Originally Posted by judmarc
"Bits is bits. They are either there '1' or absent '0' -- if the existing cable does not 'lose' bits, then no amount of cable improvement will make any sonic difference. Period." Without wanting to get into this time-honored disagreement yet again, let me gently suggest a couple of potential considerations. I am no expert, so what I'm suggesting could just be trash and ignorable, in which case, go ahead and ignore it. - Thought experiment: No music playing through the cable. Lots of electrical noise, though. Can you hear it? If we're not talking about an optical cable, then sure, it's an electrical wire. The fact that someone slaps the label "USB" on it doesn't make it suddenly immune to electrical noise. So if one cable deals with noise better than another, there's a potential difference. - Reading through a white paper by the former ESS SABRE DAC developers (now Resonessence), I saw a paragraph about how it is critical for the DAC chip to have a constant voltage level for comparison when it's evaluating whether a bit is 1 or 0 and the time when changeover occurs. So electrical noise in the power used by the DAC chip might potentially result in jitter. - Electrical noise into the DAC's clock(s) might also conceivably cause jitter. A link to some pictures and text (not by a native English speaker) that might be of interest in this connection: http://www.phasure.com/index.php?topic=2557.msg26459#msg26459
I DO NOT disagree with judmarc's caveat. Yes, noise and other interference CAN make a difference. I'm sure there is some threshold where suddenly - and it WILL be suddenly - the cable causes errors. Or a super-cheap or damaged cable is simply not up to snuff. Whether the noise causes jitter or loss, the plot of where the loss/jitter occurs will be a step, not some nice curve on a graph.
The data wires in USB are twisted pair - this eliminates most noise. Compare to speed and reach of Gigabit Ethernet where the spec is 100 meters on twisted pair. USB also includes grounded shielding whereas GigE does not. Some USB cables have ferrite cores on each end, this will also help.
One interesting factoid is that USB2.0 is limited to 5 meters simply because the spec requires a 1.5 micro-second response (turn-around time) to a command, else it considers the command lost.
There is a lot of good info on Wikipedia (take with grains of salt): https://en.wikipedia.org/wiki/Universal_Serial_Bus
This site has a fairly coherent description of async USB: http://audiophilleo.com/definitions.aspx?Asynchronous%20USB
The bottom line is that there are LOTS of other places that are more likely to cause audio degradation than the USB cable!
That being said, if using Windoze there is a tool: http://www.usblyzer.com/ that could help. It has a 33 day free trial.