2
\$\begingroup\$

I was wondering this for quite a while. Usb cables are purely digital signal so in theory no one should care about its quality if it does the job well and transfer the signal to the end (ones and zeroes). But then again... Why do some cables have double shielding for interference? And when we are talking about recording and audio interfaces I started thinking that perhaps interference could bring noise to the recorded signal or some 'impurities'? I know that gold coating is just a cheap marketing trick and practically is not important at all. Cable resistance - this I do not know (on digital cables).

enter image description here

So how (if it does) a good quality, double shielded, cable "improves" the sound?

Is there ANY advantage of using it over a cheap, single shielded (or not even shielded at all..) cable when we are talking about high quality DAC, Audio interfaces and other devices which transfer signal that is video, audio or anything familiar. Should I run and purchase a good quality usb when I'm recording? Or good quality cables are only good for reliability?

Haven't solved the mystery myself. I know that you guys are always helpful and will solve it :)

P.S. I am asking specifically for audio interfaces or DAC. I know that if you'd pay 50$ for an iPhone cable it would be simply dumb as it does not matter expensive or rather cheap. Audio interfaces and specifically audio recording is what I am not sure about.. (and didn't quite found the information I was looking for on the internet or forums)

\$\endgroup\$
  • 4
    \$\begingroup\$ Depends on if you are an audiophool or not... \$\endgroup\$ – PlasmaHH Jul 17 '16 at 19:41
  • \$\begingroup\$ It's just a "hoax". USB signals are digital. \$\endgroup\$ – bitshift Jul 17 '16 at 19:45
  • \$\begingroup\$ bitshift, you haven't fully read my question. I am aware that USB signal is digital: "Usb cables are purely digital signal so in theory no one should care about its quality if it does the job well and transfer the signal to the end (ones and zeroes). But then again... Why do some cables have double shielding for interference? And when we are talking about recording and audio interfaces I started thinking that perhaps interference could bring noise to the recorded signal or some 'impurities'?" \$\endgroup\$ – Giedrius Jul 17 '16 at 21:14
  • 1
    \$\begingroup\$ Your on the right track, the noise can often be generated by the USB cable, long wires act like antennas and will both radiate and receive all manner of RF crud flowing through the air. Shielding provides a low impedance path to ground that more or less channels this unwanted energy away (kinda like a lightning rod). However, if your gear is properly shielded & grounded inside (and the analog audio input cables have good shields) then there shouldn't be a problem. As an engineer we say: If you're unsure about something, test it for yourself. Any coupled noise won't harm the gear. \$\endgroup\$ – Sam Jul 17 '16 at 23:05
  • \$\begingroup\$ Your Audio before the nice cable 01110101010010101010 Your audio after the expensive cable 01110101010010101010 \$\endgroup\$ – user178641 Feb 19 '18 at 15:16
10
\$\begingroup\$

It's worth having a USB cable with ferrite 'stoppers' on it, to attenuate conducted emissions along the cable. This will reduce the chance of hash from the switch mode power supply and other PC generated interference 'getting into' the audio circuits of the ADC that's doing the recording. With an electrically noisy PC, and a cheap ADC, a cable without stoppers might produce breakthrough interference on audio. However, part of the cost of a good ADC should be high immunity to conducted interference.

The ferrite stoppers add a few pence (cents) to the cost of a cable, and most cables I've seen have them (the cylindrical lump close to the connector at one or both ends of the cable), only the very cheapest cables do not.

Once a cable has these stoppers, all cables either work, or are broken, there are no differences in data quality. If it can transfer a file, then it can transfer your highest quality recording. Golden ears cannot tell the difference between two identical files.

I wish I could dream up snake oil like expensive 'audio grade' USB cables and sell them for a fortune to audiophools, but I'm just an honest engineer!

\$\endgroup\$
  • 2
    \$\begingroup\$ Data transfers will attempt to retransmit bad packets. Live audio will not. Usb audio is not like a data transfer. \$\endgroup\$ – Passerby Jul 17 '16 at 20:32
  • 2
    \$\begingroup\$ Then again; data transfers are done in chunks of multi-dozen audio samples. So any faulty cable will not be a small nuance in audio quality; rather it will be a huge stutter. Noise on the VBUS of USB is the most likely culprit for audible noise. \$\endgroup\$ – Hans Jul 19 '16 at 11:31
6
\$\begingroup\$

As USB is a digital signal, no, interference will not cause the audio to change. But USB audio is normally delivered via isochronous data transfer (think UDP), a best effort non-error correction protocol. USB uses Manchester encoding on differential twisted paired data lines to reduce the effect of any interference, as well as the shielding. USB does have error checking (CRC) for transmissions, but as the audio is real time transfers, any corrupted packet is simply dropped/ignored.

You would hear audio drops if enough interference happens on the line, but no static or similar analog noise. Normal interference shouldn't affect it much if at all.

A sufficiently sub-par or damage cable may drop too many packets, or cause the line to drop to a lower speed, to make it usable. The same cable for bulk transfers, like copying a file from a hard drive, may result in corrupt transfers or transfers that take much longer than normal.

