mlsanityData to 19 Oct 2026

BridgeCode: A Dual Speech Representation Paradigm for Autoregressive Zero-Shot Text-to-Speech Synthesis

arXiv 2510.11646 ↗First posted 13 Oct 2025cs.SDcited by 0

On the map: Speech and audio language models → Vocoders and speech synthesis

Abstract (arXiv)

Autoregressive (AR) frameworks have recently achieved remarkable progress in zero-shot text-to-speech (TTS) by leveraging discrete speech tokens and large language model techniques. Despite their success, existing AR-based zero-shot TTS systems face two critical limitations: (i) an inherent speed-quality trade-off, as sequential token generation either reduces frame rates at the cost of expressiveness or enriches tokens at the cost of efficiency, and (ii) a text-oriented supervision mismatch, as cross-entropy loss penalizes token errors uniformly without considering the fine-grained acoustic similarity among adjacent tokens. To address these challenges, we propose BridgeTTS, a novel AR-TTS framework built upon the dual speech representation paradigm BridgeCode. BridgeTTS reduces AR iterations by predicting sparse tokens while reconstructing rich continuous features for high-quality synthesis. Joint optimization of token-level and feature-level objectives further enhances naturalness and intelligibility. Experiments demonstrate that BridgeTTS achieves competitive quality and speaker similarity while significantly accelerating synthesis. Speech demos are available at https://test1562.github.io/demo/.

Key facts. No citing papers in the map yet (data through 19 Oct 2026). Placed in Vocoders and speech synthesis, one of 11 directions in Speech and audio language models. Builds on 23 of its 26 references that are in the map. First posted on arXiv 13 Oct 2025.

Builds on

23 of 26 references.

Closest ancestors · 12

Same direction, newest first

  1. 2025TaDiCodec: Text-aware Diffusion Speech Tokenizer for Speech Language Modeling22 Aug
  2. 2025Spark-TTS: An Efficient LLM-Based Text-to-Speech Model with Single-Stream Decoupled Speech Tokens3 Mar
  3. 2024CosyVoice 2: Scalable Streaming Speech Synthesis with Large Language Models13 Dec
  4. 2024WavTokenizer: an Efficient Acoustic Discrete Codec Tokenizer for Audio Language Modeling29 Aug
  5. 2024CosyVoice: A Scalable Multilingual Zero-shot Text-to-speech Synthesizer based on Supervised Semantic Tokens7 Jul
  6. 2024VALL-E 2: Neural Codec Language Models are Human Parity Zero-Shot Text to Speech Synthesizers8 Jun
  7. 2024XTTS: a Massively Multilingual Zero-Shot Text-to-Speech Model7 Jun
  8. 2023Transduce and Speak: Neural Transducer for Text-to-Speech with Semantic Token Prediction6 Nov
  9. 2023FunCodec: A Fundamental, Reproducible and Integrable Open-source Toolkit for Neural Speech Codec14 Sep
  10. 2023Better speech synthesis through scaling12 May
  11. 2023HiFi-Codec: Group-residual Vector quantization for High Fidelity Audio Codec4 May
  12. 2023Neural Codec Language Models are Zero-Shot Text to Speech Synthesizers5 Jan

Other references · 11

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,599
  2. 2025Qwen3 Technical Reportcited by 9,258
  3. 2024Qwen2.5 Technical Reportcited by 5,212
  4. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 4,385
  5. 2020HiFi-GAN: Generative Adversarial Networks for Efficient and High Fidelity Speech Synthesiscited by 1,462
  6. 2022High Fidelity Neural Audio Compressioncited by 917
  7. 2019LibriTTS: A Corpus Derived from LibriSpeech for Text-to-Speechcited by 864
  8. 2022UTMOS: UTokyo-SaruLab System for VoiceMOS Challenge 2022cited by 552
  9. 2020Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Imagescited by 269
  10. 2023UniAudio: An Audio Foundation Model Toward Universal Audio Generationcited by 162
  11. 2024Generative Expressive Conversational Speech Synthesiscited by 21

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026CuteTTS: Efficient and High-Quality Speech Synthesis via Autoregressive Modeling of Continuous Latents
  2. 2025CLEAR: Continuous Latent Autoregressive Modeling for High-quality and Low-latency Speech Synthesis
  3. 2025Pseudo-Autoregressive Neural Codec Language Models for Efficient Zero-Shot Text-to-Speech Synthesis
  4. 2025Comprehend and Talk: Text to Speech Synthesis via Dual Language Modeling
  5. 2025MiniMax-Speech: Intrinsic Zero-Shot Text-to-Speech with a Learnable Speaker Encoder
  6. 2026SemaVoice: Semantic-Aware Continuous Autoregressive Speech Synthesis
  7. 2025Towards Lightweight and Stable Zero-shot TTS with Self-distilled Representation Disentanglement
  8. 2024Speech Synthesis From Continuous Features Using Per-Token Latent Diffusion