Speech to Text Support in OLPC

K. K. Subramaniam subbukk at gmail.com
Mon Sep 15 21:14:46 EDT 2008


On Monday 15 Sep 2008 5:22:34 am Mitch Bradley wrote:
> Having worked for several years for a company that was doing speech
> recognition, I'll be very interested to see how well it works on XO.  I
> must admit that I'll be rather surprised if it works well enough to be
> "usable" except under very limited conditions and low expectations.
Full speech recognition may be an overkill. XO is targeted at kids. Rather 
than Speech-to-Text, we could define the task as 'sounds to characters' and 
allow them to input text and control characters by sound. Control actions 
could be tied to specific sound signatures and invoked by sound input. 
Features of sound input (like pitch, tone, volume) could be extracted and 
made available to applications for mapping to specific actions within that 
program.

Sound input does have its limitations - will not work in noisy classroom 
environments, can add to noise in class and so on, but it would be a useful 
input method for children working alone or in small groups.

Subbu



More information about the Devel mailing list