We are experiencing voice recognition issues on some of our keywords mis-triggering when the words are not said. Voice recognitions have even been hit in quiet areas which is quite concerning, but most likely was caused by a conversation happening nearby.
These are the keywords which are posing the most issues. We have around 30 others but these seem to be the only ones being triggered on accident.
Initially we just added another keyword to complement the voice command and that seemed to worked pretty good. However, this is not ideal since we are porting over our application from another device so we'd like it to be as close as possible when it comes to usability.
We have it set to the highest confidence level but it doesn't seem to be enough when it comes to single smaller words.
I have been debating on diving into the MRTK code which I think handles the information but it might just be a hardware sensitivity problem instead.
Any help on the matter would be greatly appreciated,