend":"","description":"","filename":"InfluenceofLossfunctionsinNoiseReductionusingSAGAN.docx","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/4474e0fb64d16e0ff39e6964.docx"},{"id":96781978,"identity":"af4190e8-9d6e-487e-a8fa-020d7336c3d0","added_by":"auto","created_at":"2025-11-26 04:37:53","extension":"json","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":342,"visible":true,"origin":"","legend":"","description":"","filename":"rs8178096.json","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/f799b9778b71b695695fcf38.json"},{"id":96781981,"identity":"388d6cdb-98e0-481f-8da2-7442a4c8998f","added_by":"auto","created_at":"2025-11-26 04:37:53","extension":"xml","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":81110,"visible":true,"origin":"","legend":"","description":"","filename":"rs81780960enriched.xml","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/b17e99e639f84d468c2719d6.xml"},{"id":96781979,"identity":"5d782163-274c-4332-906a-f94d8f9db4a5","added_by":"auto","created_at":"2025-11-26 04:37:53","extension":"jpeg","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6410,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/166ece969444552f2b178612.jpeg"},{"id":96781980,"identity":"f1aacacb-d704-4c16-92ce-37947c830d09","added_by":"auto","created_at":"2025-11-26 04:37:53","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":59476,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/de437816ed72a71d3c946094.png"},{"id":96914784,"identity":"9c43e59e-4f58-4238-bba1-ca8fdb296cd2","added_by":"auto","created_at":"2025-11-27 14:06:22","extension":"jpeg","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1966,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage2.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/a22f22402dab20c010faf7b0.jpeg"},{"id":96917108,"identity":"02b7ad82-0ac8-4e62-82a9-d129f9dec597","added_by":"auto","created_at":"2025-11-27 14:09:16","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":13745,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/100e8c9f9dbbd996612c9c59.png"},{"id":96781985,"identity":"fbbcef77-2c20-4f16-93f5-8aecec828e04","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"jpeg","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1412630,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage4.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/af5ead3f974c1762a606e296.jpeg"},{"id":96781993,"identity":"ae95f807-4922-4154-ba2c-1384161a6ba4","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"jpeg","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":243027,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage5.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/11924681931e12413c31cdd4.jpeg"},{"id":96781989,"identity":"773c8505-77f4-40dc-b4b3-b0ffe1b36429","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":9,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45889,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/b3c07206884d5f71e7c58a6f.png"},{"id":96781996,"identity":"1d625115-3709-447e-9638-952b2569393e","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":10,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":48222,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/152492ffa46b17689dad2c3e.png"},{"id":96915433,"identity":"54d8bc8d-d8ff-4d47-903e-66c6bf48b08e","added_by":"auto","created_at":"2025-11-27 14:07:14","extension":"png","order_by":11,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":39261,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/ac1bc610a3ba8198ccd9ca68.png"},{"id":96781998,"identity":"a8ead660-0d89-4af5-b09b-32f992505327","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":12,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":30839,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/a50888a2aa9a4cffc14f1e1b.png"},{"id":96782002,"identity":"613a23b5-eabf-44cf-b710-841bbaf42901","added_by":"auto","created_at":"2025-11-26 04:37:55","extension":"png","order_by":13,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":1024,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/e7ff751f1625849e904e0898.png"},{"id":96782000,"identity":"ac1196b7-647a-4c5d-a26e-1a3a770378da","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":14,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":20671,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage10.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/894d642fea70b454fe186efc.png"},{"id":96781995,"identity":"e0e6cd29-b9f3-48da-a62e-0959167586b6","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":15,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":999,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/608100179e88e00afee527d3.png"},{"id":96781986,"identity":"734ed12c-0867-43a7-b7f8-e81ee0461188","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":16,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5802,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/48ce1363b5132b506de2dc5f.png"},{"id":96915674,"identity":"89437cc6-bad0-4598-b101-dbaa1c149c0d","added_by":"auto","created_at":"2025-11-27 14:07:31","extension":"png","order_by":17,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":262580,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/0cb13e1b2a5c3b50283bf46c.png"},{"id":96781992,"identity":"d6b128c8-3314-4490-b770-4b4260e1573f","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":18,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":54766,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/c603875ae8ed9e32330f8783.png"},{"id":96781994,"identity":"c366858c-86f7-4ba3-93fb-044ac9e3da74","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":19,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":45449,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/9ec5f141e278bf786dcb26c1.png"},{"id":96781982,"identity":"d68abc8e-0e3d-448a-afff-ced7c5ffb2f7","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":20,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10800,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage7.