Setting of pronunciation sensitivity/mispronunciation detection threshold in Azure Cognitive Services Speech service - pronunciation assessment

Question

As mentioned in the following site, it is stated that "Pronunciation Assessment provides accuracy scores at different levels" & "Consider setting a higher mispronunciation detection threshold". Where can I set the pronunciation sensitivity exactly and how does it work? I cannot find any information regarding the configuration in the documentation.

https://learn.microsoft.com/en-us/legal/cognitive-services/speech-service/pronunciation-assessment/characteristics-and-limitations-pronunciation-assessment

Answer 1

@Amanda The guidance in this case refers to using a threshold limit in the client application based on evaluations done by the customer on real world data and the responses from the speech service. This threshold can be set on client processing software after the response from the service is received. Usually, the accuracy score is reported on the scale of 0-100(HundredMark) or 0-5(FivePoint) based on the grading system set in the configuration parameter. This response can be evaluated against real world data to only use the assessment score if the response is above a certain threshold.

      "PronunciationAssessment":  
      {  
        "AccuracyScore" : 100.0,  
        "ErrorType" : "None"  
      }  

If an answer is helpful, please click on or upvote which might help other community members reading this thread.

Answer 2

thank you so much