Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Xing, Jingyuan, Yang, Mingru, Li, Zhipeng, Xing, Xiaofen, Xu, Xiangmin
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2510.11646
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918159652487168
author Xing, Jingyuan
Yang, Mingru
Li, Zhipeng
Xing, Xiaofen
Xu, Xiangmin
author_facet Xing, Jingyuan
Yang, Mingru
Li, Zhipeng
Xing, Xiaofen
Xu, Xiangmin
contents 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/.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11646
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle BridgeCode: A Dual Speech Representation Paradigm for Autoregressive Zero-Shot Text-to-Speech Synthesis
Xing, Jingyuan
Yang, Mingru
Li, Zhipeng
Xing, Xiaofen
Xu, Xiangmin
Sound
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/.
title BridgeCode: A Dual Speech Representation Paradigm for Autoregressive Zero-Shot Text-to-Speech Synthesis
topic Sound
url https://arxiv.org/abs/2510.11646