[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Date Index][Thread Index][Author Index]

Re: Voice recognition

ok, i suppose i should elaborate a little.

1.  almost without exception, a few hundred dollars spent in training will
save many hours of frustration.  a little bit of help goes a long way
(having a trainer durring the initial "enrollment" is really important), 
the frustration (when it does not work) can be overwhelming.  this said,
using the current available best gear (1.5ghz pc, 512ram,
naturallyspeaking), most people can dictate 100 wpm with better than 95%
accuracy.  it is not the same as talking, and some attention must be paid 
what you are doing.

2.  the above is for dictating text....executing commands is much
easier/faster/more accurate (and they should never be one word commands).
the version of the software that has good command writing capabilities is
expensive (but nicely integrated).  ...the cheap version that is designed
for straight dictation (same core as the more expensive version) can be 
with a 3rd party macro program (like macro express).

3.  i say this as someone who has trained more than 100 users in all
situations (head admistrators of hospitals, doctors, lawyers, state 
disabled senior citizens, etc).  not everyone is able to regulate their
speech appropriatly, and not everyone can make it work (accents are not
necessarily a problem).  ..i'm not sure where it would be helpful for most
musicians (perhaps quick navigation between edit points or something).  if
anyone has an application where sr (speech recognition) would be an
enhancement, i'd love to hear about it.