I have a storage room next to my listening room, so drilling a hole and drawing an ethernet cable will be easy (I have to make holes for display cables and USB anyway). However, the room is only about 1.5x2 meter and no ventilation, I wonder how warm it would get in there with a heavy-duty PC.
Btw, I got recommended by an “expert” to buy this CPU, not to expensive, 8 cores and a 3500 thread rating at base speed: https://www.cpubenchmark.net/cpu.php?cpu=AMD+Ryzen+7+5800X&id=3869
Access points by HPE (Aruba), powered via PoE. Nice small flat white device that doesn’t look disturbing. Network switches are both HPE and Cisco and feed PoE power.
I just yesterday received my Ryzen 7 5800X, I will put it together next week and then I’ll do some testing and see how it performs.
Excellent! Please report your findings, and maybe optimize HQPlayer a little extra for it 
Btw have you tried PCM to your ADI-2 with: PCM768k + ext2 + LNS15 ?
I’m much preferring these PCM settings with ADI-2 compared with DSD256 Direct (with EC modulators).
Even though I have a i9-9900K I’m much preferring PCM output mode at the moment.
One of my setups is limited to PCM anyway (multichannel Focusrite DAC for DSP crossover and room EQ) but I love HQPlayer handling this DSP, I don’t feel like I’m missing anything by not being able to do multichannel DSD (even thought my machine probably can’t handle it).
Same with my ADI-2 FS FAC (V2) 2-channel listening.
I’ve been playing with EC modulators for a few months now.
PCM mode just feels (ultra subjectively) more dynamic with better transients. Punchier.
My fanless (silent!) UpBoard Gateway with Intel Atom (running HQP OS) handles all the PCM I throw at it, including 6-channels of PCM192kHz and 2-channels of PCM768kHz.
Yes, I used to up-sample to PCM 768 with ext2 and LNS15, but DSD gives a more complete sound in my opinion (more solid soundstage, slightly less edgy etc). The only drawback I have found is that the soundstage becomes a little more narrow (still way outside speakers though). But this might also depend on the rest of the system (I have NC500 class D amplifiers and Volent VL-2 Speakers).
My digital chain is: consumer ethernet switch with SFP port -> fiber -> opticalRendu -> microRendu (with Crystek clock) -> RME ADI-2 DAC. Both rendu and the DAC is run on LiFePo4 cells which lasts way more than a DAY so I charge when sleeping.
I don’t think I can optimize anything more, but if needed I can adapt the multi-core auto-tuning like I did also for Apple’s M1.
I’m not fan of AKM’s modulators and you’d still have some correlated images. But of course from technical perspective it works.
Just remember to set PCM volume in ADI-2 to -3.5 dB when comparing. Otherwise you will have level matching issue.
By the way, which AKM chip version do you have? My both ADI-2 Pro’s are the first generation before the DAC chip was changed.
You may then like ESS the most, and maybe Cirrus Logic too. Can almost make ears bleed.
LOL that’s a bit xtr Jussi ! (extreme 
)
I have the V2 version of ADI-2 and the Focusrite 18i20 Gen3 uses AKM DAC chips too.
My Merging Anubis uses ESS chips too - no ear bleeding yet (that I know of ! 
).
I’ve never heard anyone describe any Merging DACs (example only) as bright and they’ve long used ESS chips?
I have practically never used Merging gear in PCM mode. But I have plenty of ESS / Cirrus Logic DACs. ESS has it’s own signature even with DSD, because it is not fully native.
It’s the same chip as used by EVGA NU Audio, but I certainly prefer running that card with DSD. Otherwise it sounds harsh, edgy and distorted.
But in any case, remember to compensate the PCM level by setting ADI-2 volume to -3.5 dB, otherwise your PCM and DSD Direct mode levels differ.
When I feed these DACs PCM768kHz (353k for Anubis) to compare with ASDM7EC at DSD256 (proper volume matching) it’s not a brightness/harshness/edginess thing that jumps out at me.
It’s a feeling of the punchiness and it’s not a huge difference.
But the highest rate PCM from HQP always feels a tad punchier.
It FEELS like a time domain / transients thing? I have no idea though
I always use ext2 to compare because that’s my fav.
The remaining images from PCM?
You get better idea with 1.5M PCM from Holo vs DSD256. But you need to turn up amplifier +6 dB for DSD, since Holo DSD level is 6 dB lower than PCM level.
Definitely possible since it’s there in the measurements
Can’t be anything else ?
Could be partially something else too, but that is most obvious part. Modulators are another, more complex source. End result is combination of things.
You also get some idea comparing modulators, filters, rates and PCM/DSD on a DAC that is just a DAC.
I compared ADI-2 V2 feeding superbly designed (measuring) SMSL SP200 THX AAA 888 headphone amp.
So comparing using same headphone amp section.
Maybe it’s time I demo it’s bigger brother (or dad) the Benchmark HPA4
But I know what you mean, ADI-2 is still doing some DSP when fed PCM768k, whereas Holo Spring 2 isn’t when in NOS mode and fed PCM1.5MHz.
Maybe I should demo a Holo with HPA4 LOL
Yes, and more importantly the remaining image levels at 1.5M rates with Holo’s output analog filter are very low. When you set DAC bits to 20 and use something like LNS15, then you have very low level images and distortion with PCM too. And no additional DSP. You just need to compensate for 6 dB level difference. You can turn down HQPlayer’s volume by extra 6 dB for PCM output, so if you use -3 dBFS for DSD output, set it to -9 dBFS with PCM for comparison.
P.S. The combo in my picture is very good price/performance ratio. Alternative step up is T+A HA 200 with Solitaire headphones.
I’ve been really enjoying a Spring 2 KTE with NOS, Phonitor XE amp, ZMF Verité closed headphones, no upsampling, but this is telling me I should try HQ Player Just need a quiet powerful server for this, as I’m in the fanless camp as others here. Waiting for Cirrus7 to get a US distributor…
Interestingly though, to my ears when ADI-2 is fed PCM768kHz with ext2 + LNS15, it sounds overall much like my Chord Hugo2.
DSD256 (Direct) goes away from this kind of sound signature (subjectively, to my own ears).
Hugo2 has no correlated images btw (at least to 1MHz bandwidth):
On my RME ADI-2 I put volume locked at -3.5 dB and compared PCM 768 ext2 LNS15 with DSD128 ASMD7EC and ext2 (in DSD Direct), and to me DSD gives a more refined sound with an almost floating feel, compared to a little more crude and edgy PCM.
For example, on the magic-sounding tune August on the surprisingly good Folklore by Taylor Swift, PCM was more left/right while DSD was more coherent.
