Inworld shares TTS-2 expressive voice research
What happened
Inworld AI Chief Science Officer Igor Poletaev and research engineer Julia Turc discuss how the expressive voice model TTS-2 was built and evaluated. According to the conversation, TTS-2 moves from broad emotion labels to a new approach to voice direction control, and Igor explains how Inworld measures expressiveness. Julia Turc says voice benchmark scores do not always reflect performance in real conversations, and automated evaluation still misses issues the human ear can hear.
AI-written from the coverage · updated 22 d ago
Coverage timeline
Follow the coverage to see the story from different sides.
- Inworld AIInworld breaks down how TTS-2 expressive voice was built and evaluated
Inworld AI Chief Science Officer Igor Poletaev and research engineer Julia Turc discuss how the expressive voice model was built and evaluated. Julia notes that voice benchmark scores do not always reflect performance in real conversations, and automated evaluation still misses issues the human ear can hear. Igor explains how Inworld measures expressiveness and how TTS-2 moves from broad emotion labels to a new approach to voice direction control.