Pabelson, actually the real impact of sampling at a higher frequency is not to extend the frequency response, but to make the implimentation of the anti-aliasing filter simpler, so that a filter with less passband ripple and less phase error can be used.
I agree with you that extra high frequencies above 20kHz do not make a difference.
I think this goes to the heart of your arguments ... in THEORY redbook CD is perfectly capable of producing perfect sound. Provided there is no jitter, the ADCs and DACs are perfectly linear, and most importantly, that you can implement a perfect brick wall anti-aliasing filter in the frequency domain, which creates the sin(x)/x function in the time domain to perfectly reproduce the analog waveform from the train of samples.
It is in the implementation that the CD playback falls short.
I agree with you that vinyl is a deeply flawed format, and I personally think that vinyl and CD replay are roughly on a par, with different deficiencies. I suspect that those who prefer vinyl are in some way more bothered by the distortions caused by the deficiencies in real-world digital recording and playback caused by timing jitter, non-linear DACs and ADCs and approximations to the brick-wall filter.
That vinyl should be considered "accurate" by the same people who argue endlessly about how different cartridges can sound at different tracking weights, how different tonearms sound with a different counterweights is equally ludicrous. But vinyl does have moments where it sounds more real than CDs, so I guess I kind of like the distortion introduced by the medium.
