Best USB cable to use between computer and dac?

[Mark Powell]

Were originally designed(discuss) to...

 

Make more horses.

 

[Julf]

"If you are not sure about length affecting digital cables, you should!"

 

I was trying to avoid sounding too aggressive. I certainly don't think, based on my knowledge of transmission theory, that the length of the USB cable, as long as it is within standard specifications, affects the function of the cable.

 

"Regarding cable 'direction' I talk about why some cable manufacturers speculate about. The best theory to me (but I don't apply it) is that the shielding is better truncated at the beginning of the cable."

 

Earlier in the thread I actually pointed that out. To rehash, this is what I wrote: 'Here is a great powerpoint from part of a tutorial workshop at the 2005 IEEE-EMC Symposium. that "includes Bill Whitlock's groundbreaking work on balanced interfaces, and shows why when a cable shield is connected only at one end to prevent shield current, the connection point should always be at the sending end."'

 

"But the electrons they have direction: from ground. Is like when you see a thunder getting the earth, but is the contrary, the earth getting the charge from the charged cloud. Then an optical effect."

 

As you say, it is an optical, not electrical effect. And if you re-read what I wrote, I stated that the reason the electrons don't have a direction is because the signal is AC. A lightning strike is DC.

 

"If you study electronic circuits you must know than on balanced configurations you will have two references to ground, one from the balanced connection, and other from ground pin from your AC feed. That means more ground loops, then more noise. And, some frequency increase in about 300 Hz. But I have some friends who like noise also."

 

You can have as many references to ground as you like. I think what you wanted to say was "connections to ground". And in a properly wired balanced system signal ground should not be connected to "earth" ground.

 

"Could I suggest you to remove your hearing protectors (I'm sorry but I saw it from your photo)"

 

Could I suggest you stop making silly, erroneous comments that have nothing to do with the discussion? The thing I am wearing in the picture is actually a headset, not hearing protectors.

 

[Mark Powell]

I've tried four. Two DACs, both with and without an S/PDIF converter.

 

No audible difference.

 

[Julf]

Unfortunately I am still traveling, and don't have access to my reference library, so I will just use Wikipedia, that actually has some pretty good stuff on this.

 

I suggest you start here - to quote, "While digital transmissions are also degraded, slight variations do not matter since they are ignored when the signal is received. With an analog signal, variances cannot be distinguished from the signal and so provide a kind of distortion. In a digital signal, similar variances will not matter, as any signal close enough to a particular value will be interpreted as that value."

 

Also suggest a read up on Quantization and Eye Pattern.

 

If you disagree, I think it would be more than appropriate for you to also provide references supporting your view.

 

 

[elcorso]

I'm sorry, still at the bathroom... too much to trough away!

 

BTW, headset or hearing protectors is the same, you don't listen to real music, to nobody else neither, only to yourself. Then I'll rename the artifacts you use "Echo Headphones"

 

'Reference to ground' is the correct term, I don't to use it because not everybody is an electronic engineer (or DIY) in this forum. Maybe you would need this 'reference' also, very suggested for a good landing...

 

But if Ex-lax® doesn't work I'll sue the company: Relief Guaranteed Every Time.

 

Back to the bathroom again...

 

Roch

 

[Paul R]

I do not accept Wikipedia as an authoritative source. You can just as easily find references elsewhere that dispute those findings.

 

Your assertion, as I understand it, is that since the correct values are received in a digital transmission despite minor analog distortion, there can be no effect on the reproduced sound of that signal unless an error occurs that is of sufficient magnitude to prevent reception of the signal.

 

I am not agreeing or disagreeing with you. I am simply asking you to provide authoritative references for your point of view, and Wikipedia "ain't one of those." I assume this should be easy, no? It certainly can't be an more difficult that proving that two USB cables sound different, no?

 

Whether or not I agree or disagree should not have any bearing on your assertion or opinion, nor am I attempting to prove the opposite.

 

-Paul

 

 

 

[Julf]

"I do not accept Wikipedia as an authoritative source."

 

Ah. OK. Before I go through any more effort, could you please indicate what kind of sources you would accept?

