Let me clarify my approach, with the disclaimer that I'm new at engineering and could be very wrong:
My understanding is that you can connect the audio source's ground to the +2.5v level obtained from the voltage divider. Then, the AC swings relative to that virtual ground. Of course, you still need to capacitively couple the audio signal to prevent any problems with DC offsets. I've seen a schematic actually decouple both the audio ground and audio signal inputs, but there was also way more going on that I didn't completely understand.
In contrast, the approach that you're discussing seems to be to offset the audio signal so that it's raised above (the one and only) ground even on the negative swing. It seems to me that either way would accomplish the same end result. (Assuming, of course, "my way" is actually a valid approach.)
Again, all of my audio circuits thus far have been built with center-tapped transformers for the PSU, so I've always run at +/-5 or 15v with a real ground that is naturally halfway between. I've never tried a single-supply AC circuit, so I could be completely full of it.