Open
Description
Self Checks
- This template is only for bug reports. For questions, please visit Discussions.
- I have thoroughly reviewed the project documentation (installation, training, inference) but couldn't find information to solve my problem. English 中文 日本語 Portuguese (Brazil)
- I have searched for existing issues, including closed ones. Search issues
- I confirm that I am using English to submit this report (我已阅读并同意 Language Policy).
- [FOR CHINESE USERS] 请务必使用英文提交 Issue，否则会被关闭。谢谢！:）
- Please do not modify this template and fill in all required fields.
Cloud or Self Hosted
Self Hosted (Docker)
Environment Details
Within the same reference_id, when processing identical text content in a single request cycle, the synthesized voice unexpectedly shifts between different vocal characteristics. For instance, a noticeable transition occurs from a female voice profile to a male voice profile mid-response.
Steps to Reproduce
send request with streaming
✔️ Expected Behavior
The generated voice should remain consistent
❌ Actual Behavior
The same text contains two sentences, but the voice gender switches abruptly—the first sentence uses a female voice, while the second suddenly shifts to a male voice.