Klarinet Archive - Posting 001017.txt from 2001/02

From: "Tim Roberts" <timr@-----.com>
Subj: [kl] Re: ...off topic, speech recognition
Date: Mon, 26 Feb 2001 14:32:56 -0500

On Sat, 24 Feb 2001 14:48:41 -0800 (PST), William Wright wrote:
>
> This is off topic, but have you tuned into CNN and asked your TV to
>print the closed-captioning, which, I believe, is created in real time
>by speech recognition software?
> They don't have all the bugs out of it yet. <smile>

Actually, closed-captioning is NOT done with speech recognition. For live
programs, there are specially trained court reporters called "stenocaptioners"
who use a special keyboard to enter a phonetic-based code. They are typing the
sounds they hear, not the words. They can sustaion 250 words per minute;
faster in short bursts.

This then goes into a computer, which translates it into text. This
translation is not perfect.

Many of the glitches you see in closed captioning are caused by transmission
problems. The closed-caption signal has a very small bandwidth, so there is no
room for error correction. Transmission problems which would cause an
invisible glitch in the body of the picture can render an entire line of
captioning unreadable.

Note: I stole much of this from the web site of the Oregon Disabilities
Commission's Deaf and Hard of Hearing Access Program:
http://www.odc.state.or.us/tadoc/techcp18.htm
http://www.odc.state.or.us/tadoc/techcp13.htm

--
- Tim Roberts, timr@-----.com
Providenza & Boekelheide, Inc.

---------------------------------------------------------------------
Unsubscribe from Klarinet, e-mail: klarinet-unsubscribe@-----.org
Subscribe to the Digest: klarinet-digest-subscribe@-----.org
Additional commands: klarinet-help@-----.org
Other problems: klarinet-owner@-----.org