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/50c2074a55e44d20cdec69e9.png"},{"id":96781987,"identity":"0b3ede0c-8b8c-4b7d-a7e5-10e8baa28630","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":21,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":10216,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage8.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/35fbbe18299a8aac814b56cf.png"},{"id":96782001,"identity":"cb244c08-6ab9-4bce-b52a-f30b49c839c4","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"png","order_by":22,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":11333,"visible":true,"origin":"","legend":"","description":"","filename":"Onlinefloatimage9.png","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/0494ccefe922d5d9365c2f4d.png"},{"id":96916224,"identity":"3cd692d4-50bf-47fa-a128-7725e96df481","added_by":"auto","created_at":"2025-11-27 14:08:13","extension":"xml","order_by":23,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":80037,"visible":true,"origin":"","legend":"","description":"","filename":"rs81780960structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/a67c7b5c5f82a8dbfd324eb7.xml"},{"id":96781991,"identity":"e87ea477-f4c4-4627-b5d0-630ea8a533b8","added_by":"auto","created_at":"2025-11-26 04:37:54","extension":"html","order_by":24,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":90530,"visible":true,"origin":"","legend":"","description":"","filename":"earlyproof.html","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1/02db795ad046cb5928e5bca0.html"},{"id":96922695,"identity":"37aa11c9-e334-4d8e-879c-34e391c3a26b","added_by":"auto","created_at":"2025-11-27 14:19:45","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":606940,"visible":true,"origin":"","legend":"","description":"","filename":"InfluenceofLossfunctionsinNoiseReductionusingSAGAN.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8178096/v1_covered_e3797532-0d60-4b22-ac67-093489043697.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eInfluence of Loss functions in Noise Reduction using SAGAN\u003c/strong\u003e\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Speech enhancement, generative adversarial networks, SAGAN, Loss functions, Speech denoising, babble noise, pink noise, Signal to Noise ratio","lastPublishedDoi":"10.21203/rs.3.rs-8178096/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8178096/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eRecent research has demonstrated the feasibility of using generative adversarial networks (GANs) for speech enhancement. While GANs have been applied in this domain, their advantages\u0026mdash;particularly in versatile speech enhancement and prominent noise reduction\u0026mdash;have not been fully established. A key factor in GAN-based speech enhancement is the loss function, which plays a crucial role in balancing voice quality improvement against noise reduction in speech signals. This study explores the application of the SGAN model for noise reduction, evaluating the performance of various loss functions in conjunction with the discriminator model. To assess performance, the PSNR metric is applied across varying noise levels, ranging from (+/-) 0 to 15 dB, using the Aurora-2 and NOISEX-92 databases. The study presents findings on discriminator accuracy and noise reduction effectiveness, highlighting how different loss functions impact speech quality. Additionally, it examines PSNR performance across multiple types of noise, identifying pink and babel noise as the most significantly reduced, achieving notable accuracy. Various loss functions are compared to provide insights into their effect on speech quality, discriminator accuracy, and loss functions.\u003c/p\u003e\u003cp\u003e\u003c/p\u003e\u003cp\u003e\u003c/p\u003e","manuscriptTitle":"Influence of Loss functions in Noise Reduction using SAGAN","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-11-26 04:37:49","doi":"10.21203/rs.3.rs-8178096/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"22c8c5fd-19e3-4733-892c-455757478c10","owner":[],"postedDate":"November 26th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":58424473,"name":"Acoustics"}],"tags":[],"updatedAt":"2025-11-26T04:37:49+00:00","versionOfRecord":[],"versionCreatedAt":"2025-11-26 04:37:49","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8178096","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8178096","identity":"rs-8178096","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}