I think that it could represent a leap in the utility of the software if there was a feature that enabled conversion of any given audio file into its 'binaural' form. Given that binaural recordings may be subaudible (or sound like static) but still have sound-like affects on the brain, I hypothesize that subliminal audio recordings could be made which might have applications for promoting healing/well-being.
For instance, a user could record in his/her voice some healing affirmations. Then the file could be converted to binaural beats via Gnaural, which could generate the 'binaural' form of the recording. It might sound like static, but contain subliminal affirmations, etc. for the benefit of the user.
I have no idea whether this would actually work, but since any frequency can be reproduced in binaural form, I don't see why any sound file couldn't be converted into binaural form. Some studies on subliminal messages seem promising, and I think that this could be an amazing experiment.
An idea on how this could be accomplished technically; the recorded voice/audio could be first converted to the desired beat frequency using a fourier transform:
https://forums.oracle.com/forums/thread.jspa?threadID=1270970
http://www.dspdimension.com/admin/pitch-shifting-using-the-ft/
http://en.wikipedia.org/wiki/Discrete_Fourier_transform
this should map each increment of sound data to a different frequency range, establishing the required beat frequencies for each increment of data, then the appropriate base frequency could be found by knowing the beat frequencies to play to have the binaural effect.
I'd really like to see if true subliminal messages could be delivered this way.