[opus] Opus for ASR

Young, Milan Milan.Young at nuance.com
Mon Sep 17 11:26:30 PDT 2012


> Did you have any thoughts yet how you were going to give access to that?
> I assume Nuance doesn't want to run binaries from random people on the
> Internet :-)
[Milan] We're hoping that Opus developers represent a better subset of the internet population than a random sampling might discover :).  But yes, this is definitely a concern of ours and we're planning on building a properly sandboxed environment for the service.

 
> I suspect we'll want the "quick" test for automated regression testing and a
> longer test for any experiments specifically designed to optimize ASR accuracy.
> What kind of times are we talking about here (just the order of magnitude
> would help)?
[Milan] We'll need at least 10k words to perform a smoke test.  I suspect this will take on the order of about half an hour plus a bit more for administrative overhead.

Targeted testing is probably on the order of several days.  Due to resource scarcity, I'd like to schedule those slots with the particular subset of developers who are serious about ASR optimization.  If you are interested in participating, please let me know.

Thank you



More information about the opus mailing list