So when you ask it to speak, it would randomly switch voices as it
encountered 'corrected' words? That would be unnerving and annoying,
no?
I was talking about listen, not speech. (The subject should have been
VR)
Dictate a document and have VR record your voice. When it guesses
wrong and you correct the word, place a recording of your recorded
pronunciation in the data base. Learn your voice on the fly. People
use the same words over and over. It would be customed for your
style/voice.
Cell phones should use this method as well. My old dumb phone had
about 15 numbers. One of them was "Dan" and other was "Pam" More
often than not when I said "Dan" the phone would ask "Pam?"
Instead of trying to recognize everyone's name on the planet, the
phone should learn the names/words I use.
