{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2025:BX5C343I7NURFJEIBXOKHCJSUG","short_pith_number":"pith:BX5C343I","schema_version":"1.0","canonical_sha256":"0dfa2df368fb6912a4880ddca38932a1978203fef73aa3237b3dc2c4d64e925e","source":{"kind":"arxiv","id":"2502.18924","version":4},"attestation_state":"computed","paper":{"title":"MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Bai Jionghao, Boyang Zhang, Chen Zhang, Jialong Zuo, Rui Liu, Ruiqi Li, Shengpeng Ji, Xiang Yin, Xiaoda Yang, Yi Ren, Yu Zhang, Zhenhui Ye, Zhou Zhao, Ziyue Jiang","submitted_at":"2025-02-26T08:22:00Z","abstract_excerpt":"While recent zero-shot text-to-speech (TTS) models have significantly improved speech quality and expressiveness, mainstream systems still suffer from issues related to speech-text alignment modeling: 1) models without explicit speech-text alignment modeling exhibit less robustness, especially for hard sentences in practical applications; 2) predefined alignment-based models suffer from naturalness constraints of forced alignments. This paper introduces \\textit{MegaTTS 3}, a TTS system featuring an innovative sparse alignment algorithm that guides the latent diffusion transformer (DiT). Specif"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2502.18924","kind":"arxiv","version":4},"metadata":{"license":"http://creativecommons.org/publicdomain/zero/1.0/","primary_cat":"eess.AS","submitted_at":"2025-02-26T08:22:00Z","cross_cats_sorted":["cs.LG","cs.SD"],"title_canon_sha256":"3192267fa0c074cfe9aaff91425920956a4f95fb7665a314c14ab0122bba0f8d","abstract_canon_sha256":"e0ceb8af76d98d8c1ec255ff4d8083cfa547d1a05ba44f34da097200438c6e12"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T10:40:38.857308Z","signature_b64":"aVdDWCXeezsGEXw/OCuS9aCYHo6PDlM/AvWAoGBfhPP5AszHy3Gaxf2TyNqfe39RrPVAz5wQihVGcXnbeu9dAg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"0dfa2df368fb6912a4880ddca38932a1978203fef73aa3237b3dc2c4d64e925e","last_reissued_at":"2026-07-05T10:40:38.856808Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T10:40:38.856808Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesis","license":"http://creativecommons.org/publicdomain/zero/1.0/","headline":"","cross_cats":["cs.LG","cs.SD"],"primary_cat":"eess.AS","authors_text":"Bai Jionghao, Boyang Zhang, Chen Zhang, Jialong Zuo, Rui Liu, Ruiqi Li, Shengpeng Ji, Xiang Yin, Xiaoda Yang, Yi Ren, Yu Zhang, Zhenhui Ye, Zhou Zhao, Ziyue Jiang","submitted_at":"2025-02-26T08:22:00Z","abstract_excerpt":"While recent zero-shot text-to-speech (TTS) models have significantly improved speech quality and expressiveness, mainstream systems still suffer from issues related to speech-text alignment modeling: 1) models without explicit speech-text alignment modeling exhibit less robustness, especially for hard sentences in practical applications; 2) predefined alignment-based models suffer from naturalness constraints of forced alignments. This paper introduces \\textit{MegaTTS 3}, a TTS system featuring an innovative sparse alignment algorithm that guides the latent diffusion transformer (DiT). Specif"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2502.18924","kind":"arxiv","version":4},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2502.18924/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2502.18924","created_at":"2026-07-05T10:40:38.856867+00:00"},{"alias_kind":"arxiv_version","alias_value":"2502.18924v4","created_at":"2026-07-05T10:40:38.856867+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2502.18924","created_at":"2026-07-05T10:40:38.856867+00:00"},{"alias_kind":"pith_short_12","alias_value":"BX5C343I7NUR","created_at":"2026-07-05T10:40:38.856867+00:00"},{"alias_kind":"pith_short_16","alias_value":"BX5C343I7NURFJEI","created_at":"2026-07-05T10:40:38.856867+00:00"},{"alias_kind":"pith_short_8","alias_value":"BX5C343I","created_at":"2026-07-05T10:40:38.856867+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":16,"internal_anchor_count":0,"sample":[{"citing_arxiv_id":"2606.25672","citing_title":"Joint