Originally Posted by HFat
... and that's a bogus test. You need to convert the sample yourself (and volume-match if necessary). Of course, it also needs to be a blind test.
X2
The differences (very very slight) I noticed (in a blind test, I scored 13/15) between a 24/96 sample and a 16/44.1 sample downloaded from the same site vanished when I downsampled the 24/96 myself (score exactly 50%). I just did it again - same 50% score - this time through a better USB external soundcard outputting an optical digital to an external DAC.
Ooops - my Entech DAC is only rated at 48K so I had set the sound card to 44.1 yesterday , I redid the test this time using the sound card at 24/96 but this time using the analog outs to my M^3 and Sennheiser HD580s - still the same result.
Annoyingly I cant do a blind test of 24/96 vs 24/44 of the sound card itself as you have to physically set dip switches to change mode then re-initialise the card.
Fwiw when I analysed the original 24/96 and 16/44.1 files in an audio package there were obvious (though admittedly small) visible differences in the waveforms these were far less apparent after downsampling.
YMMV of course.