|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
31.153.86.206
Following discussions in CA about the merits of a usb A to B converter versus cables, I decided to put the length of usb cable to the test.
Up to now, I have used 0.3m, 0.5m and 1m cables in preference to longer ones.
I had a Wireworld Starlight 5m cable and decided to make up 10 and 15 cm A to Bs to see if SQ changes. Plugs are Partsconnxion gold plated ones.
The exposed part of the cables have been kept to a minimum and care was taken over construction.
The 0.3 and 0.5 m cables sound similar, with the shorter ones being 'marginally' better and preferred. The home made cables are preferred to the factory terminated Wireworlds. These have half exposed stripped cables under the black plastic moulding and they sound different (and better to me) when a copper shielded is placed around the plugs.
The 15 and 10 cm lengths are another story. Sound opens up and there is obviously more attack and resolution to the music, plus better hf purity. Lf sounds different but I am not sure that it is 'better'. I am listening to the 10cm cable right now and I like it.
I am trying to make up some 5cm cables using unshielded lengths from Wireworld to 'simulate' an adapter but I am wondering what to do with the digital ground. Trying to install a copper shielding is proving to be very difficult. Perhaps I need to extract the shielded D+ and D- cable but this is also difficult. I may just try without.
Follow Ups:
Thanks for sharing your observations Fred.
10cm Wireworld W4S Dac2 DSDSe balanced out
This cable reduces hf aggression from some 2L DXD violin files and increases special definition considerably over the same 30 cm cable.
SQ has become a little more 'Yang' but placement with 10cm cable somewhat problematic. I shall need to make up a comfortable 'fit' but shall need to buy more USB plugs.
Length does matter but in this case shorter than standard.
nt
I do not doubt that you are reporting the sonic impressions you think you heard in sighted listening comparisons. I just place no credence in such uncontrolled tests and the often long winded theories that MIGHT explain possible sonic differences, as in almost all computer audio forums. Possible theories, some of them tortured and naive, are not the same as proven ones by measurement or by more controlled and unbiased listening tests.
I have to believe that somewhere engineers are making measurements of the basic data transmission performance of USB and other digital cables. But, we do not see them anywhere, except possibly in Archimago's blog occasionally. His measurements are generally about audio performance after d-a conversion in the DAC. He is just not finding anything about digital cables in his measurements, except for induced noise in nearby analog circuits.
Do measurements necessarily tell the whole story? Maybe not if we aren't measuring the right things. But, I prefer the objectivity of such tests, as opposed to listening testimonials by computer audiophiles under no controls at all.
My cable, by the way, is a 10 meter Corning glass fiber. Did I hear a BIG difference of it vs. a 5 meter metallic cable? Nope. Maybe there was an extremely subtle one at best. So, for now, it makes me a tad happier. I just do not think the sonic difference is worth getting all excited over.
I suspect that Corning chose the 10 meter minimum more for marketing reasons and price point rather than performance reasons, like longer is better.
... in Archimago's measurements?
Is he using adequate tools for the job, in your opinion?
Do you think the approach "can't hear differences - can't measure them - which explains why I don't hear them - so, everyone else can't either" is a valid one?
For something much more meaningful, you can search archives here for a post by Thorsten, I believe, where cables are measured using adequate equipment - and show substantial differences in measured performance.
You just have to love these guys with their righteous indignation when it comes to sighted listening observations. I also am amused with the love affair for measurements that have no proven correlation with the sound of the product being measured.
those in CA who go out and buy 10m plus optical cables just because a few posters say they are better. There are sound reasons why commercial optical cables are not the best transmission devices for audio signals, other than the stance that they 'isolate'. The simple facts are that the optics are powered by the same PC generated (noisy) usb voltages and that less than precision optical converters and fibres are not ideal for signal integrity. The ground lines in the Corning also have an impedance of around 3.3R, which is theoretically not good for a 'ground'.
Making my own quality usb cables from the same length and subjecting them to listening on the same PC with the same music seem more controlled to me.
Yes. I do think Archimago's is using "adequate" tools for the job. What is wrong with them? But, I am happy to keep my mind open to any critique of those tools or methods. And, I am certainly open to anyone else attempting to do a "better" job of measurement, which essentially does not exist on the Internet as far as I am aware.
