Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning

Hugging Face Daily Papers Papers

Summary

Tacit-TTS 将 IndexTTS2 的自回归文本到语义解码替换为掩码非自回归生成,并结合免训练声学长度估计与 ReFlow 蒸馏,实现无需参考文本的零样本语音克隆,在超过 5 秒的语音上比 IndexTTS2 快 10 倍以上,同时支持跨语言及非语音(如婴儿咿呀学语、合成乱码)的参考音频克隆。

TTS systems with autoregressive semantic modeling have demonstrated strong zero-shot voice cloning performance and rich expressive variation, but their sequential decoding incurs substantial latency. Non-autoregressive alternatives offer much faster generation, yet often rely on more restrictive reference conditioning, such as requiring transcripts of the reference speech during inference. We present Tacit-TTS, an efficient transcript-free zero-shot voice cloning system distilled from IndexTTS2. Our model replaces autoregressive text-to-semantic decoding with masked non-autoregressive generation, introduces training-free acoustic length estimation, and accelerates the flow-matching renderer through ReFlow distillation. Across two English and two Mandarin datasets, Tacit-TTS achieves competitive zero-shot quality while generating speech over 10x faster than IndexTTS2 for utterances longer than 5 seconds. Its transcript-free conditioning further supports cross-lingual and non-lexical references. We validate this capability using references from eight other languages, infant babble, and synthetic gibberish, where transcript-dependent systems often degrade or fail due to unreliable ASR transcripts.
Original Article
View Cached Full Text

Cached at: 10/01/26, 08:24 PM

Paper page - Tacit-TTS: From Autoregressive Decoding to Masked Prediction for Efficient Transcript-Free Voice Cloning

Source: https://huggingface.co/papers/2609.38658

Abstract

TTSsystemswithautoregressivesemanticmodelinghavedemonstratedstrongzero-shotvoicecloningperformanceandrichexpressivevariation,buttheirsequentialdecodingincurssubstantiallatency.Non-autoregressivealternativesoffermuchfastergeneration,yetoftenrelyonmorerestrictivereferenceconditioning,suchasrequiringtranscriptsofthereferencespeechduringinference.WepresentTacit-TTS,anefficienttranscript-freezero-shotvoicecloningsystemdistilledfromIndexTTS2.Ourmodelreplacesautoregressivetext-to-semanticdecodingwithmaskednon-autoregressivegeneration,introducestraining-freeacousticlengthestimation,andacceleratestheflow-matchingrendererthroughReFlowdistillation.AcrosstwoEnglishandtwoMandarindatasets,Tacit-TTSachievescompetitivezero-shotqualitywhilegeneratingspeechover10xfasterthanIndexTTS2forutteranceslongerthan5seconds.Itstranscript-freeconditioningfurthersupportscross-lingualandnon-lexicalreferences.Wevalidatethiscapabilityusingreferencesfromeightotherlanguages,infantbabble,andsyntheticgibberish,wheretranscript-dependentsystemsoftendegradeorfailduetounreliableASRtranscripts.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2609\.38658

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2609.38658 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2609.38658 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2609.38658 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles