Alternate reality. Obviously the daily news feed on CNN's website looked exactly the same no matter what. For that, don't bother with any scenic detour. Go direct and call it a day.
As to which digital gyrations mattered and which ones did not, first the no-shows. External clocking. Be it two switches clocked off an OCK-2 or SW-6 slaved to SW-10, I couldn't tell internal clocking apart. I don't know whether it's material but my Singxer SU-6 USB bridge between iMac and DAC already performs ultra-cap powered precision OXCO reclocking.
1m fibre-optic link between two switches vs. 1m CAT8a copper. Across that distance, I couldn't tell the wiring apart.
On switch count, one was good, two were bliss, three a crowd. I'm sorry to bear bad news if like me you fancy a lower box count but, two cascaded LAN distributors really did beat one. To my ears it was the third which was superfluous.
20m fibre optics vs.CAT8a copper between router and first switch? This I could definitely hear.
With my hardware and location, I'd skip LAN clock sync. Ditto short fiber. I'd probably focus on executing long LAN spurs via fiber. I could replicate the SW-6's isolated RJ45 port with SOtM's inline isolators but prefer fewer physical connections and less costs. Both points gave LHY's built-in solution the nod.
Keeping it relatively simple and using local file playback as my sonic yardstick, my final solution went all copper into our existing SW-8 through a first SOtM isolator, connected SW-8 to SW-6 via copper pigtails bracketing another inline SOtM, then exited the SW-6 over its isolated port into the iMac's network card.
If I didn't already have SOtM inline isolators, I'd go router ⇒ RJ45/SPF adapter ⇒ 20m fibre ⇒ SW-6 ⇒ isolated copper ⇒ iMac. To my ears, either solution puts the sound of cloudy files on par with my local yokels. I should add that the skinny yellow canaries route even easier than CAT8a along floorboards and under carpet. I simply don't know how tolerant optical fiber is to being stepped on; or occasionally driven over by a furniture dolly moving speakers. Our router sits in my office. To hardwire my music iMac to it and my wife's computer one further room over means that two very long LAN spurs cross a hallway beneath a thick throw rug. It's also where speakers enter and exit the sound room. Our CAT8a wiring has proven impervious to such foot and load traffic. Would fiber optics do as well?
Real alternative. With my takeaways on 'how' unpacked, time for 'what'. What was the sonic gain between direct; and detour through 2 x network distributors plus 2 x inline isolators? The absence of virtual negative feedback. Here I'm not debating NFB in general. Different circuits, executions, results. I'm on about adjustable feedback in valve amps, particularly SET. Far more often than not, their sound-not-measurements-first owners prefer either no feedback at all; or a minimal amount like 2dB. It's virtually never the maximum amount. Why? In such circuits more feedback gets ever more dry, uptight and mechanical. Less feedback gets ever more fluid, free and gushing. For today's review, my single word for the detour's benefit is juicier. Now you know how to elaborate on that. You also understand that the sound didn't break without it. Like NFB adjustments on a SET, the difference was a feel/gestalt shift and no generic treble/mid/bass or soundstage thing.
An Enleum AMP-23 drives EnigmAcoustics Mythology 1 monitors. Audirvana Studio is the Window 10/64 sound engine here in Qobuz streaming mode.
My OCK-2 sample was simply far from deflated just because my main LAN use didn't clock it. On my desktop both Singxer SU-2 USB bridge and iFi iDSD Pro Signature DAC sport 10MHz clock inputs. The former's dip-switch belly panel activates it to then confirm clock lock with a red not blue central LED. The iFi has a rotary switch on the back which opens its 10MHz BNC input.
LHY's clock cables with Neutrik BNC and Gotham wiring sell for a very reasonable $49.
