Mega-post (Click to show)
This is often true - qualities such as soundstage and instrument separation are quite difficult to pin down. This could possibly be because many cables are all but indistinguishable - in my experience differences between say one copper USB cable and another are almost impossibly to cleary tell. Much like with an SPDIF cable once you have a constant impedance (90 ohms for USB) changes are difficult to discern and to be honest from a technical standpoint this is what you would expect. After having constant impedance, some cables have silver plated conductors which apparently can improve the signal waveform integrity by creating a more constant impedancy/frequency curve. This probably depends a lot on the wire gauge, plate thickness etc and to be honest could be complete nonsense unless the cable designer shows some numbers to back this claim up. The last thing that in theory can affect USB cable performance is cable length - where a shorter cable of the same construction should give better jitter performance as predicted by the USB standard paper. The last factor is the succeptibility of the USB audio interface itself to jitter, along with other factors which may introduce jitter or noise into either the D/D converter or directly into the DAC.
You are quite right that soundstage and instrument separation are qualities which are rather easy to get confused about, and also in my experience many digital cables I have tested are very difficult to tell apart - I'm pretty sure I would fail a blind test with say 75 ohm SPDIF cables or USB cables of the same length and conductor type. Personally I have heard more differences between USB ports and galvanic isolators than between different copper USB cables. BUT I am personally quite certain I can discern between a copper USB cable and the silver plated copper cable I am using, and equally there is evidence that in theory different USB cables have different levels of jitter. BUT the effects of jitter on D/D and D/A converters is fairly controversial, in no small part because there are innumerable variables in terms of hardware, but mostly because it is relatively simple to demonstrate that relatively large magnitudes of jitter (beyond that of any decent equipment) are not audible.
As you can see I personally am in quite a precarious position - I am forced to admit that on an priorised-knowledge level one cannot clearly predict that jitter will be audible. However, based on experience I can say that elements such as USB cables have lead me to experience and articulate clear differences between USB and computer side hardware and software changes which I cannot readily explain and yet which I would not put down to cognitive bias. As I already said I like to think I am willing to concede when I cannot clearly hear differences or when I hear negative differences. When this happens I sell equipment very quickly as I would rather spend the money elsewhere. Life would be much easier if I were an electrical or audio engineer researcher and had access to lab equipment to measure either electrical waveforms or jitter or even distortion levels at the output of the DAC. Unfortunately most Audiophile tweaks are very short on technical insight and reliant upon anecdotal evidence to measure performance - and while I do not personally consider myself an objectivist - the points mentioned in this post, along with the preceding 45 pages of discussion should act as a warning not to spend too much money on digital cables.
EDIT: This is probably not that important but I feel I should mention that personally I prefer to focus on other factors such as bass bloom/tightness, midrange smoothness or grain, and imaging and focus (ie how clear the outline of an instrument can be discerned) along with upper midrange and treble fatigue. These qualities are again not foolproof, but I personally find these easier to discern that soundstange and "air around instruments". Best thing to do I would say is once you have selected a DAC -see if you can borrow a fairly short silver plated copper USB cables to compare against the generic USB cable - If you cant hear a clear difference then obviously it is not worth buying such a cable.
I like your way of thinking, you have your personal impressions, but you don't give them more weight than what they deserve.
About the jitter, when you said that the ammount of jitter wasn't audible, it made it seem like jitter directly correlates with distorted music (analog). I was under the impression that up to a certain limit of jitter introduced to a USB signal, it would still get correctly decoded and so you still get bit-perfect stream. Kind of like how you can take a perfect drawing of a face and re-draw it worse each time, until at a certain point people can't tell what it was originally anymore.