Fallback language without country code in speech recognise

Question

In some cases, languages do not have country codes, such as zh-CN can be written as simply "zh" for Chinese, or en-US can be written as "en" for English.

But speech recognise config does not support this, when I use zh for the language, the text recognised is empty.

I hope that Microsoft can update the speech recognition configuration to support languages without country codes, allowing for accurate text recognition.

Answer 1

Hi @CodeKidz ,

Thank you for reaching out to Microsoft Q&A forum!

I understand that you are facing issues with speech recognition configuration when using languages without country codes. Currently, the Speech Recognition API requires a language-locale format string to recognize speech.

We appreciate your suggestion to update the speech recognition configuration to support languages without country codes. I encourage you to submit a feature request through the Azure Feedback Forum. Here's the link to the Azure Feedback Forum: Post idea · Community (azure.com)

I hope you understand. Thank you.

---

Do click Accept Answer and Yes for was this answer helpful. And, if you have any further query do let us know.

Answer 2

Hello

but according to me

the country code is important to distinguish different dialects or versions of a language like french

there are differences between

fr-FR and fr-CA (Canadian french and ordinary French)

same is valid for chinese

https://learn.microsoft.com/en-us/azure/ai-services/speech-service/language-support?tabs=stt