[Ynot1]

I think much of the discussions have been surrounding the theoretical limits of usb cable, and I am not able to challenge the reasoning either way.
I just want to know, does it really save money for manufacturing usb cables without the data line? For all those times I've been burned using charge cables when I needed data line, I wish I had a nickel for it each time.

 

[Rockofeller]

1. I'm going to have to call you out on that one! The theory of evolution or the big bang theory are examples of scientific models to explain observations of how the world works, HOWEVER, this is NOT the case with digital audio. Digital audio theory does NOT exist as a model to explain observations of how the world works, digital audio does NOT exist in nature, DACs do not grow on trees, have to be mined or orbit stars! Rather, digital audio theory was developed, then proven mathematically and several decades later the technology was developed to implement that proven theorem. If the theory of evolution were incorrect that wouldn't change the existence of millions of fossils and observations, just our explanation of them. On the other hand, if digital audio theory were incorrect, there would be no digital audio. It's inconceivable that a university professor (of science philosophy) would not know/understand these absolute basics of science and technology, hence why I'm calling you out!


G

Actually, I have to call you out on this one. This is completely wrong, although stems from a rather common misunderstanding, so don't feel too bad. Models and theories are just that - conceptual. It doesn't matter if you're talking about a model of real world phenomena or a theoretical system to be implemented out in the world. In the end, these models, which we can describe with words, math etc are simple-in fact, that's why they're helpful to us. Because the world is infinitely complex. So if you take a model about digital audio transfer and implement it in the real world, at best you have something PREDICTIVE. Not only does your argument that "since we invented it, then implemented it" not hold for this case, it will never hold, ever, in any endeavor.

 

[theorist]

Well stated!

Take digital cables, even though I know better I still picture a digital cable: USB, Ethernet, S/PDIF, Toslink, etc like a hose pipe (NZ term where I live) with little perfectly sized 'bits' of 'on(s)' going down it between perfectly sized 'spaces' of 'off'(s). The actual reality -- the 'thing-in-itself' -- is impossible to put into actual mental representation, hence we use simple metaphors like that of the garden hose (US term, I think). Indeed, no one actually knowns how the electro-magnetic spectrum actually works or looks like (if one could sense it) in reality, hence we build theories and models about it, and if they work most of the time, we then use these theories and models in our engineering, including that of digital and analogue audio. But the thing about science is that our knowledge in never absolute, the paradigm always gets overturned -- think of the 1980s claim of CDs as perfect sound, because in the measurements used at the time it was pretty near perfect, as no one had yet discovered jitter (among other things), what jitter does to musical sound and how to measure it! Similarly, our current models will one day be overturned. Top audio designers are at the cutting edge of this and one can argue that this is what we are paying for, at least in part, in the cost of reputable high end cables, unfortunately, there are always some cowboys selling expensive snake-oil, but that is another story! There was a good editorial/comment in 'The Absolute Sound' a few months ago of the effects of latency (how things take time to start moving and stopping in response to sound waves or other physical inputs of energy, a skew rate in electronic terms, or the time delay in your internet connection before triggering a response) of microphones in recording music and how we can therefore never have perfect sound -- its a good read, if you have not seen it.

If I may make a reply to some other posts here responding to my earlier ones. Karl Popper's argument was not that science could only disprove things, rather that the scientific method is predicated on testing if proposed theories actually replicate what happens in the real world and you could not call something scientific unless it could pass the test of falsification, ie the model being proposed can be be tested as valid through observation of it working in the real world (and it would remain considered 'valid' until it was shown to not work). He famously argued in the 1950s that the then claims for Marxian science and Freudian science were not actual science, as their theories could not be scientifically tested through repeatable experimentation and observation -- he called this the test of falsification -- by comparing them to Einstein's relativity where Einstein stated that gravity can bend light and he produced a formula in 1915 predicting how much it would bend. This was shown to be an accurate prediction the next time a solar eclipse occurred about three years later when someone measured the bending of light of a star behind the sun, confirming Einstein's prediction and hence the validity of the theory of relativity itself.

I just have a knowledgeable lay-knowledge of digital audio, but a pretty good knowledge of analogue audio, as, back in the days of the Audio Amateur I used to make my own designs of audio circuits for pre-amps, amps and the like, prototype, test/measure/listen/modify/etc, layout circuit boards, etch and build them, etc. Some worked pretty well, others less so! I did a bit of digital design but back in the 1970s that was before it was used for sound reproduction, just very simple computers with 8008s and the like.

 

[Currawong]

Seriously? If you put a digital signal into an oscilloscope you will see a square wave. That is something you can mentally conceptualise. The "bits" you talked about before DO arrive perfectly. If they don't, you get audible pops and clicks. I think what you mean about the limitations of science is really the limitations of your own understanding. The means by which electronic circuits work can perfectly explain how electrical noise from one connected device can end up in another part of the circuit.

