Agree. Only works with the same observer listening on two different systems. Even then I think the experiment is flawed because you could have a system that simply is more clear in the frequency range of human voice, which could be happen with boor bass response. for example on a set of electrostats with poor low frequency response this may be easier than on a true full range system.
Nonetheless, just for kicks, I'd be happy to try. If you set up a file transfer (you have my email), I'll give it a shot. Keep in mind I am not a native English speaker though!