You are crazy, I don't know how else to express it. Downsampling is not "digital hocus pocus"; it is throwing away some information, but a file downsampled to 24/96 still has more than 4x the information than a file upsampled from 16/44.1, to say nothing of the improved sound from doing the A>D conversion in a way that avoids the need for anti-aliasing filters and with less quantization error as well. And if you want to avoid the downsampling then listen to the highest digital resolution file; no hocus-pocus involved.