I am quite convinced that sighted listening testimonials under unspecified conditions by unknown listeners with their choice of music recordings are inferior to Archimago's attempts to provide more objectivity and repeatability. Subjective listening may convince a poster that his own "tests" provide proof of some benefit, but why should I personally believe them? As I said at the outset, I am a skeptic. I have the same problem with most reviews in high end audio mags, by the way. There are some of USB cables in the current The Absolute Sound, for example, that are garbage in my opinion.
Whose approach about the applicability to everyone else are you referring to? Archimago's? I have never seen that in writing from him. I do not generally see it from subjective testimonials either. So, what is your point?
I looked but I found no measurements by Thorsten. As I said, I would welcome them as opposed to pages and pages in threads here and in CA about how much better XYZ sounded in their system. Or, I could just believe Lavorgna's review of the AudioQuest Diamond Ethernet cable, one of his advertisers, as the end all, be all sonically, even though he eventually was discovered to have induced a noise problem via EMI on the analog side of his system during his careful " listening tests".
They are simply not adequate for what he wants to measure. Hence the same results again and again. It's all a great waste of time and effort for him and 'believers' like you.
So, you believe his measurements are inadequate, and I believe, without further hard evidence which you have not provided, that they are. It seems to me we are both just "believers" in something, you just as much as me.
Just because you do not like the results that he gets does not mean his measurements are "inadequate". Maybe digital playback is in fact generally the same in sonic terms, and the claimed sonic differences are myths in people's minds, for example because of expectation bias. Maybe it is sighted and uncontrolled subjective listening that is inadequate. How can you say for sure?
There may be lots of useless idiots with cheap tape measures. It does not mean they are competent carpenters.
If someone presents measurements, it is up to him to specify the accuracy and precision of his measurements and to provide plausible arguments supporting his position. It is not up to other people to prove him wrong. Furthermore, if his measurements tend to go against another person's prior knowledge or belief it is reasonable for that other person to be highly skeptical of the measurements.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
I know, so do inmates here who are competent.
He is measuring his inadequate instruments.
"Yes. I do think Archimago's is using "adequate" tools for the job...."
Pretty fumy stuff. Much easier not having to think and just go by some useless fake technology.
Measurement tools are adequate only if they are 10x better than the devices being measured. Otherwise, one needs to be an experimental genius to make sense out of the results.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
10x? Where did that come from? Do you mean to say that the 180 mph speedometer in my car is no good above 18 mph?
This shows that you don't know anything about measurement systems.
I actually gave specific advice to him on what needed improving but he seems to insist on using the same stuff.
Waste of time.
That would be most interesting. Do you have any links to such a discussion?
So do you still think USB cables all sound the same?
I guess you might not understand the meaning of the resolution? Are you traveling 18mph or 17.999999?
I was being facetious, of course. And, resolution was not explicitly stated as an issue in Tony's post, to which I responded.
Now, here is a question for you. What is the difference between accuracy and precision? It seems to me that something can be very precise in the sense of having many decimal places in the answer, but it still might be wrong - off target.
Precision does not answer the accuracy question, and accuracy does not answer the precision question. Can an analog voltmeter with its swinging needle not be just as accurate in getting the magnitude right enough for purposes of comparison as a digital one with many decimal places? Aren't both speeds in your 18 mph example equally as accurate, with the many decimal places in the digital representation being useless for the purposes of driving the car?
I am completely open minded and just as ready as anyone else to jump on poor Archimago if indeed his measurements can be proven to be inaccurate. I am not seeing that proof, though. Do you have it? I am seeing a dismissive wave of the hand by many that seems motivated by a distaste for his results rather than the reason stated, which is his "bad" testing equipment and procedures.
I listen, too, and I do not make measurements, except for room acoustics, and I do reach my own conclusions. I do question those listening conclusions, especially if there is a perceived small difference, and try to avoid at least conscious bias. None of us can do much about eliminating unconscious bias in listening, except to try to design listening experiments or measurement protocols that reduce the possibility of unconscious bias, which is known to be pervasive in many fields from a host of experiments.
But, are anecdotal listening reports under uncontrolled conditions the gold standard? That is what many here seem to want to be the case. My own listening is often good enough for me, but I do not presume it to be good enough for anyone else unless I can offer more objective proof.
Hi and thanks a lot for the very interesting advice.
Do you have any pictures maybe ?
that would be also very interesting.
Kind regards,
bg
nt
Hi ! thanks a lot indeed. Very nice.
I am using some flat usb cables and liking them but i am afraid of much lower quality.
For sure i will keep usb connections as short as possible from now on.
Thanks again.
