반응형
[Paper 리뷰] StarVC: A Unified Auto-Regressive Framework for Joint Text and Speech Generation in Voice Conversion
StarVC: A Unified Auto-Regressive Framework for Joint Text and Speech Generation in Voice Conversion기존의 Voice Conversion model은 linguistic content의 explicit utilization을 neglect 함StarVCExplicit text modeling을 voice conversion에 integrateText token을 먼저 predict 한 다음 acoustic feature를 synthesize 하는 autoregressive framework를 활용논문 (INTERSPEECH 2025) : Paper Link1. IntroductionVoice Conversion (VC)는 ut..
Paper/Conversion
2025. 7. 3. 17:00
반응형