This stuff has a loooong history in the audiophool arena. There are good and real technical reasons why differences can be heard, but what does not follow is that if your gear produces these differences, and you can induce them, that you have a technically good system. Quite the reverse.
One point, this was made in a recent thread here, there is no interpolation occurring during error correction - the errors are completely corrected - which is why the DAW null test works. Very very occasionally a CD player will encounter an uncorrectable error - typically a few in an entire CD, so they are many minutes apart. These do get interpolated (typically) by the DAC. The S/PDIF subcodes flag the packet as bad.
What one did see in the loony end of audiophool components were broken designs. There is no shortage of insanely expensive digital components that have been incompetently designed. Often designers with some experience in analog electronics who think they can simply bolt on some digital bits. Such devices can have much greater sensitivity to jitter on the input than more mundane, but competently designed gear. Yet the fact that the DAC is so sensitive to jitter on the input is held up as a good thing - the DAC is so good that it can "resolve" the nuances of different sources and even the sort of cable used. Which is silly, at best.
And the cable? Sure. Cables can have a huge effect on jitter. If the gear they connect is badly designed. Poor termination is a good start. Poor noise rejection is another. Simply terminating into a resistor and a Schmidt trigger is guaranteed to give trouble, because most logic gates are not designed as line receivers and actually spit back energy when they transit levels.
Counter intuitively, a more lossy cable can be better, because it attenuates the energy ringing in the cable, and tends to open up the "eye pattern" seen at the receiver. Thus yielding lower jitter. Indeed, one of the mantras of good design is to have as wide a bandwidth as needed, but no more. Ultra fast logic causes more problems than it cures. Low jitter requires controlled rise times and a clear understanding of other aspects of digital design that are not found in undergraduate level textbooks.
So, I'm not at all surprised that differences can be heard. What one can't do is infer that this is due to some fantastic level of quality of the system. It is more likely to be the result of a very good backend (amp, speakers) revealing design flaws in the digital components. Flaws that can be exacerbated by messing with things like cable loss and cable length.
Swap out the DAC for a competently designed one (the usual offenders, Benchmark, Lavry, etc) and the effect goes away. Yet the audiophool complains that sound has lost a certain something. An inability to justly the insane cables being one of them. I'm also convinced that the final design tuning that occurs with these silly priced DACs partly involves tweaking to find a sweet spot where the distortion effects of jitter induced from the S/PDIF input (signal correlated) are fiddled to a point where they add a euphonic colour to the sound. This may involve simple blind fiddling with grounding and bypassing, with no understanding of why it is changing the sound, but simply providing a way of tuning it. The designer then trumpets the importance of designing by ear, and hints at mastery of deep and subtle arts, when in fact no such thing is true at all.
