Closed AravindKaay closed 3 months ago
Hi @AravindKaay.
I first want to make sure there are no bugs in the system. Can you tell me how are you configuring the speechToText
property and what version of Deep Chat are you using.
Having read your problem description - I am 99.9% sure this is to do with the Web Speech API and not Deep Chat itself. By default, we use the native browser's speech API to facilitate the speech to text functionality - which has been facilitated via the Speech To Element library. The performance of the inference can really depend on your microphone quality, how close you are to the microphone, the browser you are using and your bandwidth connection. Hence, I am leaning that this a problem in that area. You can test out the inference performance in the following website to see if you have the same result.
The only other way that the inference quality can be potentially improved is by using the Azure service.
Let me know if you face any other issues with this. Thanks!
Thanks for your reply, I have tested the speech input on that website. I get the same kind of response. You are correct the issue is not related to deepChat.
Thanks, I will be closing this issue as the functionality is not handled by Deep Chat.
When I say "hi" it is not taken as input when I say something along with Hi then only it takes the word. For example, If I say "Hi" then nothing displays on the screen so afterward if I say "Hello" then it displays as "Hi Hello". The microphone neglects hi when it is says alone. Then when we give a time gap after we spoke and again to make it active we need to say "Hi" to activate the speech recognition.