you are talking to nerds from DIYAUDIO (some of the nerdier nerds, even)
. at least 3 participants in this conversation over the last few pages.
1 and 2, i'm with you, although this isnt the development thread proper I wouldnt think, at this point. I'm already doing what i'm doing, but it sounds like a good idea if you want to turn it into a much wider project, perhaps. We have made a pretty good start on the wants and needs, but it needs more work for sure. although I wouldnt want to put in a bunch of time and then have the design not used for whatever reason, as I would think most semi-pros and pros, with limited time available to them. It depends on the prizes I guess ... lol!!! It does sounds a bit like design by committee and that almost never turns out well in my experience. better to designate leaders for different sections and work together towards a goal, once objectives and limits are discussed and decided. Thats my POV anyway. a project page could be started here in a thread, with a mud map and collected ideas, or set up a project at github
Ians stuff is good, but no reasonable current multichannel solution, is all Rpi based as well and that has no multichannel i2s (which weve discussed in detail here). He has no multichannel dacs and his 2 channels are not really set up that well for ganging together. Ive been following his project for years. It does cause o.5 seconds delay minimum, higher in lower sample-rates, by its nature, so sync with streamed video is a problem, if that is important to you.
It also isnt cheap, when all bits are added together, along with multiple power supplies and power supply/ground domains, battery boards etc etc. youve just pushed the price well over $500 all told, for 2 channel, without a case and opened a pandoras box of too many options for everything and a bit of a learning curve for setting it up, for the less experienced. It is a lot of resources spent on jitter, too IMO, which, past a reasonably easily achieved level, I think is the least important aspect.. Its relatively easy to achieve jitter in the low ps range without large FIFO buffering. actual measured analogue conversion quality of his dacs is easily matched (bettered) with any number of available solutions, they are simple, well laid out, but do not have onboard regulation AFAIK, since he favors batteries and supercapacitors. So you have to add your own power supply, or use his somewhat involved and physically large arrangement. He is the first to admit hes not an expert dac designer (he is a very humble man) and the dacs themselves are simple; it is the surrounding circuitry to provide isolation, control and clean i2s signals to his, or other designs that is where his focus lies. hes a fan of class A and his main dac uses transformers for IV conversion. linearity isnt what they are designed for (SINAD is likely to be underwhelming and not a design goal); although he does offer a simple but likely well performing opamp based IV stage as well.
his controller board and all PCBs in general are very well designed, the buffer and its peripherals do what they say on the box and very well, but there are some gotchas.
BTW, Volumio is supported on the ROCKPRO46 boards we have been talking about and it is a more powerful board than Rpi4, with more possibilities for onboard SSD etc.