Ah, I see that the thread went on-topic for a couple of posts. Apologies I posted an OT as well. All right ...
Indeed all is about "radiation" but radiation is is not just like that. It is about through air transmitted "voltage" which is captured elsewhere - an antenna. John Dyson pretty much tells the same but in better English.
If you get yourself the proper meters, it is super super easy to see how this radiation is influenced by really ALL you can imagine. Be creative at it ! It really doesn't matter what you do, it influences. For example, also where you are yourself (possibly depending on your shoes and the carpet). How cables lie is the more obvious and how they do or do not connect to PE (Protective Earth) is again more obvious. No wait, was that obvious ?
So yes. But measure it to see it.
None of this I really see back in an FFT (24 bit 192KHz) at the DAC output. Possibly in the power rails somewhere, but I never tested/measured against this.
Btw, the biggest problem with this radiation measurement is the localization of the receptor, or better, the possible harmless ones. Hard to explain in my English, but I mean: how to measure the magnetic influence of a transformer on e.g. the D/A chips; you can measure the radiation all right, but it is a most tough thing to measure "at" the places where you don't want the influence.
At some stage I bought all the shielding materials I could think of (some are dangerous to work with because conductive) and I simply gave upon applying them. Oh, they should help, but you don't even know what to protect from which. Thus, about everything radiates, and about everything can be an antenna as well. Now THAT can or should be prohibited by proper design (I'd day) but it is still not measurable (by means I know of) and otherwise you should measure a 100+ locations.
But that radiation is all over the place and that your audio gear creates it, is clear to me. And btw, it is not the gear as such (say that we forget about transformers, which can be shielded quite well) ... it is about the cables laying around. But for example, and in my own situation (never tested it elsewhere) ... unplug the USB cable at the DAC end and see what happens to the radiation figures at the end of that USB cable. And yes, I now, all logical when we think about it with some knowledge. But if I coincidentally look besides me, I see a phone connecting USB cable in a PC without the phone connected to it. That's how things go these days, and it is all over the place.
So I know it is there, I (and customers) can control it by diverse switches in the PC and the DAC (for those having both) and the result of these switches is at least very well measurable (in the DAC output - mainly the noise level or flatness (spectrum) of it). The switches "combine" for an unknown situation (no-one has connected his/her gear in the same way, and otherwise dimmers and such are around). Btw, this is now about coupled-in noise merely, which is probably why it "measures" more easily in the DAC output - it is just more severe (and more distinct). One thing: it always comes along with the radiation levels (so observe the FFT and meterS at the same time).
And just saying: if I try to find the best combination of everything (with radiation measurement) I am for about 2 hours trying combinations and then must give up because it is all to infinity. Luckily I don't use a preamp because it would multiply the number of permutations largely.
And no, I never measured radiation for CPU activity. I guess it can be done, but I just never thought about that; It is useless if you see what the "connections" already imply. Personally I don't expect a CPU to really radiate (see Marce's elaboration(s)). But, we supposedly can incur for it by an open ended cable somewhere.
Think of galvanic isolation and how that should work out for the worse.
I would say that an isolation barrier will create radiation. But somehow I could NOT prove that. So now somewhere my theories lack. I use it in the USB path and in the i2s path. Both are dead quiet (as far as again the "localization" of measurement can be applied - see earlier). Maybe it's a proof of all the voltage (current) being transferred and nothing leaks (and thus goes by air).