
【#Tech24H】NetEase Youdao has released the "Ziyue 4.0" TTS speech synthesis engine Confucius4-TTS, the industry's first open-source model that supports cross-lingual accent-free synthesis across 14 languages and enables voice cloning without requiring reference text. The model achieves internationally leading performance in key dimensions including cross-lingual voice cloning, reference-text-free modeling, emotional prosody transfer, and on-premises deployment, and is now fully open-sourced for global users. NetEase Youdao Confucius4-TTS represents a comprehensive breakthrough: users need only provide a 3-second audio sample for the model to complete voice cloning, with cloned voices achieving over 85% similarity to the original and cloning accuracy as high as 97%. Furthermore, Confucius4-TTS can automatically extract and analyze emotional features from reference audio. [ By Zhang Liyan ]
