title is not quite accurate as this would have applications outside of coding
-- for example on tablet devices and mobile devices which either don't have
keys or have keys that are difficult to press.
I guess if you don't have RSI you will probably say ' what is the point? ' And
if you have RSI then you're not going to be able to help much ( as is my
situation ).
I think even if I didn't have RSI, if such software was available, I would
train and configure it and use it! Especially on tablet/mobile devices. It
would be great to speak into the phone and have it type a message and not even
need to check whether it got it right, as it is bound to be spot-on.
to optimise something like this, I guess would require hooking into sphinx at
a pretty low level -- before the trigram level ( if I'm using the right word
-- I mean the phoneme triples ), although -- that level may be useful as for
every pair of phonemes in a spoken word there is a probability of them
following one another. for example if we have had 'b' there is a strong
likelihood of 'a' but practically zero chance of 'k'
anyway -- I think something could be got up and running without too much work
just by creating a dictionary of allowed phoneme pairs
Ohmu
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
http://spascii.wikispaces.com/
what do people think of this project concept?
title is not quite accurate as this would have applications outside of coding
-- for example on tablet devices and mobile devices which either don't have
keys or have keys that are difficult to press.
I guess if you don't have RSI you will probably say ' what is the point? ' And
if you have RSI then you're not going to be able to help much ( as is my
situation ).
I think even if I didn't have RSI, if such software was available, I would
train and configure it and use it! Especially on tablet/mobile devices. It
would be great to speak into the phone and have it type a message and not even
need to check whether it got it right, as it is bound to be spot-on.
to optimise something like this, I guess would require hooking into sphinx at
a pretty low level -- before the trigram level ( if I'm using the right word
-- I mean the phoneme triples ), although -- that level may be useful as for
every pair of phonemes in a spoken word there is a probability of them
following one another. for example if we have had 'b' there is a strong
likelihood of 'a' but practically zero chance of 'k'
anyway -- I think something could be got up and running without too much work
just by creating a dictionary of allowed phoneme pairs
Ohmu