dictation

[Eric S. Johansson]

Angelo Marra wrote:
> Hi folks
> How hobout bulding a dictation Sofware for ubuntu?
> WE need it!
> 
> QUOTE from the net:
> 
>     * TheMuso's blog <http://www.themuso.com/blogs/TheMuso>
> 
> Voce Dictation 
> <http://www.themuso.com/blogs/TheMusoUbuntu_Accessibility_Part_4_future_Ubuntu_Accessibility#comment-13>
> Submitted by NathanDBB (not verified) on Wed, 11/28/2007 - 05:05.
> 
> Many people who use Linux are spending 10+ hours in front of their 
> computers each day. Because of this, many of us have typing-related hand 
> damage. While I don't expect to be able to use the command line with 
> voice, I should be able to write email and use a mouse-grid.
> 
> Ubuntu _needs_ voice dictation software. We have not even started down 
> the LONG ROAD to having free voice software. We have a couple of 
> speech-to-text engines, but we don't have a good large vocabulary to 
> work with. This Wiki page talks about getting the collection of this 
> vocabulary started, but NOTHING has happened.
> 
> https://wiki.ubuntu.com/SpeechRecognition
> 
> As you point out, we need voice tools if we are going to get into 
> mainstream corporate and government desktops. If we are are going to be 
> an alternative to MS Vista in these markets, speech-to-text is MORE 
> IMPORTANT then 3d desktops.
> 
> I  totally agree!
> What do tou think?

as a long-term user of speech recognition, I too long for speech recognition on
Linux.  A friend of mine, Susan Cragin, is the lead QA person for naturally
speaking on wine.  She has reported that if you use NaturallySpeaking 10 plus
the latest bleeding edge wine, on the Jaunty Alpha's, NaturallySpeaking works
for about 13 minutes before croaking.

However, dictation on the wine environment alone is not useful.  We need a
gateway to communicate from NaturallySpeaking in wine to Linux applications.
I've proposed using an enhancement of the NaturallySpeaking dictation box as a
preferred user interface with dealing with remote applications.

Over the past 15 years, I've seen almost no progress in speech enabling
applications.  It's hard work, it changes every time the application changes
and, quite frankly it's almost completely a dead loss in terms of effort.
Getting an OS to use a common code base for  automatic speech recognition is
almost as completely fruitless as tweaking applications.  Why?  Because if it
was possible or valued, it would have been done by now.  So, the only way to
handle this is to use a different user interface model and  one possibility is
the dictation box cut and paste model.


---eric