Qwen-TTS now speaks dialects: new Alibaba speech synthesizer
Alibaba/Qwen
Alibaba Cloud has updated Qwen-TTS, a text-to-speech model that supports 3 Chinese dialects (Pekingese, Shanghainese, Sichuanese) and 7 bilingual Chinese-English voices. The model achieves human-level naturalness and expressiveness, available via Qwen API.
Alibaba Cloud announced the latest Qwen-TTS model (qwen-tts-latest or qwen-tts-2025-05-22) via Qwen API. Trained on millions of hours of speech, it automatically adjusts prosody, pacing, and emotional inflections. It supports 3 Chinese dialects — Pekingese, Shanghainese, and Sichuanese — with 7 Chinese-English bilingual voices including Cherry, Ethan, Chelsie, Serena, Dylan (Pekingese), Jada (Shanghainese), and Sunny (Sichuanese). Performance metrics on SeedTTS-Eval benchmark show low Word Error Rate and high speaker similarity. More languages and styles are planned. Usage is demonstrated via a Python code snippet using the DashScope API.
- Сокращения
- TTS = Text-to-Speech — синтез речи из текста
- API = Application Programming Interface — интерфейс программирования приложений
- WER = Word Error Rate — частота ошибок в словах
- SIM = Speaker Similarity — сходство голоса диктора
Source: Alibaba Qwen —
original