I’m doing the version with the nanoDigi:
Mac mini optical out > nanoDigi (crossover, time delay, EQ) > 2 DACs > active speakers + active sub.
I only crossover the sub and let the speakers play full range. I tried the SHD solution first, but there was no difference in SQ (didn’t use Dirac).
But, beware the mandatory sample rate converter in MiniDSP products that resamples everything to 96 KHz. The nanoDigi sounded strange with 44.1 material, but transparent with 48 and 96 material.
I measured the speakers in-room in REW, and you could clearly see some artifacts in the treble with 44.1 material. This did not happen with the SHD. (User March Audio did measurements of the SRC over on the ASR forum, and showed it was subpar).
In the end, I kept the nanoDigi and just resample everything in Roon to 96 KHz. That bypasses the MiniDSP’s internal SRC.
Another issue is that the nanoDigi can only delay channels up to 9 ms, whereas the SHD can delay channels up to 30 ms. This is relevant since group delay in the bass octaves is very often 20-40 ms in my experience.
This means with the nanoDigi, you cannot time align the sub though you can of course phase align the system at the crossover region, which is the most important. (Time coherence over time consistent drivers in the bass).
A final issue is that the SHD works as Roon endpoint through Volumio, so no need for a Mac/PC in the chain and using its optical out.
With the SHD, you do pay for more than just Dirac. But the nanodigi solution is just way cheaper and sounds good.
PS. I am eyeing the Okto Dac8 as well…but right now, I cannot fault the setup (apart perhaps from not being able to play DSD and MQA, which is meh when you get to have tight, deep, and even bass response). If I can do it all in Roon including time delay and crossover with the Okto, then I will prob shift to that.