They might have thought their USB implementations were good enough but obviously they made a mistake. Actually, they made two mistakes: 1. Their engineering design was flawed AND 2. Their quality control/testing was flawed. They should have done those experiments as part of their testing and identified and fixed that engineering flaw BEFORE the product was ever released!

Please explain, in detail, exactly what experiments and tests they should have done.

 

[theorist]

Do digital bits arrive perfectly? Then why do we have jitter (time smear) in everything digital. That's why some sophisticated USB inputs re-clock the arriving signal before passing it on in the circuit for reprocessing it to sound.

Further, if you look at a digital signal in an oscilloscope of greater resolution and one capable of capturing a narrower (ie faster) slice of time it will be far from square (the inherent slew rate alone must cause this, but other factors, such as capacitance interaction with the cable's insulation (a dielectric can never be a perfect insulator, this can only occur in a pure vacuum) and every physical change of material (circuit to plug to socket to wire to socket to plug to circuit -- even without any bad soldiers) the digital signal goes through also further creates capacitance ringing even with digital bits [like the ripples of water from a dropped stone on a pond]). The image you see is not that of the 'bit', it is simply something generated by the oscilloscope to a set of rules wired into the circuit as it reacts to the received signal and the rules in standard oscilloscopes are normally set to remove any fuzziness of the image made by very low magnitude error (often the noise of oscilloscope itself, but also some of the initial signal noise). Basically what you see is just another kind of metaphor, given as a 'standard' oscilloscope image, but it is not the 'bit' as a 'thing-in-itself'.

 

[chef8489]

Do digital bits arrive perfectly? Then why do we have jitter (time smear) in everything digital. That's why some sophisticated USB inputs re-clock the arriving signal before passing it on in the circuit for reprocessing it to sound.

Further, if you look at a digital signal in an oscilloscope of greater resolution and one capable of capturing a narrower (ie faster) slice of time it will be far from square (the inherent skew rate alone must cause this, but other factors, such as capacitance interaction with the cable's insulation (a dielectric can never be a perfect insulator, this can only occur in a pure vacuum) and every physical change of material (circuit to plug to socket to wire to socket to plug to circuit -- even without any bad soldiers) the digital signal goes through also further creates capacitance ringing even with digital bits [like the ripples of water from a dropped stone on a pond]). The image you see is not that of the 'bit', it is simply something generated by the oscilloscope to a set of rules wired into the circuit as it reacts to the received signal and the rules in standard oscilloscopes are normally set to remove any fuzziness of the image made by very low magnitude error (often the noise of oscilloscope itself, but also some of the initial signal noise). Basically what you see is just another kind of metaphor, given as a 'standard' oscilloscope image, but it is not the 'bit' as a 'thing-in-itself'.

You are trying to apply analogue rules to digital signals and you can not do that. A digital signal either arrives or it does not. The dac see a 1 or a zero. and then converts them in to a physical signal.

The USB protocol (unlike SPDIF) does not have a single problem with jitter. In Asynchronous mode, the audio device does the clocking. It tells the PC to regulate its speed. As the clock is running on the audio device this solution can be implemented in such a way that the the DAC receives a signal with a very low jitter.

What you do have to worry about is using a dac that takes power over usb versus through a separate power brick. Then you might get electrical interference if your source is not good.

 

[theorist]

You are trying to apply analogue rules to digital signals and you can not do that. A digital signal either arrives or it does not. The dac see a 1 or a zero. and then converts them in to a physical signal.

The USB protocol (unlike SPDIF) does not have a single problem with jitter. In Asynchronous mode, the audio device does the clocking. It tells the PC to regulate its speed. As the clock is running on the audio device this solution can be implemented in such a way that the the DAC receives a signal with a very low jitter.

What you do have to worry about is using a dac that takes power over usb versus through a separate power brick. Then you might get electrical interference if your source is not good.

A good discussion on jitter is at http://www.enjoythemusic.com/magazine/manufacture/0509/ Its a bit old as the reference to 192/24 makes clear, but still nice and comprehensive.

 

[chef8489]

That data is out of date and inaccurate.

 

[theorist]

That data is out of date and inaccurate.

As the web page starts: in regards to jitter "there are believers and unbelievers", and we are all entitled to our opinions. Ultimately, I believe in what my ears tell me.

 

[Arpiben]

@theorist
For a better understanding of Jitter and where it applies, I will suggest you to read Julian Dunn's paper.
When dealing with USB data entry the criteria is Max Tolerable Input Jitter.
Basically you apply jitter to data entry until having errors....
There is no issue at all with it since the USB chip inside DACs tolerates huge values of it.
That is the reason why some posters keep telling you there is no issue at all with USB cable jitter.
Rgds.

