Use Cases
I am using FUTO Voice Input for Chinese voice dictation on Android.
Issue:
When I select Chinese and use offline voice input, the recognized text is output in Traditional Chinese. I could not find an option to force Simplified Chinese output.
Expected behavior:
Please add a setting for Chinese output script, for example:
- Chinese Simplified / zh-Hans / zh-CN
- Chinese Traditional / zh-Hant / zh-TW / zh-HK
Actual behavior:
Chinese voice input currently outputs Traditional Chinese for me, which makes it difficult to use in Mainland Chinese / Simplified Chinese workflows.
Why this matters:
FUTO Voice Input is very easy to configure and works well offline, but without Simplified Chinese output it is hard to use as a daily Chinese voice input tool.
Device / environment:
- Android version:
- FUTO Voice Input version:
- Model used: Multilingual-74 / Multilingual-244 / other
- System language: Simplified Chinese, China
- Keyboard used: Yuyan IME / other
Suggestion:
If the ASR model cannot directly control script output, a post-processing option using zh-Hans normalization or OpenCC-style Traditional-to-Simplified conversion before inserting text would solve the problem.
Proposal
No response
References
No response
Use Cases
I am using FUTO Voice Input for Chinese voice dictation on Android.
Issue:
When I select Chinese and use offline voice input, the recognized text is output in Traditional Chinese. I could not find an option to force Simplified Chinese output.
Expected behavior:
Please add a setting for Chinese output script, for example:
Actual behavior:
Chinese voice input currently outputs Traditional Chinese for me, which makes it difficult to use in Mainland Chinese / Simplified Chinese workflows.
Why this matters:
FUTO Voice Input is very easy to configure and works well offline, but without Simplified Chinese output it is hard to use as a daily Chinese voice input tool.
Device / environment:
Suggestion:
If the ASR model cannot directly control script output, a post-processing option using zh-Hans normalization or OpenCC-style Traditional-to-Simplified conversion before inserting text would solve the problem.
Proposal
No response
References
No response