Saved in:
Bibliographic Details
Main Authors: Torgashov, Nikita, Henter, Gustav Eje, Skantze, Gabriel
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.13518
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912965865766912
author Torgashov, Nikita
Henter, Gustav Eje
Skantze, Gabriel
author_facet Torgashov, Nikita
Henter, Gustav Eje
Skantze, Gabriel
contents Full-stream text-to-speech (TTS) for interactive systems must start speaking with minimal delay while remaining controllable as text arrives incrementally. We present VoXtream2, a zero-shot full-stream TTS model with dynamic speaking-rate control that can be updated mid-utterance on the fly. VoXtream2 combines a distribution matching mechanism over duration states with classifier-free guidance across conditioning signals to improve controllability and synthesis quality. Prompt-text masking enables textless audio prompting, removing the need for prompt transcription. Across standard zero-shot benchmarks and a dedicated speaking-rate test set, VoXtream2 achieves competitive objective and subjective results against public baselines despite a smaller model and less training data. In full-stream mode, it runs 4 times faster than real time with 74 ms first-packet latency on a consumer GPU.
format Preprint
id arxiv_https___arxiv_org_abs_2603_13518
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle VoXtream2: Full-stream TTS with dynamic speaking rate control
Torgashov, Nikita
Henter, Gustav Eje
Skantze, Gabriel
Audio and Speech Processing
Computation and Language
Human-Computer Interaction
Machine Learning
Sound
Full-stream text-to-speech (TTS) for interactive systems must start speaking with minimal delay while remaining controllable as text arrives incrementally. We present VoXtream2, a zero-shot full-stream TTS model with dynamic speaking-rate control that can be updated mid-utterance on the fly. VoXtream2 combines a distribution matching mechanism over duration states with classifier-free guidance across conditioning signals to improve controllability and synthesis quality. Prompt-text masking enables textless audio prompting, removing the need for prompt transcription. Across standard zero-shot benchmarks and a dedicated speaking-rate test set, VoXtream2 achieves competitive objective and subjective results against public baselines despite a smaller model and less training data. In full-stream mode, it runs 4 times faster than real time with 74 ms first-packet latency on a consumer GPU.
title VoXtream2: Full-stream TTS with dynamic speaking rate control
topic Audio and Speech Processing
Computation and Language
Human-Computer Interaction
Machine Learning
Sound
url https://arxiv.org/abs/2603.13518