Best USB cable to use between computer and dac?

[4est]

If you ever disclose any factual content as to the cause of said phenomena, I might address it. As to your motivations, I didn't comment on them, but stated that your attitude is a patronizing one in the context of this thread. I could most likely guess why, but it would not be polite nor appropriate here.

 

[Julf]

"If you ever disclose any factual content as to the cause of said phenomena, I might address it."

 

OK, I made clear statements about what would be expected based on laws of physics, electrical engineering and communications theory. Instead of addressing them, you decided to lecture me about my attitude. Sure, feel free. I just don't call that a discussion.

 

As to more factual content about "said phenomena", maybe you might like to address the following page that Jim Lesurf has compiled: Cable Sounds and uk.rec.audio?

 

[4est]

As I stated, most everyone here understands that there really should NOT be differences in digital signal transmission unless there are errors. That is the basis of this entire discussion.

 

[Julf]

"most everyone here understands that there really should NOT be differences in digital signal transmission unless there are errors."

 

Great! It was just my impression that not everybody shared that opinion.

 

OK, so if we start with the assumption that USB cables cause real, audible differences (instead of ascribing the observations to psychoacoustic effects), there are a number of theories and experiments that could help isolate the causes.

 

One theory is that the cable conducts ground loop noise and EMF into the DAC circuitry. This would be easily tested for by using an USB isolator. Is anyone aware of any listening tests where the cables have been tried both with and without an isolator?

 

Another is that cable inductance/capacitance causes attenuation that actually affects the transmission of the data. If so, that would show up when comparing different lengths of the same cable, as the cable inductance and capacitance increase linearly with cable length.

 

USB is also special in that both the signal and power are carried close to each other in the same cable. There has been reports that separating out the power feed conductors from the cable has produced improvements in sound quality. It would be interesting if listening tests show differences between the same cable connected to a bus-powered and connected to a self-powered DAC.

 

Any other major theories?

 

 

[Paul R]

Assuming people are not crazy is liable to get a lot of people who are interested working with you Julf.

 

It's nice to see you guys actually communicating instead of poking sharp objects at each other!

 

-Paul

 

 

[Jud]

Any other major theories?

 

Depends which sort of USB digital signal transmission we are talking about. With async USB, I'm assuming any jitter due to cable characteristics re timing of the signal will not show up in the sound. With adaptive USB, however, timing characteristics of transmissions through the cable not bad enough to cause actual dropouts might show up as jitter.

 

With async USB, barrows has mentioned that the ground loop noise and EMI you referred to might affect the DAC's clocking and thus result in jitter.

 

[4est]

I truly feel that the average poster here is at least somewhat cognizant of this. With that in mind, I feel the safer assumption would be to give us the benefit of the doubt in that regard, and focus on the engineering as you have described. I apologize for being offensive at times, as I have felt talked "down to" on many occasion as of late. Many of us come here to share our experiences in hopes to get to the bottom of things such as this.

 

Frankly, it makes me shake my head in disbelief, and I have often considered my equipment as "broken" due to these phenomena. I do not know the answers, but I know what I hear. I am open to all possibilities, including my system being "broken". The only thing that I know for sure are the net effects that it has on my stereo, as it is repeatable and discernible by others here and elsewhere. Lastly, just to be clear, my hopes and expectations were not met.

 

[wgscott]

Was the generic USB printer cable the better one? (You never said which one was better.)

 

[Jud]

...with apologies to Mr. Zimmerman.

 

I am open to all possibilities, including my system being "broken".

 

PeterSt has been critical of cable "tweaking," and has said (at least as I paraphrase what I think he meant) that in a proper system, cable and various other sorts of changes should not matter. I interpret this as meaning, for example, that if ground loop noise and EMI transmitted through the USB cable can change the sound quality, one really ought to have measures in place elsewhere in the system to avoid creating such noise, or to avoid the noise creating problems once it has reached the DAC input.

 

Since I do hear changes in sound when I change USB cables, and I don't have anything like an almost perfectly low noise source or an almost perfectly noise-immune DAC, I'm very willing to think that these sound changes are a reflection of my system being, to use your word, "broken" to some extent.

 

[Julf]

"With adaptive USB, however, timing characteristics of transmissions through the cable not bad enough to cause actual dropouts might show up as jitter."

 

Indeed. So we should see clear differences between the cables when connecting async-capable DACs vs. adaptive ones.

 

"With async USB, barrows has mentioned that the ground loop noise and EMI you referred to might affect the DAC's clocking and thus result in jitter."

 

Right. That's where testing with an USB isolator should show differences.

 

Jitter is also reasonably easy to measure.

 

[elcorso]

If you are on a Mac, you need a Canon USB printer cable.

 

But, if you are on a Windows PC it's better an HP printer cable, or a Lexmark if you love deep bass.

 

Roch

 

[Julf]

"I apologize for being offensive at times, as I have felt talked "down to" on many occasion as of late."

 

Likewise!

 

Let's try to sort out this one in a constructive manner! I know I will occasionally fall back into electrical engineer mode, so apologies in advance...

