I'll discuss a key flaw in his interview, which Chris has also pointed out. When he claimed that higher sampling rates allowed for more samples(especially at high frequencies)and therefore better "resolution", I almost stopped listening because this indicated a complete misunderstanding of the digital sampling process, regardless of his experience in the industry. It was proved mathematically by Nyquist and others over eighty years ago that as long as there were at least two samples at a given frequency that the analog waveform could be reproduced precisely, not just approximated. This has of course also proved to be true in practice, including military and industrial processing where accuracy is a bit more crucial than in audio reproduction.
Some who lack an understanding of the basis of the sampling process view the samples as if they were involved in a sort of "connect-the-dots" procedure where the more dots you had, the closer you got to the original analog waveform, but you never could get there with full analog completeness. Again, nothing of the sort is going on; only two samples are necessary to establish unambiguously the precisely frequency involved at that instant. The perhaps 2000 samples available around 20Hz(at a 44.1Kb/s sampling rate)in no way add more accuracy than the as little as 2 samples taken around 20KHz, and the 20KHz wave is reproduced just as precisely(rather than being simply a straight line connecting 2 dots).
With his misunderstanding coloring his subjective views, this doesn't allow me to take seriously any of his claims about "amazing", "unbelievable"(at least as he intended the term), etc. results he experienced.
