Text To Speech Synthesis with Bidirectional Long Short Term Memory based Recurrent Neural Networks

Text To Speech Synthesis with Bidirectional Long Short Term Memory based Recurrent Neural Networks | 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 Text To Speech Synthesis with Bidirectional Long Short Term Memory based Recurrent Neural Networks S. Tharun Govind, M. Dinesh, K. Bindusri This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7079674/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 According to recent studies, feed-forward Deep neural networks (DNNs) perform better than text-to-speech (TTS) systems that use decision-tree clustered context-dependent hidden Markov models (HMMs) [ 1 , 4 ]. The feed-forward aspect of DNN-based models makes it difficult to incorporate the long-span contextual influence into spoken utterances. Another typical strategy in HMM-based TTS for establishing a continuous speech trajectory is using the dynamic characteristics to constrain the production of speech parameters [ 2 ]. Parametric time-to-speech synthesis is used in this study by capturing the co-occurrence or correlation data between any two points in a spoken phrase using time aware memory network cells. Based on our experiments, a combination of DNN and BLSTM-RNN is the best system to use. Upper hidden layers of this system use a bidirectional RNN structure of LSTM, the low layers use a simple, one way structure followed by additional layers. On objective and subjective metrics, it surpasses both the traditional decision-based tree HMM’s and the DNN-TTS system. Dynamic limitations are superfluous since the BLSTM-RNN TTS produces very smooth speech trajectories. Theoretical Computer Science Bidirectional Long Short-Term Memory(BLSTM) Deep_Neural_Network Recurrent_Neural_Network Statistical parametric voice synthesis hidden Markov model 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-7079674","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":482699735,"identity":"68ab9b3e-92b2-4249-bbec-c53597f2b911","order_by":0,"name":"S. Tharun Govind","email":"data:image/png;base64,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","orcid":"https://orcid.org/0009-0000-5404-8737","institution":"Sreenidhi Institute of Science and Technology","correspondingAuthor":true,"prefix":"","firstName":"S.","middleName":"Tharun","lastName":"Govind","suffix":""},{"id":482699736,"identity":"f150ccaf-bb76-46a3-8a47-86c09f714ac1","order_by":1,"name":"M. Dinesh","email":"","orcid":"","institution":"Sreenidhi Institute of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"M.","middleName":"","lastName":"Dinesh","suffix":""},{"id":482699737,"identity":"619bd335-13fa-477c-b7a6-1cd4ddfc766b","order_by":2,"name":"K. Bindusri","email":"","orcid":"","institution":"Sreenidhi Institute of Science and Technology","correspondingAuthor":false,"prefix":"","firstName":"K.","middleName":"","lastName":"Bindusri","suffix":""}],"badges":[],"createdAt":"2025-07-09 04:37:17","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-7079674/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7079674/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":86377448,"identity":"53fd0a53-fcc5-40e0-be24-c84c0ab9a3fe","added_by":"auto","created_at":"2025-07-10 03:19:29","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":600398,"visible":true,"origin":"","legend":"","description":"","filename":"SNP.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7079674/v1_covered_f3828506-dfc9-4e5f-9b7f-e72068785616.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eText To Speech Synthesis with Bidirectional Long Short Term Memory based Recurrent Neural Networks\u003c/p\u003e","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Sreenidhi Institute of Science and Technology","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":"Bidirectional Long Short-Term Memory(BLSTM), Deep_Neural_Network, Recurrent_Neural_Network, Statistical parametric voice synthesis, hidden Markov model","lastPublishedDoi":"10.21203/rs.3.rs-7079674/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7079674/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAccording to recent studies, feed-forward Deep neural networks (DNNs) perform better than text-to-speech (TTS) systems that use decision-tree clustered context-dependent hidden Markov models (HMMs) [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e, \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e]. The feed-forward aspect of DNN-based models makes it difficult to incorporate the long-span contextual influence into spoken utterances. Another typical strategy in HMM-based TTS for establishing a continuous speech trajectory is using the dynamic characteristics to constrain the production of speech parameters [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Parametric time-to-speech synthesis is used in this study by capturing the co-occurrence or correlation data between any two points in a spoken phrase using time aware memory network cells. Based on our experiments, a combination of DNN and BLSTM-RNN is the best system to use. Upper hidden layers of this system use a bidirectional RNN structure of LSTM, the low layers use a simple, one way structure followed by additional layers. On objective and subjective metrics, it surpasses both the traditional decision-based tree HMM\u0026rsquo;s and the DNN-TTS system. Dynamic limitations are superfluous since the BLSTM-RNN TTS produces very smooth speech trajectories.\u003c/p\u003e","manuscriptTitle":"Text To Speech Synthesis with Bidirectional Long Short Term Memory based Recurrent Neural Networks","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-07-10 02:47:24","doi":"10.21203/rs.3.rs-7079674/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":"163a05ca-39dd-4672-bcda-404f2284ff4a","owner":[],"postedDate":"July 10th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":51251757,"name":"Theoretical Computer Science"}],"tags":[],"updatedAt":"2025-07-10T02:47:24+00:00","versionOfRecord":[],"versionCreatedAt":"2025-07-10 02:47:24","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7079674","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7079674","identity":"rs-7079674","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}