 

[elcorso]

"...With async USB, barrows has mentioned that the ground loop noise and EMI you referred to might affect the DAC's clocking and thus result in jitter..."

 

BTW, I was reading form some DAC advertising as one his main points:

 

• Digital Ground Isolation - Noise Reduction

 

I don't completely understand (at least in digital) but believe this is one of the big clues regarding USB and FireWire cables, and I implemented it the best I can with good results (to my ears).

 

Roch

 

[The Computer Audiophile]

"Jitter is also reasonably easy to measure."

 

Hi Julf - I'm thrilled to see you and others working constructively in this thread. It makes the hobby much more enjoyable :~)

 

Regarding your statement I quoted above that "Jitter is also reasonably easy to measure.", I disagree 100%. Without question every audio engineer I've ever talked to has suggested jitter is extremely hard to measure. On a very basic level it may be simple to obtain a measurement of jitter, but when talking about thoroughly digging into jitter it's another story. Most engineers even suggest it's extremely difficult to figure out what and how to measure it.

 

Demian Martin from Auraliti is an extremely knowledgeable person on this topic. I cross my fingers that he can add a professional opinion as he has been doing this at the highest levels for decades.

 

 

 

[Julf]

OK, let's agree to disagree. Measuring jitter is a fairly straightforward and routine thing to do. All you need is a suitable test signal (the J-test is the most well-known) and a spectrum analyzer. Play the test signal, and look at the output - any extra intermodulation sidebands are caused by jitter, and directly proportional to the amount of jitter.

 

One of the classic references is Jitter: Specification and assesment in digital audio equipment - Julian Dunn, AES 1992 - still useful after 20 years.

 

[4est]

That paper, if I read it correctly, is looking at phase noise/jitter in the 20 nano-second range, not the pico-second range. It has been shown (Gordon at Wavelength IIRC) that it can be audible into the pico-seconds. MSB recently introduced a clock that is rated in the femto-second range FWIW.

 

[Julf]

Sure, but the method is the same even if you look at the picosecond range. And yes, one of the most quoted papers, High Resolution Audio DACs (Ivar Løkken, 2005) states:

 

"We can see that the audibility threshold decreases from 500ns at low frequencies to as little as 20ps at 20kHz. Especially when using formats or converters with high sample-rate this will be a major issue."

 

 

 

[Jud]

You can ask Gordon what the cost is of the measuring equipment he mentions in the following comment:

 

http://www.computeraudiophile.com/content/Some-additional-thoughts-about-measuring-jitter#comment-14084

 

[4est]

I have heard there are issues with sub 10ps jitter levels, and along the lines of Jud's post, that measuring jitter (and the types thereof) is not as straight forward as you are making is seem. I do not know the answer myself, and so this is purely hearsay. Either you are mistaken/over simplifying or some of the (better) manufactures are lying.

 

[Julf]

Well, yes, when you get to sub-10ps jitter you do need very precise measuring equipment. The measuring method doesn't change, but the required precision definitely becomes a limiting factor.

 

Also, just to clarify, I am talking about measuring the *amount* of jitter (as that is what we are interested in in the context of cables). If you actually want to analyze the jitter mechanisms, and determine thinks like the frequency of the jitter, you need more advanced measurement systems.

 

So, let me qualify/restate my statement - for jitter at levels found in most practical systems, measuring the level of jitter is fairly straightforward to measure, and only requires a J-test signal and a spectrum analyzer.

 

[kaka]

I've watched this thread with a sense of curiosity. I'm running firewire dacs now, but I was in the USB camp. The posts that interest me are of the "I tried X,Y and Z and they sounded the same" ilk.

 

I have the USB cables that came with Red Wine Audio and Empirical Audio gear, plus a couple of the specialist Locus cables. That was probably before the Wireworlds/Audioquests etc had come into existence.

 

Playing via an Offramp 3 (which was not async) all those cables sounded different, and the Locus cables outperformed the basic ones by quite some margin. I've still got three of them, so if I find myself with async USB gear I look forward to finding out if the cable still makes a difference.

 

[Mark Powell]

There is a school of thought that thinks that the better your DAC is the less sensitive to the cable it should be.

 

 

 

[Talk2Me]

Maybe we should measure the pro's?

 

[Talk2Me]

You are such a tease,kaka.

 

[pilgrum]

This is a tough subject but theoretically there are better conductors out there than what a freebie usb cable has to offer. I really don't notice an audible difference till you start working with hi % silver as a conductor and am a skeptic about how that relates to data other than it traveling more efficiently. It is important to keep the USB cable as short as possible due to the nature of how it transfers electricity the longer it is the slower it is. Grab an audioquest usb cable for like 20-50 bucks and see what you think. I personally use a Dcoax i made out of an audioquest silver series video cable and terminated BNC/rca on it since my sound card and dac have different connections for Coax spdif. Did I notice a difference between testing a $300 AQ custom cable and generic rg6, nope but it looks much cooler. Now as far as speaker wire and component interconnects that is where I have heard substantial improvements with using silver but any copper variants past psc+ no so entry level rca cables seem to be the most cost effective I am a fan of AQ cause of their prices.