Let IT Begin

TED-TTS: Training-Free Intra-Utterance Emotion and Duration Control for Text-to-Speech SynthesisControllable Text-to-Speech는 여전히 fine-grained intra-utterance expression 측면에서 한계가 있음TED-TTSCausal masking과 monotonic stream alignment를 combine 한 segment-aware emotion conditioning를 도입추가적으로 local duration embedding steering과 global EOS logit modulation을 combine 한 segment-aware duration steering strateg..