I responded point by point to Clever Hans and of course, true to form, he can't rebut a we reasoned response so he deleted this:
"You've deleted technical posts by myself and others in the past. Let's see how this one goes (I'll be sure to screen scrape right after I've posted).
At
8:01 you get into jitter and phase on the PAM encoding on the Ethernet signal. If you have a powerful enough measurement device you will find deviation from TX side of things. This is especially true as cable lengths increase and toss in poor terminations, and as BlueJeans cable proved, with many no name cables you get out of spec solutions. But you have a failure in your understanding and you are attempting to teach that which you don't know.
Your first issue is that the switch isn't the biggest indicator of signal integrity. It's the interconnects.
Your second issue is that with interconnects that are in spec wrt to signal integrity at each end all that data is buffered and copied, multiple times. You should read up on clock domain boundaries.
At
12:45 you start talking about buffering. You seem to be conflating various inputs (S/Pdif, AES-EBU, USB) of a 2011-2013 design and the ability to buffer. If those inputs have ground tie you could have an issue that isn't related to buffer. BUT Ethernet cabling isn't tied to any ground in unshielded applications. Also we are not talking about AES-EBU vs TOSLink vs USB vs SP/DIF co-ax.
We are talking about one upstream switch vs another. Again you are confusing your inter-connects with regards to buffering. Now if you want to make the argument that one switch vs another is going to introduce some constant analog noise just by virtue of being plugged in please let me know. I have a response all ready to that. And yes I've had the eR, and Audio Quest, WW, Nordost Ethernet cabling. "
He uses a Chord QDB76 HDSD as an example because it has a 3 second buffer and he can tell the differences in whether the input is AES-EBU, TOSLink, S/Pdif, USB. That DAC is from 2013 and:
1. It doesn't have an Ethernet input
2. He's talking about inputs right on the DAC itself
3. How can he fail to realize we can take the Chord QDB76, connect a computer to it's USB and on the other end feed the computer via Ethernet and swap out the upstream switch!
He can't be that daft, can he?