Re: voice recognition by emacs

[Tim X]

David R <angel_ov_north@tiscali.co.uk> writes:

> Hi,
> I am doing editing work on files which have thousands of normal words, like
> proof-reading journalism in LaTex, not C code or serious programming.
>
> What I want is a voice recognition system which will run on cygwin winXP.
> I have found lots of programming voice macros, the most used of which seem
> to be voice commander created of Hans van Dam.  This seems operate along
> the principle that I say "move cursor next paragraph" "perl-mode" and emacs
> responds to the commands, but i understand that great advancements have
> been made in recent years with speech synthesis and recognition software.
>
> It seems obvious to me that some academic researchers would have used
> emacs.  Is there no project on emacswiki or open source project like this?
>
> I understand that emacspeak is designed to describe screen content, not to
> type on a screen the words which are being said.  I can't compile emacspeak
> because I don't seem to have a c++ compiler as part of gcc on cygwin : and
> it is doubtful that doing so would assist me greatly anyway.
>
> Anyway, any pointers in a useful direction would be appreciated.
>
> David R

Emacspeak will be of no use to you for your objectives. 

There is nothing available specific to emacs that I'm aware of. I think
this is because the speech recognition stuff you are talking about needs to
occur at a different level of abstraction i.e. more OS/Envrionment level
than at an application level. 

I think you will hve problems doing what you want via cygwin. You would be
better off looking at solutions based on a specific OS. For example, there
is the sphinx 2 project, which is a speech recognition system for Linux
that I  believe runs under X windows. I'm not a Windows user, but believe
XP and Vista in particular, comes with some degree of speech recognition
capability. There are also commercial products, like Dragon and many
others. 

If you get one of these speech recognition packages, they will all be able
to provide a way of getting spoken input into an emacs buffer (or any other
editor/word processor for that matter. 

Tim


-- 
tcross (at) rapttech dot com dot au