 

[gregorio]

Actually, I have to call you out on this one. ... Not only does your argument that "since we invented it, then implemented it" not hold for this case, it will never hold, ever, in any endeavor.

Obviously, if you're going to call me out on this one you've also got to call out the hundreds of millions of people who've transferred countless petabytes of data via USB over the last 20 years, bit perfectly AND without having to rely on audiophile grade USB cables to achieve this feat!

A good discussion on jitter is at http://www.enjoythemusic.com/magazine/manufacture/0509/ Its a bit old as the reference to 192/24 makes clear, but still nice and comprehensive.

Why would a self proclaimed university professor be quoting marketing material rather than actual scientific papers? And, not only are you quoting marketing information rather than actual reliable data/evidence but marketing information which completely contradicts your claims? You stated "It was jitter being created in the cable" but the article you describe as "nice and comprehensive" states "Digital Cables don't actively add jitter to the signal".

I don't get it, you're not only proving that you were lying about being a university professor but also that your claims about USB cables were false? If you really are a university professor then you can easily substantiate your claim that "So USB cables do sound different, they can create temporary distortion to do with burn-in, and this is observable in a very audible and measurable manner.".

Please explain, in detail, exactly what experiments and tests they should have done.

No idea, I was quoting YOU!
"Remember that the Wyrd, and subsequently the Eitr, come from highly experienced and competent designers who were wyrd-ly surprised by their experiments and their results, having previously thought that the USB they had designed was good enough."
"The manufacturers I've spoken to, two of whom have posted here about it, thought that their USB implementations were good enough until they either did experiments themselves (with electronics, not cables) and/or had feedback from customers that they could be better."

G

 

[theorist]

Why would a self proclaimed university professor be quoting marketing material rather than actual scientific papers? And, not only are you quoting marketing information rather than actual reliable data/evidence but marketing information which completely contradicts your claims? You stated "It was jitter being created in the cable" but the article you describe as "nice and comprehensive" states "Digital Cables don't actively add jitter to the signal".

I don't get it, you're not only proving that you were lying about being a university professor but also that your claims about USB cables were false? If you really are a university professor then you can easily substantiate your claim that "So USB cables do sound different, they can create temporary distortion to do with burn-in, and this is observable in a very audible and measurable manner.".


G

First off, as any prof knows, undergrad students, even in their own field of study, have difficulty reading academic journal articles -- they are difficult to understand without a lot of prior specialized knowledge, which is largely developed in graduate school doing their masters and then PhD -- that's why you have u/g textbooks, as they are easier to read and always explains the basics as was the case of the article I posted a link to (journal articles assume the reader knows all the basics and the specialised terminology used within it).

Also journal articles are peered reviewed by academic peers to ensure good scholarship. One of the sins of bad scholarship that peer reviewers are there to expose is the quoting of other literature incorrectly, or out of context. The full quote you partially stated is: "Cables don't actively add jitter to the signal, however they can slow the signal transitions or "edges". When the edges are slowed, the receiver or buffer at the cable destination is less likely to detect the transition at the correct time with certainty, which results in jitter." [technically this should be an indented quote as its more than one sentence -- but, hey, this is a discussion forum!]

In the case of part of the burn-in period of the USB cable I referred to, it was generating extremely high distortion with certain kinds of music, which I assume was jitter related (what else could it be since it was clock bandwidth, ie time, sensitive), replacing the USB cable with another in the system removed this jitter, returning it to the system returned the jitter. Putting the Auralic Vega DAC clock bandwidth setting to 'course' allowed the DAC to process this jittery data and play without audible distortion, on 'fine' it could not handle it without producing extremely un-listenable distortion -- I note there was no drop-outs in the music, it played continuously, but with extreme distortion, especially when the music got more complex with lots of transients. Perhaps the reason for this was the one stated in the quote about "edges" (aka rise time/slew rate) due to some specific attribute of burn-in, perhaps it was some other factor. I do not know, or particularly care as it went away with another 50 hours of burn-in. But it was clearly audible and, in its sensitivity to the 'bandwidth' setting, it was clearly measurable.

Finally, I am not a prof in the field of audio engineering or IT, nor have I claimed to be, my research is in how scientific and other knowledge is applied and used (often incorrectly through misunderstanding/overly simplifying it, etc, -- indeed, the world is actually very complex) within government policy. I largely draw on European philosophy and post-structural theory in doing so. I won't be more specific because in my very small country of New Zealand I can be too easily identified and I don't feel like having my house turned over for my gear!

 

[Currawong]

[...] I assume was jitter related [...]

