VoXtream：极低延迟的全流文本到语音
VoXtream: Full-Stream Text-to-Speech with Extremely Low Latency
AI总结：
VoXtream通过高效架构实现极低延迟的全流式文本到语音系统，其在多个指标上超越大基线，且在GPU上达到102毫秒的初始延迟。
摘要：
我们提出了VoXtream，这是一种完全自回归、零样本流式文本到语音（TTS）系统，适用于实时使用，能够从第一个词开始说话。VoXtream通过单调对齐方案和有限的前瞻机制，直接将输入的音素映射到音频标记。该系统围绕增量音素Transformer、时间Transformer预测语义和持续时间标记，以及深度Transformer生成声学标记构建。据我们所知，VoXtream在公开可用的流式TTS中实现了最低的初始延迟：在GPU上为102毫秒。尽管是在中等规模的9000小时语料库上训练的，它在多个指标上与更大的基线相匹配或超越，同时在输出流和完整流式设置中提供有竞争力的质量。演示和代码可在https://herimor.github.io/voxtream上获得。
英文摘要：
We present VoXtream, a fully autoregressive, zero-shot streaming text-to-speech (TTS) system for real-time use that begins speaking from the first word. VoXtream directly maps incoming phonemes to audio tokens using a monotonic alignment scheme and a limited look-ahead that does not delay onset. Built around an incremental phoneme transformer, a temporal transformer predicting semantic and duration tokens, and a depth transformer producing acoustic tokens, VoXtream achieves, to our knowledge, the lowest initial delay among publicly available streaming TTS: 102 ms on GPU. Despite being trained on a mid-scale 9k-hour corpus, it matches or surpasses larger baselines on several metrics, while delivering competitive quality in both output- and full-streaming settings. Demo and code are available at https://herimor.github.io/voxtream.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
