Speech recognition

Speech recognition


I am currently trying the speech recognition feature of the Intel SDK beta out and have a few questions:

- When I run the sample app (voice_recognition), I get messages displayed which I guess are the possibilities, are they really it? And if yes, is there any way to access it in my code? It seems to be only debug message but I could use them. They look like this:

Words : 3

RECG: one <0> RECG: game <1> RECG: fame <3>

- The confidence value of the keywords recognized is always 0, is that normal?

- I can't find any doc for the class UtilCmdLine, is there any?

Thanks a lot,


4 posts / 0 new
Last post
For more complete information about compiler optimizations, see our Optimization Notice.

First two are known bugs of speech recognition module in Beta2 release. In next release the speech module will not print any debug messages and will report valid confidence through recognition handler.

There is no documentation specifically for UtilCmdLine class (except inline comments) because it used only as utility for sample applications. In function UtilCmdLine::Parse, application specifies list of supported options in parameter ‘options’ (all in single string) and passes ‘argc’ and ‘argv’ into two other parameters.

Hi Mikhail,

Thanks for your answer.

There's no way to access the narrowed set of possible recognized words? If no, is it gonna be implemented on the next release also?

About the UtilCmdLine I was just interested in the public members, but names are quite explicit.

Finally is there a date for the next release?


Speech recognition is reducing the need for manual transcription but speech recognition software is still not accurate enough to replace a human transcriptionist.

Transcription Services

Leave a Comment

Please sign in to add a comment. Not a member? Join today