When something puts a lot of ultrasonic stuff, there is every opportunity for it to generate IMD products in the audible range, which may or may not account for perceived differences. There might be something else to it that we don't know.

Before you mentioned this, I went back and A-B'ed a DSD file with no filtering vs 50khz filtering, in order to see whether no filtering was audibly better.   To my mild surprise, I found that the 50khz filtering sounded slightly better than no filtering.

Then I read your comment above, which explains this (if the order had been the reverse, confirmation bias might have been involved).  Not scientific, of course, but I think it is more confirmation that people should use some filtering.

I agree with you that audio stuff is not necessarily fully understood, and that "end-game"/"wire-with-gain"/"the-one-to-rule-them-all" gear sounds sort of mythological ... But in my mind that does not mean developers have no idea WTF they are doing, and great sound is purely a result of random chance and divine providence.

The epicycles reference was not referring to gear development - which is where theory and measurements are most important - but rather to the existence of sound differences that are contrary to any existing theory.  If one has an outlook that is 100% based on science, then the history of science clearly shows that in 2050 or 2100, we will likely know a lot more about digital audio and digital circuits.  So, the Objectivist viewpoint "you can't hear better bass in 24/192 because the theory doesn't support that" is actually unscientific.

Who's coming up with such silly ideas anyway?

It's quite simple to check if there's something to it: first ABX with careful level matching and if successful describe the differences and also visualize the difference between both signals.

All those myths out there are accepted blindly as subjective truth (heh) because some people are not able to think (rationally). They want to hear difference, no they need to hear differences. Their hobby isn't hi-fi in the literal meaning but hearing differences.


Anyway, even 24/96 > DSD from a purely technical point of view.

PS: If you just click [smite] and leave without a response you are just a coward, like the other ~40.

Yeah some DACs/amps definitely do seem to have troubles with ultrasonic content.
Now we just need to convince people that distortion during reproduction is bad... 

I don't worry about the number of smites, I just "dislike" the people that hit and run.

I thought somebody worked out that DSD64 is roughly the equivalent of 24/117 or so.

The thing I hate most about DSD is how much of my life was spent tracking down 2 PS3s that could rips the things, mod the shit out of the cooling on the second one so it wouldn't die on me like the first, and then ripping my SACD collection.

Consider that the alternative used to be recording the output manually through an ADC. DVD-A was by far the more convenient format thanks to DVD-A Explorer. Usually the more convenient format is the one that wins, but not in this case. Neither format really "won" anything, but DVD-A is as dead as HD-DVD.