A chip is not a circuit. The circuit defines the usage of the chip. A chip can be used in many ways, for a DAC some of the more basic ones are disabling on-chip digital filtering and using FPGA. Dual mono can mean different things for different purposes, and it's not just about noise. There can be features only available in dual mono, an example is the PCM1794, which can only have external filtering in dual mono configurations.
For things like layout and routing, there are differences, particularly on the digital side, to ensure that additional noise is not injected before the D->A conversion.
Barring that, on the digital side, there are sonic impacts from filtering algorithms. This is one of the differentiators on the high end devices, where and how filtering is done.
Parts can be related like this:
Circuit schematic - architecture plans. This includes how the chip is to be utilized, what features, etc. i.e., one or two story house, concrete slab, basement, attic, etc. and materials to use, copper plumbing vs PVC, etc. A chip can be instrumented in multiple and various ways.
Floorplan/layout - execution of architecture plans. The exact parts to use, model of wallboard, accessories, etc., and how they are installed, how many nails to use, how to seal the seams, etc.. How many houses have you been in where there were issues that don't show up in the plans? Lots, if you've ever bought a house!
Audible, sure, using analogies. A house with paper walls vs drywall...wiring to code vs. a fire hazard. The list goes on. The flip side will be, are you able to recognize what it is you are hearing? Most people cannot, particularly at first.
THD isn't everything, but on the other hand, removing unnecessary distortion surely isn't a bad thing. Surely the "best" of designs would have both low THD, excellent impulse response, excellent phase distortion, as well as sonic pleasantries, etc.
People always argue about "what's on paper" and then saying only ears matter. In my experience, those that say only ears matter tend to be ignorant of how to interpret data or completely misinterpret what is in the specifications. Ears are one of the worst measuring devices ever, because the brain gets in the way. Sound I "enjoyed" 10-20 years ago, is most definitely not at the same enjoyment I have today. Measurements, on the other hand, won't change as long as the process is deterministic...i.e., not probing the proper place by luck.
Why don't you try and build something and see for yourself? To properly have an opinion on something, you need to understand it first; otherwise it's as the saying goes...opinions are like assholes.... There's no better way to understand than doing it. You will quickly learn what matters the most, and how things gradually improve with subtle changes.
Your questions are all fine and dandy, but what are your intentions? Is it to involve yourself in some silly flame war, and comment without any basis of experience, or are you sincerely going to try and understand, and experiment?
Another thing to note. Just because there are audible differences, it doesn't necessarily make one more preferable to another. Differences are just that, a notable change. It doesn't necessarily make one better than the other, that will be imprinted on you over time. Enjoyment of music doesn't necessarily relate to good design and good execution.
