Here's the thing: you can't
listen to an ethernet switch unless it makes a noise itself. If you put an ethernet switch into a system you must listen to the systems, which means "the sound that comes out of your speakers". Ideally,
ideally, when you make one change in the chain you could hear the difference between before and after. However, making that change involves a bunch of emotions ($650 tends to do that) and expectations, and it takes so long to do it that once you find your favorite song and try to listen for micro differences you've literally forgotten what the system sounded like before you made the change. Humans are brilliant at recognising sounds but crap at comparing two sounds once a few seconds have passed between them. The
McGurk effect is a brilliant example that our hearing is not a particularly clever sense compared with other stuff we're equipped with (like sight).
Other than that, the switch thing claims to reclock the digital signal (which is unnecessary because it will be reclocked in the DAC anyway), to reject "external leakage currents" (which Amir shows is wrong), and ensure "maximal signal integrity" as if normal switches leaked packages left and right (they don't). And so on.
Amir did measure the TDAI-3400 and,
contrary to his bias he found that it wasn't as good as he hoped. It's not a surprise that he has preconceptions and that might bias his reasoning and judgement of a product. Preconceptions have been described in philosophy since literally the beginning of philosophy. If there ever was a universal human trait, that seems to be the one. However, what Amir is doing is to
explain why he reaches the conclusion he does, by showing a) what he measures, b) how he measures, and c) what he thinks it means. Since he does all those three things, you can either dismiss his choice of measurement, his method of measurement or his conclusions based on those measurements.
That's why you can look at the graphs of the TDAI-3400 measurements and go "hm. They kinda look wonky, but it still sounds good. That might mean that the TDAI-3400 could be better, but it still sounds good. Maybe the measurements are more thorough than I reasonably need." Or you can go "I'm frickin paying thousands, I should get -120 dB distortion peaks!" At that point, it's your choice.
Sorry for the rant and I don't mean to come across as hostile. As Voltaire or someone (don't have time to look it up) might have said, "I would have written shorter/less hostile but I didn't have the time".