Please Note:
It is our considered opinion that the use of inadequate hardware is the
greatest single reason why many health care and legal professionals consider
voice recognition software "Not ready for prime time." Proper
hardware configurations allow the software to deliver the results you want.
The following is a recent
email reprinted with the permission of its author, J. Craig Barnett, MD, F.A.C.C., of Orlando, Florida, in response to an inquiry on the medspeech mailing list for medical speech recognition about whether
a quiet computer was worth its additional cost.
"I would like to
add my voice to this with years of experience -- you cannot put
voice-recognition software on "Your mother's" computer. There
must be at least 10 times each week that I either personally run into a
colleague or by e-mail field questions from professionals who attempt to
purchase cheap voice-recognition software, place this on their computer and
become immediately frustrated with this. Of course, since the
professional is never wrong (our egos would never allow this) the
voice-recognition software and technology is always blamed as the culprit.
It is my strong opinion that this is the single reason why voice recognition
software is considered "Not there yet." Certainly, this is the
attitude of most health-care professionals, legal professionals and in the
insurance industry -- our fields of sale.
Readers are encouraged to look at the total system as has been suggested --
saving $100 or $200 on any part or motherboard will simply poison the
approach.
J. Craig Barnett MD, F.A.C.C.
Orlando, FL"
(If the use of voice
recognition in health care is of
interest to you, you can subscribe to
this mailing list by sending a blank email to: medspeech-subscribe@yahoogroups.com)
Another "rant"
on the subject was published in
Lockergnome whose website is:
http://www.lockergnome.com/
One Size Doesn't Fit All (Rant)
Scribbled by John Belanger
I steer people away - far away - from the mass-marketed, 'one size fits
all' computers in the stores (and advertised on TV - ed.) today. Everything is proprietary. Hardware; the
graphics adapter is a cheap chipset built onto the proprietary motherboard, as
is a sound chipset and a $3 modem. It's all hardwired, and it's hard to
upgrade.
(Note: Due to its length,
the entirety of this article can be found on our Hardware
Rants page.)
Desktop Computers
For consistently fast
and reliable performance, VoiceTranscription™ recommends a quiet
computer specifically designed to provide
an optimal environment for speech recognition. In addition to the
"glamor" components of high MHz CPUs, lots of RAM, and large (but
slow) hard drives used by major manufacturers who put more of their
money into their costly
advertising than your computer components, the VoiceTranscription™ computer puts those dollars into components
designed to transform the spoken word into the written word
faster and with better recognition accuracy.
The use of a properly designed computer saves you many HOURS every
week (not to mention the frustration of correcting misrecognized words). One source has calculated the use of
an optimized computer combined with a
Custom Language Model as equivalent to saving 30 to 40 hours of training time.
Multiplied by your professional rate, this alone would more than offset the cost of
your system.
VoiceTranscription™
has voice-recognition optimized computers built to its own specifications by a
shop with whom we've had a professional relationship for over 14 years.
Prices and configurations
vary monthly, so please Email
voice@dbic.net
for current details.
Notebook Computers
Until the introduction of the
Mobile Centrino™ processor, we had our best experiences with the Dell Latitude™
Pentium 4 series.
Since Dell currently does not
currently offer a Notebook with anything other than a Mobile Centrino™
processor which we have not found to give optimal performance and
speech recognition, we're currently recommending two other favorites, Sager
and Acer notebooks. Contact us or email
voice@dbic.net for current specifications and prices.
Digital
Recorders
Yes,
there finally are recorders "good enough" for accurate voice
recognition in the professional environment.
NEW!!!
Over the years, VoiceTranscription has been asked to test a variety of
digital recorders by their manufacturers but found most don't deliver acceptable speech
recognition quality for the professional environment. Two that do are the
Sony ICD-BM1 and Olympus DS4000 Models:
Olympus DS4000 Features:
VoiceTranscription™
