FELLE：基于逐Token粗到细流匹配的自回归语音合成
FELLE: Autoregressive Speech Synthesis with Token-Wise Coarse-to-Fine Flow Matching
AI总结：
本研究提出自回归语音合成模型FELLE，融合语言建模与逐Token粗到细流匹配，通过调整先验分布、分层生成提升连续值Token建模的相干性与合成质量，实验验证其可显著改善TTS生成效果。
摘要：
为推进连续值Token建模与时间相干性增强，我们提出FELLE——一种融合语言建模与逐Token流匹配的自回归模型。借助语言模型的自回归特性与流匹配的生成效能，FELLE可有效预测连续值Token（梅尔频谱图）。针对每个连续值Token，FELLE通过引入上一步信息调整流匹配中的通用先验分布，提升相干性与稳定性。此外，为提升合成质量，FELLE引入粗到细的流匹配机制，以语言模型输出为条件分层生成连续值Token。实验结果表明，将流匹配技术融入自回归梅尔频谱图建模具备潜力，可显著提升TTS生成质量，详情见https://aka.ms/felle。
英文摘要：
To advance continuous-valued token modeling and temporal-coherence enforcement, we propose FELLE, an autoregressive model that integrates language modeling with token-wise flow matching. By leveraging the autoregressive nature of language models and the generative efficacy of flow matching, FELLE effectively predicts continuous-valued tokens (mel-spectrograms). For each continuous-valued token, FELLE modifies the general prior distribution in flow matching by incorporating information from the previous step, improving coherence and stability. Furthermore, to enhance synthesis quality, FELLE introduces a coarse-to-fine flow-matching mechanism, generating continuous-valued tokens hierarchically, conditioned on the language model's output. Experimental results demonstrate the potential of incorporating flow-matching techniques in autoregressive mel-spectrogram modeling, leading to significant improvements in TTS generation quality, as shown in https://aka.ms/felle.
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。
