Bub
So now we're arguing about A / B listening?
LOL!
Here's a thought ... maybe you can hit the DIFF button, then load up the Noise Removal plugin in Sound Forge, capture the difference, and remove it? I dunno.
I can think of a few dozen reasons why I'd want to hear the difference ...
More options are better?
No no no no no.
The reason it's not good to have the DIFF option is it expressly encourages people to do things that are REALLY REALLY STOOOOOOOOOPID.
As I said before, it COMPLETELY IGNORES the fact that the stuff that is taken out is taken out because of psychoacoustic masking. So it's something that sounds like a smart idea to people who don't see that when you are taking away the stuff that's doing the masking and trying to make decisions based on it, you can't hope to succeed. So it encourages people to screw things up. It's not like other places where maybe you can actually learn things from a DIFF signal - because in most cases any masking is more incidental, not THE VERY BASIS of what is included in the DIFF signal.
And the fact that they graphing the DIFF signal also encourages dumb behavior. You often find people who want to take DIFF signals and use them to compare codecs, thinking that a codec with a smaller DIFF must be "better". But again, THIS IGNORES ALL OF THE PSYCHOACOUSTICS which is the very basis of how lossy codecs work. A codec that can throw out MORE and still be transparent is the superior one. Seeing what or how much it throws out is irrelevant. It's only the results that matter.
The big problem is it's MUCH easier for people to screw around based on the DIFF signal, because it's all right there and easy to hear, whereas when actually listening to the encoded audio people may well hear no differences. This is especially true at higher bit rates, where people (even the ones who think they are "special") are generally unable to tell encoded audio from the original in double blind or ABX testing, despite what they may believe.