Evaluating live streaming speech synthesis (TTS) requires assessing fine-grained, highly expressive prosody such as emotion, intonation, and energy which tra…
机构:阿里
来源:arXiv 2609.20124 | AI4Papers 论文推荐平台