http://www.tnt-audio.com/clinica/jitter1_e.html
this explains the issue with timing and waveform at the stage the signal is reconstructed to analogue.
edit: high end dacs wont even work with cheap cables. there too tight a tolerance.
I have a 10meter optical cable that works fine. try a cheap 1 meter and the dac spits out intermittent corrupt signals until it declares it an error and stops processing.
decent audio components have tight operational tolerance so these effects beneficial or not can be heard.
in a senario where the square wave is on an analogue carrier signal any spikes or ripple will offset the surmised data and move the intended values to reconstruct the sound wave.
theres no information in a digital recording regarding the clock rate it was captured at or on the consistency of its timing. so the dac responsible for reconstructing the given information relies on it being theoretically perfect.
probably not the case out of all but the finest recording studios
edit: the only computer capable of true real time generation of an analogue signal is the high end apple mac.
a PC with no budget concious approach would need some reconstruction of the OS and to employ GPU cards capable of tesla calculations.
the whole PC audio is still in its infancy. an I7 with full instruction set has the advantage of simultaneous memory addressing. so large banks of ram with small overall capacity in memory terms is the most cost effective step forward in the current technological state.
I have an aerospace company currently mulling over designing a emi cage for a usb audio card. in theory it will give the card optimal environmental conditions to operate in. the company that makes the card are keen to have a sample sent to them for approval yet they state from there knowledge base the card in question has all it needs to perform. yet there still open to the better understood knowledge base of the aerospace company.
much like the PC no one person can understand every subject that arises. people within specific fields contribute to a whole. so we experiment in an effort to push results forward.
we can debate for another 50 pages with our combined limited knowledge base but I doubt it will lead to nothing more than an agreement to difffer.
thats why theres companies throughout with different design philosophies.
if we provide links to the information we have taken on board then we can reach our own conclusions.
I take on board what you pose because I want to understand more