As a producer of a musical recording that is on the itunes store I can tell you that the 24-bit 192 khz version definetely sounds better than the 256 kb AAC version for sale on the iTunes store.
The 24/192 version sounds much closer to how it sounded when I recorded it in my studio, the version for sale on the iTunes store isn't terrible but if I had to A-B I could tell them apart instantly - I probably listened to that darned thing a thousand times in the mastering process.
Ok but that's because you listened to it thousand times. How about something you listened only once? Do you think you can differ between AAC and 24/196 in a A/B blind test?