When first time using startContinuousRecognitionAsync Safari prompts correctly for microphone access, I have a 'microphone active' display next to the url, and speech gets recognized.
As I am integrated GPT3 I am using stopContinuousRecognitionAsync to stop it until GPT 3 respond and after that I start again using startContinuousRecognitionAsync to listen the user but at that time I do not get a microphone access prompt, and there is no 'microphone active' display next to the URL.
We are using SpeechSDK.TranslationRecognizer's recognized method for the capturing the user input but it's never go into that function second time as microphone is not active.
Expected behavior
Once I stop using stopContinuousRecognitionAsync and then I again start it using startContinuousRecognitionAsync it should show the microphone active again and go to recognized method again.
Describe the bug
When first time using startContinuousRecognitionAsync Safari prompts correctly for microphone access, I have a 'microphone active' display next to the url, and speech gets recognized.
As I am integrated GPT3 I am using stopContinuousRecognitionAsync to stop it until GPT 3 respond and after that I start again using startContinuousRecognitionAsync to listen the user but at that time I do not get a microphone access prompt, and there is no 'microphone active' display next to the URL.
We are using SpeechSDK.TranslationRecognizer's recognized method for the capturing the user input but it's never go into that function second time as microphone is not active.
Expected behavior Once I stop using stopContinuousRecognitionAsync and then I again start it using startContinuousRecognitionAsync it should show the microphone active again and go to recognized method again.
Version of the Cognitive Services Speech SDK I am using https://aka.ms/csspeech/jsbrowserpackageraw to load the latest SDK.
Platform, Operating System, and Programming Language