I'm not following this whole thread, but would just like to agree (strongly) with one logical point that's been raised about the way some people's opinions have been formed on this topic. If you take a digital audio recorder and you record one performance at one sample rate, and another performance at another sample rate, you may think that you're comparing "the sound" of the two sample rates. Many people have done this and have come away with strong opinions; to them, their opinions are based on first-hand experience, and seem valid.
But other people have recorded and listened to identical signals at two different sampling rates. Many in this second group of people also feel that there are audible differences (particularly after learning what to listen for, which doesn't stand out to most people)--but quite a few people don't hear any definite difference when the identical source is used for both sampling rates--particularly if someone helps them out so that they can listen without knowing which recording they're hearing at a given moment. Those people tend to have strong opinions less often, though their experience lends greater credibility to their conclusions (or lack thereof), since there are fewer uncontrolled variables in the comparison.
Now here's the thing: Even back in the early 1980s when digital audio was still new and there were no high-rate A/D converters--the highest was Tom Stockham's 50 kHz Soundstream system--it was widely acknowledged that some converters sounded better than others. And that's huge as far as this discussion is concerned. If different converters and recorders can sound different when running at the same sampling rate, then another recorder or converter could always come along and sound better (or worse) than anything else you've ever heard at that same rate, and reset your expectations.
Conclusion: NOBODY can ever be sure that they've heard what any particular sampling rate "sounds like." Sampling rates in themselves can't be shown conclusively to have a "sound"; only particular implementations of them can, on specific recorded material.
One classic case of this occurred back in the 1980s, and I think that a lot of pro audio people formed their opinions on its basis. The first-generation Sony studio DAT machine (PCM-2500 if I recall correctly) was the first widely available digital recorder with a front-panel switch to control the sampling rate. Those decks were widely judged to sound not so great at 44.1 kHz, and better at 48 kHz. So a lot of people decided, aha: 48 kHz sounds a lot better than 44.1 kHz.
But in those decks, whenever you switched sampling rates, you also switched the analog anti-aliasing and anti-imaging filters. So the quality of those filters was a variable as well, and the 44.1 kHz filters had particular problems. As some of you may know, Apogee got their start in big-time audio by supplying lower-distortion aftermarket filters for Sony professional DAT and DASH recorders. With Apogee filters in place, the 44.1 kHz performance in particular became cleaner, and most people no longer heard any big difference between 44.1 and 48 kHz sampling on those decks.
But here's the thing: The great majority of people who had formed an initial opinion never made the follow-up experiment with comparable-quality implementations at the different sampling rates. So they still held on to their initial opinions, because no experience of theirs ever contradicted them. Plus, by then certain people had staked part of their reputation on their ability to hear the difference between sampling rates (can you actually imagine--people having egos in the music business)?
And that is the situation we find ourselves in. In general, the strength of anyone's opinion about the "sound" of different sampling rates (or many other topics such as tubes vs. transistors, discrete vs. op-amp, etc.), and their judgment of how relevant their listening experience actually is to the question at hand, may be inversely proportional to the amount of truly relevant listening experience that the person has.
--best regards
