arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.05770eess.AS
收藏

零样本文本到语音合成与增强音频提示:2026 Wildspoof挑战赛TTS赛道的BSC提交

Zero-Shot TTS With Enhanced Audio Prompts: Bsc Submission For The 2026 Wildspoof Challenge TTS Track

Barcelona Supercomputing Center(巴塞罗那超级计算中心);Universitat de Barcelona(巴塞罗那大学);DFKI GmbH(德国人工智能研究中心)

Jose Giraldo, Alex Peiró-Lilja, Rodolfo Zevallos, Cristina España-Bonet

AI总结:

本文提出通过增强音频提示和多阶段噪声抑制提升零样本文本到语音合成的鲁棒性和语音质量。

摘要:

我们评估了两种非自回归架构,StyleTTS2和F5-TTS,以解决野外语音的自发性。我们的模型利用灵活的持续时间建模来提高语调自然度。为处理声音噪声,我们实现了一个多阶段增强流程,使用Sidon模型,其在信号质量上显著优于标准Demucs。实验结果表明,微调增强音频可获得更强的鲁棒性,达到最高4.21 UTMOS和3.47 DNSMOS。此外,我们分析了参考提示质量和长度对零样本合成性能的影响,证明了我们的方法在生成真实语音方面的有效性。

英文摘要:

We evaluate two non-autoregressive architectures, StyleTTS2 and F5-TTS, to address the spontaneous nature of in-the-wild speech. Our models utilize flexible duration modeling to improve prosodic naturalness. To handle acoustic noise, we implement a multi-stage enhancement pipeline using the Sidon model, which significantly outperforms standard Demucs in signal quality. Experimental results show that finetuning enhanced audios yields superior robustness, achieving up to 4.21 UTMOS and 3.47 DNSMOS. Furthermore, we analyze the impact of reference prompt quality and length on zero-shot synthesis performance, demonstrating the effectiveness of our approach for realistic speech generation.

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。

↑