 

"Your assertion, as I understand it, is that since the correct values are received in a digital transmission despite minor analog distortion, there can be no effect on the reproduced sound of that signal unless an error occurs that is of sufficient magnitude to prevent reception of the signal."

 

Precisely. And that view is supported by any basic textbook of digital transmission systems, as well as the sources quoted in the Wikipedia articles (and that should be easy for you to check):

 

Tocci, R. 2006. Digital Systems: Principles and Applications (10th Edition). Prentice Hall. ISBN 0131725793

 

Jayant, Nikil S.; Noll, Peter (1984), Digital Coding of Waveforms: Principles and Applications to Speech and Video, Prentice–Hall, ISBN 9780132119139

 

Gregg, W. David (1977), Analog & Digital Communication, John Wiley, ISBN 9780471326618

 

Stein, Seymour; Jones, J. Jay (1967), Modern Communication Principles, McGraw–Hill, ISBN 9780070610033

 

 

[tarquineous]

Mark, which ones did you try ?

 

Were any of those cables the same ones mentioned by other members, such as Wireworld, Audioquest Carbon, or Silver Sonic ?

 

I mention these three, because they were favored by the person doing the testing/comparisons.

 

[tarquineous]

Julf, For the practical side of the basics, how do you determine these things in your own system ? That is: "Errors of sufficient magnitude" and "No effect on reproduced sound" ?

 

In other words, besides listening tests, do you use bit rate check, oscilloscope, distortion analyzer on the DAC output, or other equipment ?

 

[Paul R]

Digital Systems: Principles and Applications (10th Edition)

 

I have the 11th edition, though in paper, and with just a quick scan, I did not find anything that directly backs up the assertion I restated in the previous message. At least, not specifically in the case of audio, and directly specifically in the case of USB transmissions. It's great if you want a reference on flip flops though. (Doesn't mean it isn't in there, just I could not readily find it in 20 minutes.)

 

The other references are a tad bit old, don't you think? CA in 1967 involved AM radios and line printers.

 

I'm flexible on references- Ron Tocci and Greg Moss pretty much known their stuff.

It doesn't have to be peer reviewed necessarily. It can even be from industry standards.

 

In general, I am looking for what references you feel make your assertion unassailable. For instance, the USB specs are clear that "An essential issue in audio is synchronization of the data streams. Indeed, the smallest artifacts are easily detected by the human ear. "

 

To me this introduces enough questions to ask a few questions. What exactly are these "artifacts" and how precisely are they produced? Further, since these are synchronization artifacts, do they always include data loss? What studies have been done to show this?

 

USB 2.0 was signed off on by some very major, and very smart players who know their stuff.

 

Here's the Management Overview, reproduced below. Emphasis is mine.

 

Universal Serial Bus Device Class Definition for Audio Devices

Release 2.0 Release 2.0

May 31, 2006 May 31, 2006 1

 

2 Management Overview

 

The USB is very well suited for transport of audio ranging from low fidelity voice connections to high quality, multi-channel audio streams. The USB has become a ubiquitous connector on modern PC’s and is well-understood by most consumers today. As such, it has become the connector of choice for many peripherals and is indeed the simplest and most pervasive digital audio connector available today. With the advent of the High Speed USB, consumers can count on this medium to meet all of their audio needs today and into the future. Many applications from communications, to entertainment, to music recording and playback, can take advantage of audio features of the USB.

 

In principle, a versatile bus specification like the USB provides many ways to propagate and/or control digital audio. For the industry, however, it is very important that audio transport mechanisms be well defined and standardized on the USB. Only in this way can interoperability be guaranteed among the many possible audio devices on the USB. Standardized audio transport mechanisms also help to keep software drivers as generic as possible. The Audio Device Class described in this document satisfies those requirements. It is written and revised by experts in the audio field. Other device classes that address audio in some way should refer to this document for their audio interface specification.

 

An essential issue in audio is synchronization of the data streams. Indeed, the smallest artifacts are easily detected by the human ear. Therefore, a robust synchronization scheme on isochronous transfers has been developed and incorporated in the USB Specification. The Audio Device Class definition adheres to this synchronization scheme to transport audio data reliably over the bus.

