Sorry to respond with commercial information here in the newsgroup,
but the mail address didn't work.
We have a speech recognizer that runs on PDAs. It is quite small in
footprint (compared to our competitors) yet offers high accuracy,
speaker independent, continuous speech recognition.
Some background on TEMIC Speech Processing: We have been delivering
speaker independent, continuous speech recognition and processing
solutions to the automotive market since the mid nineties. In 1996 we
brought the world's first in-car speech recognition system to the
market. Since 1997 we are also active in the telecommunication market.
Capitalizing on our experience in robust speech recognition in adverse
conditions and under tight memory and CPU constraints, we have started
in 2001 to make speech recognizers for mobile units such as PDAs and
mobile phones. Our products support the most important European and
American languages and soon Japanese.
Please contact Stefanie Dolderer at stefanie...@temic-sp.com for
more information. She is our sales person for the mobile market.
Best regards,
Marcus Hennecke
--
Dr. Marcus E. Hennecke
TEMIC Sprachverarbeitung GmbH
Head Algorithmics Dept. E-A Tel +49-731-3994-117
Soeflinger Str. 100 Fax +49-731-3994-61-117
89077 Ulm - Germany marcus....@temic-sp.com
> What speech engines exist for PDAs?
Wondows CE, SVOX
But you have to ask, it's a very new project
and probably not on the website at this time.
Andy
> We have a speech recognizer that runs on PDAs. It is quite small in
> footprint (compared to our competitors) yet offers high accuracy,
> speaker independent, continuous speech recognition.
Are you talking about large vocabulary continous speech recognition
(several 10,000 words) or about command and control (10-50 words)?
Chris
"HansenAlan" <has...@wiredpocket.com> a écrit dans le message de news:
3C694827...@wiredpocket.com...
I am not talking about dictation. We do command and control, but
certainly more than 10-50 words. Typical command and control
applications have vocabularies on the order of several hundred. Some
applications (e.g. navigation) need several thousand words.
It does depend on the resources available. Some mobile devices (e.g.
cell phones) are very limited in resources. On the PC we can
demonstrate recognition of several 10,000 words and in the telecom
market we do have running applications of that size. This would
probably also run on an iPAQ, but to be honest we haven't had such
large applications on a PDA yet.
We will be demonstrating our technology at the CeBIT (March 13 to 20
in Hannover, Germany) and until then I would rather not disclose any
details on just exactly what applications we are working on. I hope
you understand. 8-)
Regards,
Marcus
Actually, this should read: http://www.svox.com/
Their speech synthesis is quite good, with very low resource
consumption. Their specialty is the multilingual synthesis: It is
truly multilingual, meaning that in the same sentence you can switch
languages and have it all speak in the same voice. Of course, they are
a Swiss company and in Switzerland you almost have to have a
multilingual synthesis.
Marcus
Here are some interesting PDA applications:
http://www.ectaco.com/software/item.php3?ssmodid=270&softid=261&scat=8
http://www.ectaco.com/software/item.php3?smodid=271&softid=270&scat=8
Best wishes,
James
Check www.voxtec.com
SW+HW solution, noise canceling mike, background noise mike, 70dB SNR Audio channel
Lawrence (Larry) Ricci
Applied Data System
lri...@applieddata.net