It’s probably a little early for burn in to be complete but as @Driftwood has noted above I knew I’d be keeping it shortly after trying it out with DSD 512. Anyway I’d like to let @Rugby and @MusicEar know about initial impressions as I’m breaking it in.
Firstly Daniel’s post above about CPU grunt required to upsample to DSD 512 is spot on (as usual 
). I opted to build a new music server in an ITX mini case which I have named MUSE. That project nearly went adrift when some bodgy RAM had me foxed, but it’s cooking with gas now. It has some fan noise so it sits in the room adjacent to the stereo connected to the router by Ethernet and thence to the microRendu.
The GTX 970 CUDA assist seems necessary. Without it I get some pops from using the i7 Kaby Lake alone (can’t remember what filter). With it the whole thing hums along between 19 and 28% CPU at 55-60 C convolving Room EQ in Roon and upsampling from Redbook to DSD 512 (22.5792 mHz) using closed-form filter and AMSDM7 512+fs in localhost HQP.
I prefer closed-form because it gives both the biggest and tightest sound stage on my system. And I have found I am a sound stage junky. Fortunately the Holo Audio Spring and closed-form DSD 512 deliver my daily fix.
Some of the reviews that can be found from Guy’s links above seem to suggest that the sound stage from the Spring using NOS can be almost “too big”. Rubbish. In the same way that you can never be too beautiful, too rich or have too many backups, there is no such thing (in my humble opinion) as a sound stage which is “too big”.
DSD 512 is qualitatively different from DSD 256 and in the same direction that DSD is generally heading. I like the description organic. For me it is an absence of digital sibilance. I just don’t hear it any more. At last the dreaded bugbear of “CD sound” (remember perfect sound forever ?) has been banished. It sounds as analog as I have ever heard a digital system sound. And it’s not rolled off highs. The air is all there like with a good moving coil cartridge (I use a Koetsu Black).
I can understand, however, those who might prefer PCM. The DSD sound is still a bit sucked out in the mid-bass, lacking the presence of PCM and might also have a shade less dynamics. Fortunately the mR prevents those defects from going too far and becoming annoying. They are there but for me the sound stage and air make it worthwhile.
If I seem to be talking about DSD rather than the Spring, it’s because I don’t really think the Spring has a “sound” as such. The detail is still there compared to the Vega, but where the Vega pushed the detail to the fore, with the Spring it is just there in a more natural way. I don’t think of the Spring as warm or cold. I think it is just dead accurate. For $1,700 USD (Level 2) it is some of the best value in Hi Fi at the moment. You can have mine when you pry it from my cold dead hands.
Edit: I’ve now listened more with a poly-sinc-shrt-mp filter in HQP and I believe my comment about “a shade less dynamics” was referring to the closed-form filter rather than the Spring.