\$\endgroup\$
  • \$\begingroup\$ So if I understood correctly a good quality cable might reduce the audio latency (input-PC-live.output) over a really cheap one? \$\endgroup\$ – Giedrius Jul 17 '16 at 20:06
  • 1
    \$\begingroup\$ No, any cable that comply with the USB norm would behave the same. But ultra cheap counterfeited cables may not comply and create issues like described. Also low cost cable may wear much faster than high end cables. \$\endgroup\$ – Blup1980 Jul 17 '16 at 20:12
  • \$\begingroup\$ If we're talking about a bus-powered device, though, could a sub-par cable change the power delivery enough to cause artifacts? I know that the bus that is actually supplying the power would be a bigger deal here, but could that have any effect on, say, how the preamps or ADC for a mic function, or how steadily Phantom Power is supplied? \$\endgroup\$ – sbergeron Jul 7 '17 at 15:28
  • \$\begingroup\$ Only if the device pulls more than 500mA or has very fast incharge. \$\endgroup\$ – Passerby Jul 7 '17 at 16:27
1
\$\begingroup\$

Yes it can. Even if you have lets say isolation barrier between USB procesor and DAC chip (inside the "DAC box"), EMI modulated on USB data transfer (and thus heavilly corolated) will get thru and onto I2S data lines to the DAC modulating his silicon substrate with jitter and affecting the audio quality. This is true for each UAC standart transfer (asynchronnous, isynchronnous and adaptive).

On the other hand...USB audio cable is the last thing you want to mess about and its a "voodoo" in lot of senses. There is technical (measurable) as well as hearable difference (both of them are very small), but you are better saving those USD500 for better DAC, then spending it on a cable. Double shielding isn't really going to help you much. It has some technical backround, but doesnt solve the problem and its mostly to sell cables for ridiculous prices.

Thing that usually helps is a USB hub between PC and DAC which is powered from external power supply. It repackes the USB frames resulting in lower jitter and has much cleaner supply for the DAC.

\$\endgroup\$
  • \$\begingroup\$ Thank you :) I really don't plan on spending more than 20$ on a cable. I'm probably comparing 2$ cable and a 10$ cable (or to put it otherwise - single shielded vs double shielded) \$\endgroup\$ – Giedrius Jul 17 '16 at 20:54
  • 1
    \$\begingroup\$ I would like to see some support for that statement about jitter... like written by other people, unless it's a VERY bad cable, jitter is well within tolerance and therefore not affecting the result. Jitter it's like the rest: either too much, or not affecting at all. There is no linearity with digital devices. \$\endgroup\$ – FarO Jul 17 '16 at 21:06
  • \$\begingroup\$ Olaf: That is not true at all...its not about the data correctnes, that will alsways be OK, its about when DAC is converting audio signal to analog domain. If you have jitter there, its always converting at different time interval then before and you get wrong analog value - distortion. \$\endgroup\$ – Jaroslav Dohnal Jul 18 '16 at 5:31
  • 2
    \$\begingroup\$ Except that all sensible USB systems have a small receive buffer in the controller, usually one frame, so their actual DAC subsystem is driven from an internal clock and not affected by the precise timing of the arriving frames. e.g. PCM2704 has a 1ms buffer specified in the datasheet. \$\endgroup\$ – pjc50 Jul 19 '16 at 11:16
  • 1
    \$\begingroup\$ That is of course true (I have written my own USB audio class firmware, so I know some stuff), but the jitter will get thru anyway...you always have more digital pins than grounds on a given device...which means they will modulate each other. e.g the jittery input will modulate "clean" output. Only thing that can solve this is a reclock with multiple single edge triggered flip flops. \$\endgroup\$ – Jaroslav Dohnal Jul 19 '16 at 21:46
0
\$\begingroup\$

No. That shielding is ineffective against conducted EMI from the computer anyway which is the major source of interference. Unless it's in the audio spectrum, you won't be able to hear nor measure it though.

Shielding in general is a way to sell audio-whatever at a higher price. Think balanced XLR with transformer for ground-lift in one end. Now change the cable and listen again. Did the shielding make the audio warmer, more emphasis in the midrange or clarity that just wasn't there for the hi hats?

As for the USB part of it, the specification requires sheilded cables but this has nothing to do with audio.

\$\endgroup\$
  • \$\begingroup\$ I'm afraid that you didn't quite catch (or I did not explained until the edit briefly enough) what I was looking for. I am interested specifically for audio recording. I know there are expensive cables that are made specifically for audio purposes. Again - wondering what's all with that. Just a hoax? \$\endgroup\$ – Giedrius Jul 17 '16 at 19:38
  • \$\begingroup\$ Yes, it's a hoax. \$\endgroup\$ – winny Jul 17 '16 at 20:14
0
\$\begingroup\$

Besides EMI issues as explained in the other answers, there is also jitter, but it is rather subtle.

Isochronous audio transfers allow the DAC/ADC master clock to be either the host, or the device. This depends on the chosen mode which is a decision made when implementing the device:

  • synchronous, where the stream’s clock source is linked to the USB’s clock, such as through the use of USB frames;
  • adaptive, where the stream’s clock source varies according to the buffer requirements of the stream.

In these two cases, the device has to recover the master clock from USB frame timings or data rate. This is usually done via a PLL, and its implementation may be good, or not. The clock created by the PC is usually very jittery, so the ability of the PLL to create a clean clock is important.

  • asynchronous, where the stream uses a clock independent from the USB clock source;

In this case the device runs its own clock (presumably clean) and sends feedback packets to the PC to keep it in sync. This is the preferred mode for high quality.

In both cases, the influence of USB cable quality on clock jitter seems very dubious. Implementation issues like PLL or local oscillator quality, or noise leaking from the CPU/USB chip into sensitive parts would matter much more. I would not put a lot of trust in claims of USB cables making a huge difference in jitter.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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