What type of study is this?

This is a Quantitative Study study.

October 8, 2025Open Access

Zero-Shot Streaming Text to Speech Synthesis with Transducer and Auto-Regressive Modeling

HSHaiyang SunChina University of Petroleum, East China SHShujie HuChinese University of Hong Kong SLShujie LiuHarbin University of Science and Technology

Key Points

SMLLE achieves high-quality speech synthesis with minimal latency, enhancing user interactions and experiences.
The system utilizes a transducer for real-time text processing and duration alignment, allowing for efficient streaming.
Experimental validation shows SMLLE surpasses existing methods while matching quality with traditional sentence-level TTS systems.
The Delete < Bos > Mechanism offers a novel solution to reduce delays in autoregressive speech generation.

Abstract

Zero-shot streaming text-to-speech is an important research topic in human-computer interaction. Existing methods primarily use a lookahead mechanism, relying on future text to achieve natural streaming speech synthesis, which introduces high processing latency. To address this issue, we propose SMLLE, a streaming framework for generating high-quality speech frame-by-frame. SMLLE employs a Transducer to convert text into semantic tokens in real time while simultaneously obtaining duration alignment information. The combined outputs are then fed into a fully autoregressive (AR) streaming model to reconstruct mel-spectrograms. To further stabilize the generation process, we design a Delete Mechanism that allows the AR model to access future text introducing as minimal delay as possible. Experimental results suggest that the SMLLE outperforms current streaming TTS methods and achieves comparable performance over sentence-level TTS systems. Samples are available on shy-98. github. io/SMLLEdemoₚage/.

KI fragen

Bookmark

View Full Paper