Having trouble setting up a more custom implementation of this SDK where I want to manage transcription with the Open AI whispers for on-device transcription so I can also utilize that API for a wake word to activate voice commands.
https://github.com/Macoron/whisper.unity
However since there isn't an option for the microphone to not be recording while the script is active, there is a conflict in terms of getting access to the microphone. I tried workarounds for this where the Wit voice experience game object was enabled and disabled based on when the wake word was triggered. This would work for the first command, but then the activation would cease to function like it should when the 2nd voice command was issued.
I know this SDK offers Activate(String) as a function, so I want to just fully manage transcription via the whispers API and then send the string to WIT for processing after the wake word is triggered.
Version 66 of the Oculus voice SDK
Having trouble setting up a more custom implementation of this SDK where I want to manage transcription with the Open AI whispers for on-device transcription so I can also utilize that API for a wake word to activate voice commands.
https://github.com/Macoron/whisper.unity
However since there isn't an option for the microphone to not be recording while the script is active, there is a conflict in terms of getting access to the microphone. I tried workarounds for this where the Wit voice experience game object was enabled and disabled based on when the wake word was triggered. This would work for the first command, but then the activation would cease to function like it should when the 2nd voice command was issued.
I know this SDK offers Activate(String) as a function, so I want to just fully manage transcription via the whispers API and then send the string to WIT for processing after the wake word is triggered.
Version 66 of the Oculus voice SDK