BelchFire
10-19-2004, 10:36 AM
I'm hearing impaired, and I don't use the telephone well. I really enjoy closed captioning on the TV, and I was wondering if we could build a telephone to work the same way. That caused me to do some research.
1) Closed captioning is still done just the way you would expect. Someone is typing in all that stuff. On live telecasts, it's full of errors, but off-line, a stenographer can take all the time necessary to get it right.
2) Voice recognition software must be trained to recognize the voice of the user. You usually get a que card to read into the mic at the prompt, thus training the software to your particular dialect, and accent.
3) TTY is really the same thing as closed captioning; there's a stranger listening to your conversation, and typing everything they hear for you. That just doesn't appeal to me.
However....... I can hear to an extent, just not well enough to be considered normal. Likewise, the voice recognition software might "get" some of a stranger's voice without being trained. Between the two, it just might be enough.
Now to test this theory..... I need to get my phone conversations into the sound card so that I can try to see what the voice recognition software can do with a strange voice. How to do that?
Are there ready build devices that send telephone conversations into a mic jack? How difficult would it be to build?
1) Closed captioning is still done just the way you would expect. Someone is typing in all that stuff. On live telecasts, it's full of errors, but off-line, a stenographer can take all the time necessary to get it right.
2) Voice recognition software must be trained to recognize the voice of the user. You usually get a que card to read into the mic at the prompt, thus training the software to your particular dialect, and accent.
3) TTY is really the same thing as closed captioning; there's a stranger listening to your conversation, and typing everything they hear for you. That just doesn't appeal to me.
However....... I can hear to an extent, just not well enough to be considered normal. Likewise, the voice recognition software might "get" some of a stranger's voice without being trained. Between the two, it just might be enough.
Now to test this theory..... I need to get my phone conversations into the sound card so that I can try to see what the voice recognition software can do with a strange voice. How to do that?
Are there ready build devices that send telephone conversations into a mic jack? How difficult would it be to build?