[Athen] Best way to train the user that dictates one word at a time.

Doug Hayman dhayman at u.washington.edu
Wed Feb 27 10:27:43 PST 2013


I wonder if using something like the VOR function that some digital
recorders have would help.

Those are meant to save tape/storage space and record when they hear sound
above a certain level...then stop recording the silent gaps...and
re-activate again on the next sound.

And like noise-gates used in audio chains, getting a good setting can be a
bit tricky.

So if the user could dictate with VOR a

"string-----of-----words------dictated" separately, then play
them back all linked together as "string of words dicatated" to appear
linked more closely in time.

Just thinking along the lines of how one would use an audio/DAW
application to string together words/phrases once unwanted content is
edited out, so that it sounds like the speaker made no "ums", "errs" and
so on.



Doug Hayman <dhayman at u.washington.edu>
Technology Specialist
DO-IT Program (Disabilities, Opportunities, Internetworking, Technology)
UW Technology Services
Box 354842
Seattle, WA 98195
(206) 221-4165
http://www.washington.edu/doit

On Wed, 27 Feb 2013, Kenneth Elkind wrote:


>

> Has anybody worked with this issue?

>

> The research that I have been able to find suggests to bypass the general training and do training by using the accuracy center tools. The clinical research that was able to find this extremely old 2002 and 2004. That makes it Dragon NaturallySpeaking version 6 or possibly older.

>

> This one of the articles that I was able to find. Voice-Recognition and Aphasia: Can Computers Understand Aphasic Speech? Disability and Rehabilitation 2001

>

> Kenneth Elkind

> Assistive Technology Specialist

> (617) 287- 5243

> Kenneth.elkind at umb.edu<mailto:Kenneth.elkind at umb.edu>

> Skype User Number: adaptiveumb

>

>

> Adaptive Computer Lab

> Maximizing Learning Potential

>

> Learn about the Adaptive Computer Lab<http://www.umb.edu/it/getting_services/computer_labs/adaptive_computer_lab/>

>

>

>




More information about the athen-list mailing list