>
> Voice recognition is NOT a silly notion for the SAM.
Yes it is.
> I was not expecting to use it for dictation, afterall.
> There were even some little robot toys several years
> ago that could distinguish enough to accept about 6
> commands. You might also be impressed by some of the
> PC voice rec software. If you use a limited vocabulary,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
This is the problem, plus the way people's accents vary. Speech from someone
from, say, Newcastle will be more than a little bit different from an
American accent.
> then the software has only to make the best match and
> could even intelligently guess from the phrase list.
>
>
OK, so you've got 512k on your SAM. To get a decent sample you need 44kHz
(this will sample frequencies up to 22kHz, and your average person's voice
will be well within this range) and 8 bits per sample, I guess.
So, that's 44k a second.
So, you'll fill up your available memory in roughly 10 seconds.
And you've got to _process_ that data, not just sample it.
No way.
Nige
