Qwen-Audio-3.0-TTS Flash Multi-Voice Style Demo

This piece is aimed at developers who need to quickly validate TTS performance; it uses a single voice to demonstrate the naturalness and instruction-following ability of qwen-audio-3.0-tts-flash in Chinese spoken-word scenarios. In terms of composition, the piece uses only one built-in voice, longanhuan_v3.6, publicly available on the author's line. The text is a short product-description-style narration, and the instruction field is used to specify tone and pacing in natural language (steady, with a slight broadcast feel). rate and pitch remain at the default 1, and the output is a 24kHz wav for direct listening and further processing. seed is fixed to ensure reproducibility. Applicable scenarios: audible evaluation of voice and style before integrating the DashScope Speech Synthesizer API; as a baseline sample for real-time scenarios such as order announcements, alert notifications, and assistant voice output; as reference material when comparing different voices within the same family or different rate/pitch combinations. Note: the schema currently offers a limited set of selectable voices for this endpoint; this piece does not involve voice cloning or text-based voice design; when professional-grade audio quality is required, the Plus version should be weighed.