As I have a fairly decent headphone setup for some time now. I decided to do a double-blind test1 between lossy and lossless music streams. Picked the quick 5 samples test (5 songs, 5 tests per song). The outcome? Got 74% right (statistically valid) so potentially I should be able to hear a difference. I think it helped that the test included one of my favorite artists though. And even with this outcome I still doubt how much is plain luck vs really being able to hear a difference.
It would be interesting to do the 20 sample version someday. But even from the quick test, I learned a lot.
First of all, it’s incredibly hard to spot any difference (or believe to hear a difference). But more importantly, I needed to focus on super small and minor details of the songs and concentrate to unpleasant levels. So all in all differences are super small and negligible for me. A high-quality lossy codec is impressively good (way better than I expected).