 

This document contains all necessary information for a designer to build a USB-compliant device that incorporates audio functionality. It specifies the standard and class-specific descriptors that must be present in each USB audio function. It further explains the use of class-specific requests that allow for full audio function control. A number of predefined data formats are listed and fully documented. Each format defines a standard way of transporting audio over the USB. Provisions have been made so that vendor-specific audio formats and compression schemes can be handled.

 

Many of the changes introduced in Version 2.0 of the USB Specification for Audio Devices take advantage of the new features provided in the USB 2.0 Specification. With the additional bandwidth made available, high speed USB operation allows the transport of multiple channels of high bit rate audio. This expands the range of solutions provided by USB audio devices but also challenges the way in which they operate. In addition to supporting the additional bandwidth, the specification supports new codec types for consumer audio applications, provides numerous clarifications of the original specification and extensions to support various changes in the core specification. The changes are not generally backwards compatible to 1.0 because that would too severely limit this new class of devices.

 

 

 

-Paul

 

 

[Talk2Me]

How about no cable at all? Anyone try this?

INFRARED, USB Infra-Red Wireless Adapter IrDA - 115kbps Transfer Rate (SIR mode)

 

[Julf]

"do you use bit rate check, oscilloscope, distortion analyzer on the DAC output, or other equipment?"

 

Mainly oscilloscope to check eye pattern, but verified with frequency analyzer and logic analyzer on the i2s link. DAC output is not interesting for this discussion, as we are talking about the digital part.

 

[Julf]

Basic digital transmission theory has not changed since the days of Claude Shannon.

 

Would this reference satisfy your requirements? Intersymbol Interference (ISI) and the Eye Diagram

 

To quote:

 

"A transmitter converts a message, or sequence of bits, into a series of analog pulses to create the signaling waveform. A receiver recovers the bitstream by periodically sampling the signaling waveform and comparing the sample to a threshold value to decide "1" or "0." Sinc-shaped pulse do not interfere with adjacent bit slots, provided that the bit slots are sampled at the correct instant in time."

 

That document contains a great picture illustrating the eye pattern:

 

 

 

As long as both the noise margin and the timing error sensitivity is > 0, the original signal is reconstructed perfectly.

 

Looking at the USB spec you quoted, it is important to remember that the "Management Overview" part is marketingspeak. The real technical info is elsewhere. But:

 

"An essential issue in audio is synchronization of the data streams. Indeed, the smallest artifacts are easily detected by the human ear. Therefore, a robust synchronization scheme on isochronous transfers has been developed and incorporated in the USB Specification. The Audio Device Class definition adheres to this synchronization scheme to transport audio data reliably over the bus"

 

So, as synchronization errors are *easily* detected by the ear, the protocol has been defined so that synchronization errors don't happen.

 

" What exactly are these "artifacts" and how precisely are they produced? Further, since these are synchronization artifacts, do they always include data loss?"

 

If jitter (or noise) is so bad that reliable reception is not possible, then yes, you get data loss. If the synchronization error is on a higher level (not bit-by-bit jitter, but buffer underrun), you get gaps followed by delayed sound.

 

[Julf]

Unfortunately the speed required by 96/24 audio is about 40 times the 115 kbps... Even CD-quality 44.1/16 requires 12 times the bandwidth of the IR link...

 

[Paul R]

But it appears to be a classroom assignment. To me it appears designed, like any good classroom assignment, to steer the student to an understanding of an accepted fact, not so much to test out boundary conditions in real world environments. I lie Labview.

 

I would rather see something a bit more targeted to hi-res audio, but I am not sure it exists.

 

By the way, Digital Signal Theory has changed a lot in the past 50 years or so.

Truthfully, I wonder if Paley-Weiner space and its accompanying reconstruction algorithms might be a better choice in some ways. Haven't spent much time on it, but it is interesting. A lot of math comes out about this stuff every year.

 

-Paul

 

 

[Julf]

"I wonder if Paley-Weiner space and its accompanying reconstruction algorithms might be a better choice in some ways."

 

I think the Hoffman transform would be more appropriate.

 

[Paul R]

An orbit that consumes the lowest delta-v? Huh?

