Yes, the Amy Winehouse CD is a great example of a bloody awful mastering job. The sound is terribly distorted in many places. This is one of the CDs that lead to the original post in this thread, by the way.
No, this is not my setup. When I hear a younger user like Rasmussen claim that nobody can hear clipping and that the problem is probably my setup, I begin to fear for the future of recording...do our younger users even know what a properly recorded album sounds like? It's quite possible that none of the music they enjoy has ever been recorded properly.
ReplayGain (using the clipping option) definitely reduces the perceived distortion (even after volume matching to compensate for the reduced volume that using ReplayGain causes). My understanding is that digital clipping is caused by the data in the signal reaching the maximum value...ReplayGain would certainly prevent this from happening if it were implemented correctly. But it's easy to prove: listen to the Winehouse CD (or any other where you notice distortion from clipping) with and without ReplayGain...to me, listening with ReplayGain certainly sounds a lot better than listening without it. By the way, I am not using any DSPs in my current Foobar setup (and my ReplayGain pre-amp gain settings are both set to 0db)
I fully understand that ReplayGain will reduce the resolution of the digital recording (by scaling the values in the data by a certain amount), but in the case of a very poorly recorded CD this seems to improve the sound. And, in the case of overly compressed data in the recording, reducing the resolution a bit probably isn't going to matter much because the signal is probably at or near the max for much of the recording anyway (i.e. there is very little dynamic range).
Despite claims on HydrogenAudio to the contrary, I've found that the "Advanced Limiter" in Foobar2000 is almost useless for reducing distortion caused by the clipping problem, by the way. Also, the clipping indicator in Foobar appears to be broken (or representing something other than what I expect it does). I remember that then one in the E-Mu PatchMix panel seemed a lot more accurate (although I'm not currently using an E-Mu soundcard so I can't verify this).
I guess some artists might actually want there to be distortion in their recordings (so the comment about just accepting recent CDs for what they are is appreciated), but in many cases it just seems like the engineer didn't really know any better (or was instructed to just make the recording sound "loud" at all costs).
The problem is so widespread that I'm beginning to wonder if there isn't some sort of conspiracy going on...something that will make us want to purchase some new format of music in the future that isn't sold with criminal misuse of compression in the digital data. The thought of having to re-purchase (and rip) my collection in "CD 2.0" format is not very appealing (and it will almost certainly contain the type of DRM crap that SACD has in order to prevent ripping in the first place). Perhaps this is yet another unfortunate decision by the recording industry that will backfire on them (instead of increasing future sales, maybe it will lead to more declines in record sales).