Tencent ships three speech models: ASR, TTS and real-time voice-preserving translation
solyarisoftware · x · 2026-11-02
Tencent released three speech models at once: Tenlai-ASR for multi-speaker meeting transcription, Tenlai-TTS for more natural speech generation, and Tenlai-LiveTranslate for simultaneous translation that preserves the original speaker's voice. The reposter notes that multi-party interruptions, latency, and naturalness remain the hardest problems for real-time voice agents, and is eager to see the actual latency and quality of the live translation.
More from Multimodal
- Testing AI models via game-ready 3D assets: mix Blender with Tripo and Meshy — chongdashu · 2026-11-02
- PiVo: a full TTS stack with vocoder in just 1.4M parameters, running locally on CPU — pmttyji · 2026-11-02
- Midnight Run: an underground racing short film made with Seedance 2.5 — Hot-Decision9790 · 2026-11-02
- Sonic 2 VFX animator shows hybrid Maya + AI rendering pipeline — KarimHann · 2026-11-02
- Fibo 2.0 image editing ships a 2-step distilled version on Hugging Face — Illustrious_Row_9971 · 2026-11-02
- 4B Qwen3.5 fine-tune generates game-ready 3D assets with one-click GLB export — linoy_tsaban · 2026-11-02