 

-Paul

 

 

[jackfish]

I'm going to try the .5 meter DH Labs Silver Sonic USB cable.

 

[Mark Powell]

Two generic, of different makes, one 4 metres long, the other 1 metre.

 

A Chord.

 

An Audioquest Carbon.

 

To me, no difference at all.

 

[Talk2Me]

I have been using a Wireworld glass toslink from my MacBook Pro to my Wadia S7i. It sounds awesome. But I keep reading , "No, toslink does not sound good" (cough-Gordon-cough). Ok, methinks, what is all the fuss about usb about? How much for a usb cable, on a new platform I have never herd? Oh, those Locus Design cables look so bold and muscular. They are how much? Hmmmm, read a article raving about a inexpensive litte usb dac by Mhdt. Surely they now about usb cables too. How much for theirs? $45? Sold me right there. I will start with that, get used to usb, learn more about it, then see what other cables there are.

 

[audiozorro]

When another audiophile tried to convince me that power cords can make a sonic difference, our experimental experience and conclusion was that 7 out of 10 cords sound the same. 3 out of 10 sound different.

 

We tried out about 60 different cords and the evaluations took between one and two months. I believe we burnt in each cord 24/7 for at least one week before listening and comparing.

 

Good critical comparisons are time consuming and expensive, something that many audiophiles are not willing to do. They often want quick A/B comparisons without due consideration to the many factors that may change their conclusions.

 

A good example is the Mytek DSD Stereo DAC that has gotten a lot of positive feedback on this site. From what I gather from current owners the sonic quality of the Mytek changes and improves over the first four months. Some cables require 400-800 to reach their optimum sonic state.

 

As others often state YMMV, so if you have few miles invested your conclusions may not be complete or accurate. There may not be a difference with the two USB cables that you tried for the periods that you tried them, which for you may be all that matters. But IMO you would be selling yourself and audio short if you concluded that all USB cables sounded the same, when a fair number of other well respect posters have observed sonic differences.

 

Right now for me, I have bigger fish to fry. But if I ever find the time and money I would love to do some critical comparisons on some of the common digital cables recommended in this thread.

 

[Talk2Me]

Color me stupid, but why don't some of you (people who genuinely care) get together with burned in cables, and have a free for all shoot out comparing usb cables, on the neutral, best system? Am I asking for way too much? It rally gets my goat, and it should not, why the excuse is the cable has not been burned in yet.

 

[Paul R]

I mean, a lot of folks here wind up investing $8k to $20k or more in their audio systems. Taking the time to burn in a cable, even if nothing changes is nothing less than smart money management in my book...

 

Paul

 

 

 

[elcorso]

Then the problem could thee burn in regarding cables?

 

I do agree definitely and absolutely with you, everything should be burnt in, also the marriage (not always for good, the last one), in order to make then the comparisons.

 

Every component in you music chain has capacitors or capacitance, that varies with the use, so the cables.

 

I believe in this, and spent a good part of my audiophile life burning. Sometimes the better the capacitors in your gear (or capacitance in cables) lengthly the time to get it under the manufacturer specs. Is like a good guitar, violin or cello, it get better with the age (I use this as an example only).

 

Why? Maybe I know, but not a discussion I want. But maybe I don't know, then please don't ask me.

 

Roch

 

[audiojim]

Buy and try is way, way, way more difficult today than in the 90s when I got into this hi end audio subculture. But, of course, back then the only way to get a "deal" was to be a total hard ass , which might get you say 15% off list from the dealer. Now, with audiogon and the internet web sites of dealers listing "demos" it is a whole other world and you can get substantial discounts on new with warranty stuff, if you try very hard, even for the most supposedly exclusive dealers. But, so far as I can tell, it is really hard to borrow and try.

 

Another thing: it was explained to me by a large store in NYC that hi-end stores revnues are mostly driven by casual consumption buyers,particularly people installing home theatre and/or super duper systems, or walk-in buyers picking up a decent low to midlevel system. They view "audiophiles" who are highly likely to complain, want to buy and try, etc. as a pain in the ass.....Makes some sense to me, but maybe that was one store.