First, AP Mastering posted another YouTube regarding the fact that the viewers who took his test could not tell the difference between the different portions he was playing. He had over 1,000 responses and his results are extremely statistically significant. That doesn’t mean his test proved what he set out to test, only that viewers who tried to determine how many cuts were in his loopback were unable to do so.
https://www.youtube.com/watch?v=LbcEgxm6RTU
Second, my own opinion is that there are small differences between DACs that may show up on a particular system set up in a particular way. I think the biggest differences are the different types of conversion be it what kind of chip vs r2r vs fpga. I heard a friend playing a very cheap DAC which sounded alright but adding a more expensive (not that much more) DAC opened up the soundstage significantly.
I A/B’d a T+A Dac vs an ifi iDSD pro and couldn’t really tell the difference. This was not a blind test either. I think AP Mastering has it correct that most people could not tell the difference between dacs if they are volume matched and in a blind test if the dacs use the same conversion method.