Pronunciation assessment

@Yagit

Thanks for reaching out to us, if you are willing to develop an app to do so, I would suggest you try Pronunciation assessment feature which is just for this scenario. And you don't need to develop an app.

Fortunately, you can do it directly in Speech Studio if you have the recording.

Reading and speaking scenarios

For pronunciation assessment, there are two scenarios: Reading and Speaking.

• Reading: This scenario is designed for scripted assessment. It requires the learner to read a given text. The reference text is provided in advance.
• Speaking: This scenario is designed for unscripted assessment. It requires the learner to speak on a given topic. The reference text is not provided in advance.

Please take a look on above information and let me know if that is your case. If you want to develop your, you may want to consider custom speech feature which you can leverage it to do the scenario, you will know how to evaluate pronunciation with speech to text through the Speech SDK. To get pronunciation assessment results, you apply the PronunciationAssessmentConfig settings to a SpeechRecognizer object according to below document -

https://learn.microsoft.com/en-us/azure/ai-services/speech-service/how-to-pronunciation-assessment?pivots=programming-language-csharp

I hope this helps.

Regards,

Yutong