🚀 go-pugleaf

Keith D. Tyler (romulus@shore.net) mused:
: Arnold Hendriks (segmentation.fault@core.dumped.org) wrote:
: : speech-to-text conversion would be more interesting. But as long as "play
: : record 5" takes a few seconds to consider for even mainframes, I doubt
: : that will come in reach soon :)

: Speech-to-text conversion progams are fairly common, though they are
: probably pretty expensive and need to be calibrated to the individual
: user. But lots of people who have lost the use of their hands/fingers to
: type use speech-to-text converters to write books, term papers, etc.

: Then of course there's Stephen Hawking, who writes books with his
: handy eye-blink-to-text converter.

: Now here's a damn silly idea: Get a speech-to-text converter, go on ircm
: and have the person on the other end using a text-to-speech converter.

Not such a silly idea if you have plenty of cpu power at each end and want
to save bandwidth in between.  Lots of people have tried to compress actual
speech waveforms in real time and some of the results are pretty impressive
but to the best of my knowledge none come close to the efficiency of a speech
to text conversion.