Residual Reweighting for Classifier Free Guidance in Flow-Matching Zero-Shot TTS","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2606.19325","citing_title":"Reference-Driven Multi-Speaker Audio Scene Generation from In-the-Wild Priors","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.18323","citing_title":"Reliable Neural-Codec Text-to-Speech by ASR Self-Verification and Distillation: Near-Zero Catastrophic Failures Across Models and Codecs","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.20650","citing_title":"EmoInstruct-TTS: Dual-Path Instruction-Guided Emotional Speech Synthesis","ref_index":10,"is_internal_anchor":false},{"citing_arxiv_id":"2606.07080","citing_title":"dots.tts Technical Report","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2606.06928","citing_title":"VoxCPM2 Technical Report","ref_index":17,"is_internal_anchor":false},{"citing_arxiv_id":"2606.03455","citing_title":"WavTTS: Towards High-Quality Zero-Shot TTS via Direct Raw Waveform Modeling","ref_index":38,"is_internal_anchor":false},{"citing_arxiv_id":"2605.30993","citing_title":"SwanVoice: Expressive Long-Form Zero-Shot Speech Synthesis for Both Monologue and Dialogue","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2606.25672","citing_title":"Joint Residual Reweighting for Classifier Free Guidance in Flow-Matching Zero-Shot TTS","ref_index":11,"is_internal_anchor":false},{"citing_arxiv_id":"2605.22083","citing_title":"RobustSpeechFlow: Learning Robust Text-to-Speech Trajectories via Augmentation-based Contrastive Flow Matching","ref_index":30,"is_internal_anchor":false},{"citing_arxiv_id":"2506.23552","citing_title":"JAM-Flow: Joint Audio-Motion Synthesis with Flow Matching","ref_index":19,"is_internal_anchor":false},{"citing_arxiv_id":"2507.01936","citing_title":"The Thin Line Between Comprehension and Persuasion in LLMs","ref_index":2,"is_internal_anchor":false},{"citing_arxiv_id":"2604.13067","citing_title":"From Seeing it to Experiencing it: Interactive Evaluation of Intersectional Voice Bias in Human-AI Speech Interaction","ref_index":14,"is_internal_anchor":false},{"citing_arxiv_id":"2604.00688","citing_title":"OmniVoice: Towards Omnilingual Zero-Shot Text-to-Speech with Diffusion Language Models","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.12292","citing_title":"CoSyncDiT: Cognitive Synchronous Diffusion Transformer for Movie Dubbing","ref_index":18,"is_internal_anchor":false},{"citing_arxiv_id":"2604.11283","citing_title":"Multimodal Large Language Model-Enabled Video Translation: A Role-Oriented Survey","ref_index":79,"is_internal_anchor":false}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG","json":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG.json","graph_json":"https://pith.science/api/pith-number/BX5C343I7NURFJEIBXOKHCJSUG/graph.json","events_json":"https://pith.science/api/pith-number/BX5C343I7NURFJEIBXOKHCJSUG/events.json","paper":"https://pith.science/paper/BX5C343I"},"agent_actions":{"view_html":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG","download_json":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG.json","view_paper":"https://pith.science/paper/BX5C343I","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2502.18924&json=true","fetch_graph":"https://pith.science/api/pith-number/BX5C343I7NURFJEIBXOKHCJSUG/graph.json","fetch_events":"https://pith.science/api/pith-number/BX5C343I7NURFJEIBXOKHCJSUG/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG/action/timestamp_anchor","attest_storage":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG/action/storage_attestation","attest_author":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG/action/author_attestation","sign_citation":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG/action/citation_signature","submit_replication":"https://pith.science/pith/BX5C343I7NURFJEIBXOKHCJSUG/action/replication_record"}},"created_at":"2026-07-05T10:40:38.856867+00:00","updated_at":"2026-07-05T10:40:38.856867+00:00"}