People can choose to believe what they want, regardless of facts.
Also, it's important to notice that hearing is believing.
By hearing I mean the feeling that you have heard.
It is impossible for a human to discern whether something really has been heard or whether it is imagined, when we are talking about signals at the treshold of audibility.
As such, the only way to test for this is a proper AB-sub-type test (depending on research question, either ABX, ABC/HR or something else), even with all the problems related to sensory blind evaluation.
I for one can claim that after nearly three years of training, very good setup and very good absolute level of hearing (in terms of audiologist measurements), I find it very hard to hear artifacts at 128kbps in most material when properly compressed either by LAME/MP3 or Apple/QuickTime/AAC at certain settings.
Yes, there are instances, where the artifacting can be considerable, but in probably c. 90% of cases it's complete opposite for me. And I've trained, using most of the freely (and for-pay) available lossy compression material.
Hearing the artifacts has been proposed to be a function of training time. The artifacts produced by lossy compression are not similar to those produced by bad speakers, lousy amps or inferior recording techniques. Your auditory system needs to be trained to be able to discern them more acutely.
I also helped to organise a double blind test (using randomized cd tracks) and with high end audio (stereo/speaker) gear using four serious audio enthusiasts with years of listening experience.
None of us could consistently discern AAC/128kps (Apple's implementation) or Lame/128kbps (slightly modified alt-preset) in statistically significant manner on ALL of the tracks. On some yes, but not on most. WMA9 was a slitghtly different matter though
This test will be published. It's not just some guy blabbering on a forum
This of course does not mean you cannot hear differences yourself, but you'd probably be surprised if you really did a proper blind test (or several of them) to yourself using a software that doesn't allow you to cheat.
So, if you don't hear any problems immediately at 128kbps (HQ/Stereo from latest iTunes), it doesn't mean you are deaf.
Psychoacoustic compression does work, when done properly, to an amazing degree.
But you don't have to believe me, you can keep on using any which compression (or lack of it) you desire and be happy.
That's what counts after all, isn't it?
Being happy and enjoying the music
regards,
halcyon
PS Just so that nobody will use my words as ammo for his/her own crusade: you cannot prove the non-existence of something. Also, sensory evaluation is not easy to do right. It is easy to do a null test where almost nobody hears any differences even with really badly compressed material. All you need is a bunch of people who don't believe they'll hear any differences and don't really want to put an effort into trying to hear. Self fulfilling prophecy: no differences! A useless test.