I won't dispute the fact that you can distinguish, 160k vs 320k. I'll give you the benefit of the doubt, I'm ready to believe people can hear artifacts at 160k

I won't dispute the fact that you can distinguish, 160k vs 320k. I'll give you the benefit of the doubt, I'm ready to believe people can hear artifacts at 160k even with the best encoders, especially if these people are trained for these artifacts, have a good headphone setup, and are listening to particular problematic passages (higher frequencies, transients, etc). However, just for information, there is a lot more to lossy audio compression than just bitrate. A song encoded with modern LAME with VBR at around 64k or 96k possibly sounds better than the same song encoded with a napster-era encoder at 192 CBR. Modern formats such as AAC or Vorbis are even more performant than LAME. I don't have hard data to back it up, but I think 99% of people would not differentiate between lossless and and lossy with a mid high quality setting (lame at v3-ish, whatever the equivalent is in the best aac encoder these days), even on a great headphone setup. Maybe if they were specifically looking for artifacts in particuliar problem samples.