Cowling, Mark wrote:
> I would like to introduce an extra question: because my wife is deaf
> we watch live events on the television with subtitles. [snip] I would
> be interested to know how this is done? If the solution is a purely
> technological one, is it one that we could imitate in some way?
As I understand it, there are two main ways that live subtitling is done
1) Palantypy - where an individual using a chorded keyboard with
phonetic system of some kind types what is said by speakers at a very
high speed. I know the ones for deaf people's speech to text
transliteration are expected to do 180 wpm at 90% accuracy
(Signature/CACDP course spec).
2) Speech recognition, where an individual who is watching the live
broadcast repeats what the speakers are saying into a speech recognition
programme such as Dragon. I assume these are professionally trained
individuals who happen to have good speech pickup from Dragon et al. I
think there may be phonelines involved somewhere so the text is uploaded
straight into the subtitle stream.
I'm partially deaf myself - I almost never watch live subtitled content
as the lag between the sound, lip-patterns and the subtitles gives me a
headache. I have similar difficulties with text to speech
transliteration systems, I'd much prefer BSL/SSE to supplement my
residual hearing for live speakers despite my sign not being that good.
Interestingly I think my deafness and my resultant mild (most people
don't notice) speech and language impairment is what has made pre v10
Dragon unusable for me. Mark, can your wife use Dragon? Has she ever
I need to play with my copy of v10 when it's not nearly end of term and
see if I can arrange training from someone experienced with working with
deaf users (if such an obscure thing exists). In fact if any of you lot
know of a suitable Dragon trainer who has worked with deaf or speech
impaired individuals I'd be very interested to hear about them!
I lend copies of dragon to disabled students via our loan pool and I
find their success and happiness with it varies, some get on well with
it, others really don't. I always advise them to put the learning time
in, but recognise that it may not be a solution which works for them or
that they may need proper training which I can't provide.
I wonder if there's a correlation between 'verbal fluency' and success
with speech to text system....