[M]y research is in how scientific and other knowledge is applied and used (often incorrectly through misunderstanding/overly simplifying it, etc,

My father wrote a couple of books on pretty much the very same thing, though directed at ordinary people, which you might appreciate. That being said, as rude as this is going to sound, I think your assumption is an example of your research.

 

[gregorio]

[1] First off, as any prof knows, undergrad students, even in their own field of study, have difficulty reading academic journal articles
[2] The full quote you partially stated is: "Cables don't actively add jitter to the signal, however they can slow the signal transitions or "edges". When the edges are slowed, the receiver or buffer at the cable destination is less likely to detect the transition at the correct time with certainty, which results in jitter."
[3] In the case of part of the burn-in period of the USB cable I referred to, it was generating extremely high distortion with certain kinds of music, [4] which I assume was jitter related .. [5] Putting the Auralic Vega DAC clock bandwidth setting to 'course' allowed the DAC to process this jittery data and play without audible distortion, on 'fine' it could not handle it without producing extremely un-listenable distortion -- [6] I note there was no drop-outs in the music, it played continuously, but with extreme distortion, especially when the music got more complex with lots of transients. [7] Perhaps the reason for this was the one stated in the quote about "edges" (aka rise time/slew rate) due to some specific attribute of burn-in, perhaps it was some other factor. [8] I do not know, or particularly care as it went away with another 50 hours of burn-in. [9] But it was clearly audible and, in its sensitivity to the 'bandwidth' setting, it was clearly measurable.
[10] Finally, I am not a prof in the field of audio engineering or IT, nor have I claimed to be, ...

1. But according to you, you are not an undergrad, you're a prof and should have less difficulty reading a journal article than an undergrad! If uni professors relied solely on consumer review magazines for their knowledge and research rather than on peer reviewed papers, there would be no science and we'd still be in the dark ages ... that's why peer reviewed scientific papers exist, as any undergrad should know, let alone a real uni prof!
2. No! The "slowed edges" result in what's called an "eye pattern". The transition times/shape of this eye pattern is part of the USB specifications. So, either a cable can transfer the signal within these USB specifications, in which case it is a USB cable, or it cannot, in which case it is not a USB cable! So even a 99c no brand USB cable can transfer a USB signal perfectly. Likewise with a DAC, either it can handle a USB specified (eye pattern) signal or it cannot, in which case it's not a USB DAC. Therefore, if "the receiver or buffer at the cable destination" is unable to accurately detect the transition of a USB specification signal and deal with it appropriately, then it is NOT a USB device (or it's a faulty one).
3. NO! USB cables do NOT carry music, only on/off voltages representing zeros and ones. A USB cable does not know it is transferring audio, let alone know that audio is music, less still music of certain genre. It's just digital data, zeros and ones, which could represent anything from a word document to a feature film or the schematics for a jet liner. What if you send a selfie of your private parts down an un-burnt-in USB cable, does it generate "extremely high distortion" of that too or does it only wait for the zeros and ones which represent "certain kinds of music"?
4. Even an undergrad should know that you can't claim anything based only on assumption. In fact, that would be the very antithesis of science, as you would know only too well if you really were a uni prof dealing with scientific knowledge!
5. If your Auralic Vega DAC cannot adequately handle a USB specification signal without extreme distortion, then it is not a USB DAC (or it's a seriously faulty USB DAC). And, what is it you think your burnt-in audiophile cable is doing to the zeros and ones which a standard USB specification cable is not, to cause this "clearly audible" difference?
6. A USB cable does not transfer any transients, ONLY zeros and ones and, it's does not know what those zeros and ones represent and change it's behaviour accordingly! It's just a few strands of metal wire with no intelligence, no computational power and no magic!
7. Again, either the "rise time/slew rate" of the signal transferred by the cable is within the USB specifications or it is not a bona fide USB cable!
8. Even a moderately rational high school student should not be making claims if they "do not know", let alone an undergrad and heaven forbid an actual uni prof!!!
9. Again, pure speculation/assumption, something a real uni prof would NEVER base a claim upon! And, even an undergrad should know that something clearly audible/perceivable does not necessarily mean clearly measurable. It is trivial to perceive significant differences where in fact there are none and again, this is part of a fundamental tenet of science, which apparently as a uni prof you don't know!?
10. What does being a prof in audio engineering have to do with anything (?), USB cables do not carry audio, just zeros/ones and you don't need to be an IT prof to know and understand this basic fact of digital data!

G

 

[Clive101]

Just had a quick read through the blog.

I am not a scientist nor have magic ears.

I do know what sounds good to me, I have tried and tested several makes and expensive and cheap USB leads and despite what people say on this thread they do make a difference.

I found noise being injected into my system via the network the better USB leads helped with this.

All in all if you enjoy the cheap USB sound perhaps you have no noise or do not realise it is present.

I now go direct USB to DAC no network no noise.

If you'd want more detail just post