Something I apparently didn’t mention yet (I thought I did) is that the noise I get is periodic to the block length of my signal. So at 16 samples, the noise I get is 44.1/16 = 2.75kHz. So it’s either coming from the ADC/DAC or it’s the supply needs spike that results from having to process that block. I think the first scenario is more likely. I also noticed that I get the usual buzz in addition to louder digital noise when I unplug the input. Likely due to the high resistance „pull down“ at the input and induced noise.
My reference voltage is mostly clean now. That leaves the two options noise from the OpAmps supply or induction due to breadboard. I already moved my ground away from the noisy supply line though, it’s better then before.
Regarding a "more compact and well layed-out implementation". I'll probably use THT and perf boards for the final board. So the layout will mostly follow what's easiest to implement. Though I probably will redesign the circuit to move the noisy supply line as far away from Ground and anything else as possible.
Maybe I'll just design a PCB anyway. Would be my first try though and I have the strong suspicion, that I'll manage to mess up the first try.
I might have messed up earlier. I tried to RC filter the supply to the opAmps again and this time I managed to completely remove the noise, at least to the point where the ugly digital noise was masked from the general analog noise floor. So 100Ohm / 100uF RC filter, that drops my voltage supply from 9V down to ~8.5V according to spice any multimeter.
I think I still got noise earlier because of induced noise. simply by separating supply voltage and ground, this isn't a big issue anymore. I now get around 90dB SNR, sufficient for a first-try DIY project.
Thanks for all the help regarding that topic!