I have found all types of audio cables to be very necessary to be optimized for the best sound quality of the system. Once the sound quality of the system is brought up to at least a good level (and the audiophile ear becomes a little trained in identifying these effects), then sonic differences due to the cable design and construction become obvious and very important. This is regardless of whether blind tests were done to verify the cable evaluations. Blind tests are not reliable in finding the subtle effects of conductor purity and crystal size, dielectric, conductor construction (solid core, Litz,, ribbon), cable topology, etc. The trained human ear-brain system is vastly more sensitive than conventional electronic instrumentation. The meter-reader mindset ignores this fact.
Concerning the question as to whether digital cable quality is unimportant beyond a $300 or so retail price: The answer is that digital cable quality is very important, despite there being error detection and correction designs with many digital cables. The explanation for this in my experience is that the sonic differences with digital cables are even more important than with analog cables, and (unfortunately) the benefits of very expensive designs are immediately apparent.
I think the reason for this sensitivity of digital cable sound characteristics to the quality of the design, is that much or most of the sonic distortion in digital cables is due to pulse timing issues, which are generally ignored in conventional mass-market digital audio cable design. This is where the error logic of the cable interfaces is looking for the presence or absence of pulses within allowable time windows. Accordingly, fine errors in pulse timing within the overall error margins are not corrected for. Exact timing of the data bit pulses is very important to cable sonics, but low cost digital cables don’t try to optimize this parameter.