I am currently reading about Digital-Analog-Converters in a magazine. In one section about connecting an iPod, it says that the sound via a USB cable such as the one from Audioquests Cinnamon is significantly smoother and more velvety compared to the official USB cable from Apple.

Also, a standard USB cable: In the description of Audioquest "Coffee", they are claiming that there is a huge difference. And they cost about 280€! Or their "Diamond" version costs about 550€! Here is some review, again claiming that there is a difference. Here is another review. And here another one.

How can that be? The signals which go through the cable are digital. How can there be any difference?

(Just for reference, I also asked the same question on Apple.SE with similar answers.)

  • 2
    "Sound coloring" due to USB cables is just not possible or plausible. Digitally encoded information (especially using error detection and correction schemes built into the codec) is either transmitted as a 100% faithful copy of the packet or it is dropped and retransmitted. These are con men preying on naive consumers.
    – horatio
    Nov 8, 2013 at 21:57
  • 2
    @horatio - in fairness, it looks like not all USB audio protocols have error correction. Some modes only have detection, but still, packets will either get there or not. You might get some lost packets that get skipped, but it won't otherwise color the sound since error detection would show that the packet was no good. Technical Details
    – AJ Henderson
    Nov 9, 2013 at 5:20
  • 1
    There is one single word that brings 550€ USB cables and the like perfectly to a point: Audiophoolery.
    – leftaroundabout
    Nov 9, 2013 at 10:53
  • 1
    @AJHenderson: My understanding is that USB protocol itself has CRC error checking in all modes, but what happens upon detection of error depends on the modes chosen by driver software. For typical audio, I think the error check might fail, but no retransmission occurs. (I think this supports your comment)
    – horatio
    Nov 12, 2013 at 15:29
  • @horatio - yeah, that's exactly it. It won't use a bad packet, but it may miss one in low latency modes that don't retransmit. So yes, we seem to be in agreement.
    – AJ Henderson
    Nov 12, 2013 at 15:43

3 Answers 3


A good digital signal should either get there or not. Minor differences in cable quality are not going to have any impact at all other than latency. Latency is simply the delay in how long it takes to get to point a to point b and won't impact the sound quality significantly. USB is a digital signal, so a "premium" USB cable is a load of crap. There is no real difference between a well made $2 USB cable and a $50 premium one. You get the exact same result.

If the cable is too long or improperly shielded and you use a transfer mode that doesn't have packet retry, it's possible that cable induced jitter (ones being detected as zeros and vice versa) could result in packet loss which would impact the sound, but you should be able to detect if that is happening and a buffering format that can retry packets will prevent any impact from this, however that increases latency. If you absolutely need super low latency, then it might make a difference, but otherwise, not so much.

For analog signals on the other hand, it's a very different situation, though the "premium" cables are rarely, if ever, worth the cost, even in the professional market.

As for the reviews, they work better in those cases because they are paid by the manufacturer from the boat loads of profit that they make when a sucker buys a cable that cost them $.50 to make for $500. The review just reads like something someone in a poorly thought out marketing department wrote. It's an ad not a review. (Update: One of the reviews appeared like it could potentially be legit, but they were also talking about using it in ultra-low latency environments and I'm still not sure that I buy that the reviewer knew what he was listening for. It wouldn't be the first time that placebo effects impacted the perceived listening experience. Until they can identify which cable is which in a blind test, I don't buy in to it, certainly not for sound quality, maybe for latency. (Though even that I kind of doubt.)


I find such arguments generally invalid. The whole point of digital is that the signal is transmitted as-is, without loss or coloration. Now, one might make an argument from the standpoint of interference (e.g., if you're running your USB cable alongside a thinly insulated power cable with a 10kw load), but such arguments are silly as well.

If someone out there can point me to articles which give actual scientific proof as to why cables would in any way alter the transmitted digital content (i.e., its frequency or content), I'd be very interested in taking a look.


It is theoretically possible for a USB cable to impact the audio quality, however for any practical application it is extremely unlikely, to the point of impossible.

For instance, a really cheap, badly made cable will potentially drop out altogether as the materials and mechanical quality can be awful and conductors completely come away from the plug intermittently, but if the cable wasn't impossibly cheap, this is very unlikely.

For a bus-powered device, if the cable shielding isn't up to scratch and the amplifier/DAC/whatever has inadequate power supply filtering, noise coupling onto VBUS will cause an issue, although often the connections inside a PC are the culprit for this sort of noise (ie buzzing when a USB stick is accessed on a neighboring port).

Finally, it is possible that the geometry and materials that make up these cables change their transmission line properties (propagation velocity etc) to such a degree that communication over a longer distance than the 5 metres in the USB specification would be possible, although above 5m USB is the wrong system to use anyway.

As for these reviews saying the sound quality gets better, I would love to see the same cables connected to an oscilloscope and a signal generator to see if the signals see any meaningful change, I suspect the answer would be no.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.