I am working on a pronunciation testing tool for an english academy. The test uses chrome web speech API, it compares the original writting with the transcript and gives a user a final score. It is very useful for students of english as a foreign language, and the same method could be used for other languages as well
I need two new improvements for my web app:
- Recognize the accent of the user. For this, I need to launch at the same time an instance of webkitSpeechRecognition for every english accent, and the compare all of them. The max score will tell me what accent do you have, but this is turning out difficult and I am not sure of it would be possible to make it work. Can I initiate several instances of webkitSpeechRecognition for the same speech?
- Also, Recording the speech will be necesary for students. Any ideas on how to record the speech?
Any ideas on this?
I am using SSL so it doesn't ask for permission every time.