That confirms my expectation and belief of a technically oriented person. However, in my setup still all sources (CD, Qobuz/ROON, Qobuz/Audirvana) sounds differently. So what causes it? Some weird EMI / RFI or what? For the record, I play Rotel CD -> Yggy SPDIF or MacMini -> Qobuz / ROON -> PI2AES -> Yggy AES or MacMini -> Qobuz / Audirvana -> Yggy Unison USB. CD still rules them all. Differences are not huge, but quite easily noticeable.
That's a good question. One thing is that I only checked these specific devices and none of yours. We can't say for certain that yours are all bit perfect as well without checking.
Different devices produce different levels of jitter and electrical noise. Supposedly the more jitter the DAC has to remove, the more noise it produces itself.
In your case, using different interfaces is another compounding variable. I use a Mutec MC-3+ USB reclocker in my main system, fed by a TOSLINK switch. I tried all outputs the MC-3+ USB has into the Yggdrasil OG and preferred BNC (I did not compare USB). But if different inputs can sound different, presumably outputs can as well, even on the same device.
And then there are cables. The RCA and BNC coax cables I used are comparable, same cable, same lengths, just different connectors - but RCA is known not to be ideal when you want to maintain 75 Ohm. The AES cable I tried is also from BlueJeans and also 5 ft long, but of course a different cable with an extra conductor and 110 Ohm. And TOSLINK is a wholly different technology (does not apply in your scenario). When I compared short HDMI cables for I²S use between a Singxer SU-1 KTE and a Holo Audio Spring 2 KTE I also heard differences. Basically there is no way to make things perfectly identical when using different input types.
Then there is the matter of cable length. Electrical or optical, attenuation can play a role, or noise picked up along the way, and there can even be flipped bits. A 25 ft TOSLINK cable I have was mostly bit perfect, but not quite. Even with more reasonable lengths, I've read claims about the possibility of reflections if the impedance isn't quite right (hence the 5 ft length, though from what I understand that recommendation is dependent on the sample rate, and thus the signal frequency, and thus the wave length).
For my comparison, I'll move the source end of a 5 ft RCA to BNC cable between the two devices to keep that part completely identical. I briefly thought about getting an Inday TOSLINK/coax switch that I had my eye on for a while anyway. One variant of it has an RS232 interface that I could theoretically use with an ABX testing tool. But that's an active device, so it itself is likely to reduce the differences between the sources somewhat, so that actually doesn't make sense for this test. At best I should use a passive switch, but I don't think one for analog signals (like the Schiit SYS) is actually advisable and switches for BNC seem incredibly rare, leaving various F-type switches as an option. But then you read about insertion loss and it becomes clear I wouldn't really be comparing what I'm trying to compare - how different these two devices sound when directly connected to a DAC in the same way.
So basically everything matters a little bit, it would seem. I wouldn't even be dismayed to hear differences.
What I'm really hoping is that I can make them disappear with a reclocker, and possibly additionally by using the TOSLINK out of the streamer instead of coax and converting the coax out of the Blu Ray player to TOSLINK. Start with a bit perfect signal, use TOSLINK for perfect electrical isolation, and finally use a reclocker to treat jitter.
I can compare all that treatment stuff even with the same source, ultimately, but it would be more fun with two different sources that produce the same bits, but sound different.
Since the sources are bit perfect, in principle I should be able to eliminate/drastically reduce all differences this way - except for the absolute speed. All reclockers have to adapt to the source speed, otherwise you'd need a considerable buffer, which is problematic in conjunction with separately reproduced video. A universally useful audio reclocker would have to be a video reclocker as well to keep things in sync, but with the various types of data transferred over HDMI in conjunction with copy protection that's definitely not trivial.
I could create a test signal that is exactly 1 hour long, play it twice or more from each source, and record the analog output of an attached DAC with an ADC to check whether the sources have absolute speed differences, just to try and account for that. Hm, I'd have to burn it to a CD, I suppose.
Been a long time.
No, I don't have a life, thanks for asking.