{"paper":{"title":"SemaVoice: Semantic-Aware Continuous Autoregressive Speech Synthesis","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"SemaVoice adds a foundation-model alignment step to continuous speech representations so autoregressive TTS can keep semantic meaning without losing acoustic quality.","cross_cats":[],"primary_cat":"eess.AS","authors_text":"Haoning Xu, Hui Lu, Huimeng Wang, Jiajun Deng, Shiyin Kang, Shuhai Peng, Xueyuan Chen, Xunying Liu, Youjun Chen, Zhaoqing Li","submitted_at":"2026-05-16T12:37:06Z","abstract_excerpt":"Continuous autoregressive speech synthesis has recently emerged as a promising direction for zero-shot text-to-speech (TTS). However, existing methods still suffer from a fundamental mismatch between semantic-prosodic modeling and reconstruction-driven continuous speech representations. This mismatch causes TTS models to focus excessively on low-level acoustic textures at the expense of high-level semantic coherence, further exacerbating error accumulation in autoregressive generation. To address this challenge, we propose SemaVoice, a semantic-aware continuous autoregressive framework for hig"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"SemaVoice introduces a Speech Foundation Model (SFM) guided alignment mechanism that refines continuous speech representations to better capture both local semantic consistency and global structural relationships. These representations condition a patch-wise diffusion head within the autoregressive framework for high-quality speech synthesis, achieving an English WER of 1.71% on the Seed-TTS benchmark.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The core premise that an SFM-guided alignment step can resolve the fundamental mismatch between semantic-prosodic modeling and reconstruction-driven continuous representations without introducing new artifacts or error accumulation in autoregressive generation (stated in the abstract's problem formulation and solution description).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"SemaVoice adds SFM-guided alignment to refine continuous speech representations in autoregressive TTS, reporting 1.71% English WER on Seed-TTS and competitiveness with open-source SOTA.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"SemaVoice adds a foundation-model alignment step to continuous speech representations so autoregressive TTS can keep semantic meaning without losing acoustic quality.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"88a40c56d1a58a5e057ce7fe1927efded781077ffcc0a7504e0d766981bb5496"},"source":{"id":"2605.16964","kind":"arxiv","version":1},"verdict":{"id":"e388dd5e-62ed-4459-8fc7-8bc9aeb59b77","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T18:54:12.330602Z","strongest_claim":"SemaVoice introduces a Speech Foundation Model (SFM) guided alignment mechanism that refines continuous speech representations to better capture both local semantic consistency and global structural relationships. These representations condition a patch-wise diffusion head within the autoregressive framework for high-quality speech synthesis, achieving an English WER of 1.71% on the Seed-TTS benchmark.","one_line_summary":"SemaVoice adds SFM-guided alignment to refine continuous speech representations in autoregressive TTS, reporting 1.71% English WER on Seed-TTS and competitiveness with open-source SOTA.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The core premise that an SFM-guided alignment step can resolve the fundamental mismatch between semantic-prosodic modeling and reconstruction-driven continuous representations without introducing new artifacts or error accumulation in autoregressive generation (stated in the abstract's problem formulation and solution description).","pith_extraction_headline":"SemaVoice adds a foundation-model alignment step to continuous speech representations so autoregressive TTS can keep semantic meaning without losing acoustic quality."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.16964/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"cited_work_retraction","ran_at":"2026-05-19T19:51:57.835056Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"citation_quote_validity","ran_at":"2026-05-19T19:50:12.302521Z","status":"completed","version":"0.1.0","findings_count":0},{"name":"doi_title_agreement","ran_at":"2026-05-19T19:01:18.863763Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T19:01:00.392389Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T18:41:56.228084Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T18:33:26.313664Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"d5f1f760704a15549281e8728d93c868ff763ba147e67732562869a7c34bf194"},"references":{"count":56,"sample":[{"doi":"","year":2023,"title":"Audiolm: a language modeling approach to audio generation.IEEE/ACM transactions on audio, speech, and language processing, 31:2523–2533, 2023","work_id":"ec7df9ed-3b74-45ef-8d96-020ad7bda143","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2023,"title":"Speak, read and prompt: High-fidelity text-to-speech with minimal supervision.Transactions of the Association for Computational Linguistics, 11, 2023","work_id":"8389f213-b661-4ef1-bfb5-7a196c3ea657","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2024,"title":"CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens","work_id":"e5ad925a-4045-49b5-b301-208bcbf3eca8","ref_index":3,"cited_arxiv_id":"2407.05407","is_internal_anchor":true},{"doi":"","year":2025,"title":"Neural codec language models are zero-shot text to speech synthesizers.IEEE Transactions on Audio, Speech and Language Processing, 33: 705–718, 2025","work_id":"78b05e17-f13b-4cd0-80d7-758e20d40df2","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2025,"title":"Spark-TTS: An Efficient LLM-Based Text-to-Speech Model with Single-Stream Decoupled Speech Tokens","work_id":"31f99dad-40ae-4a19-aeff-eafa54f5b42a","ref_index":5,"cited_arxiv_id":"2503.01710","is_internal_anchor":true}],"resolved_work":56,"snapshot_sha256":"320cc6efc6c804d015f8c87d45f5fca277c12d5bd9a3e36d3477f8d90c0f0ccd","internal_anchors":8},"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"}