cornelyus
2010-01-07 13:54:01 UTC
Meaning..
I have built the grammar i want to use, but when speaking it catches things
i say and understands it like the commands i want to recognize. So i thought,
if i have some words on my grammar, sound-alike the ones i already have, the
SRE will recognize them and i'll treat it as garbage. Does this makes sense?
Because since i started reading about speech recognition , one of the rules
to me was, smaller grammars, better accuracy, and this seems to go totally
against this.
Anyone to share some tips, experiences is more than welcome!
I have built the grammar i want to use, but when speaking it catches things
i say and understands it like the commands i want to recognize. So i thought,
if i have some words on my grammar, sound-alike the ones i already have, the
SRE will recognize them and i'll treat it as garbage. Does this makes sense?
Because since i started reading about speech recognition , one of the rules
to me was, smaller grammars, better accuracy, and this seems to go totally
against this.
Anyone to share some tips, experiences is more than welcome!