Kind regards,
bg
Seems completely wrong to me where shortest is always "best" but that might be making things too simple.
Just as Robert Fulton's assertion that speaker cables need to be some outlandish lengths (30+ feet) it could be the same for USB. Though I know not to that extent.
The author of a highly regarded player thinks two meters is best.
I wonder if the perception of greater transient response could be something other than what one initially perceives? Only time will tell.
There is the other aspect of whether it is such a good idea to have the DAC that close to the computer?
But affecting SQ, assuredly so. Just which direction, shorter or longer, is best, remains the question.
Shortest is not best where SPDIF is concerned. This has to do with bit rates and signal propagation on cables. USB is a packet system and the data rate is vastly higher, hence the required minimum cable length is vastly shorter if the goal is to space reflections between data transitions.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Each to his own; I am just carrying out a more or less controlled experiment.
By all means use 10m as Martin Colloms said of a cable. I tried it and it sounded inferior. May be someone had a huge stock to move..
I am not arguing with what you have found.
My inclination is to always use the shortest possible cable.
I was surprised to hear what I was told - but, my setup would not allow a cable as short as those you are using.
My reason for commenting is to remind (as I am always reminding myself) that only time will tell what is best. Too many times have I thought I had found something good only to have it drive me crazy a week or two later.
I am glad you continue to try things.
I am now pretty sure that this sounds better than longer ones, especially
by way of clarity and lack of hf the last bits of aggression.
I made this up using Wireworld D= and D- cable, with twisted pair V+ and V Ground connection.
Sounds ok but no where like the Wireworld 10cm one
d
Edits: 07/01/15 07/01/15 07/06/15
With digital transmission, is not characteristic impedance and length combine to dictate the time of reflections which affect jitter? Does this apply to USB cables as it does to COAX transmitting things like SPDIF or AES/EBU? Are USB decoding chips more or less immune to this phenomenon than audio-specific data formats?
More questions, more wonder...
Cheers,
Presto
This affects both and JS also highlighted the effect of packet noise in usb audio.
It's no good blaming the dac or the transport. It's bothe the messenger and the recipient together.
In the (unlikely) event that the source, source connector, cable, destination connector and sink all have the same impedance there won't be any reflections to worry about. However, there will still be cable losses and these will be frequency dependent, limiting the bandwidth of the cable, with a greater limitation for longer cable runs. The effect of longer cable runs will be reduced signal amplitude at the receiver and increased rise time. This will spread out the transition time between "0" and "1" and this will make determining the exact time of the transition more dependent on any noise at the receiver which adds/subtracts to the point at which the receiver decides whether it saw a 0 or a 1. That is to say, more jitter...If there are reflections, particularly reflections at both ends or from some discontinuity in the middle of the cable, then the reflections will affect the receivers decision making process if they take place around the time of transitions. The reflections also depend on the past data bits that have been sent, making the process difficult to understand.
With SPDIF/AES as normally used the sample clock is derived from the incoming waveform and thus can affect the timing of digital to analog conversions. With async USB the sample clock comes from a local oscillator and (theoretically) is not affected by the timing of events on the USB cable. However, in practice, the required processing by the USB receiver creates noise and this noise can (and does) couple into the DAC sample clock circuitry, so USB signal quality can still affect results.
With a sufficiently narrow bandwidth phase lock loop an SPDIF receiver can (theoretically) filter out jitter in the incoming clock so that the sample clock is not polluted. In practice this doesn't work perfectly. For example, the designers of the SABRE chip explain in their white paper that the SPDIF receiver circuitry on the input part of the chip creates noise and this noise can couple into the clock circuitry on the output side of the chip. The result may be a second order effect, but it can be measured and heard.
The USB receiver circuitry is vastly more complex and runs at much higher speed than lowly AES/EBU/SPDIF signaling. This means that the USB receiver is likely to create more noise. This will have to be isolated from the critical DAC circuitry and this is difficult and costly.
Tony Lauck
"Diversity is the law of nature; no two entities in this universe are uniform." - P.R. Sarkar
Edits: 07/01/15
and one were to use a USB Regen in the USB path, would you recommend as short a cable as possible between the Regen and the DAC?
The Regen is supplied with a solid adapter to the DAC. Next choice is a short cable.
It's not a good idea to 'optimise' using an adapter, which is not only not 'solid' but also contains unshielded wires and disconnected usb connector grounds.
The latter alone makes a sonic difference as you can try on any decent usb cable.
This lets me sit in my listening position with my laptop in my lap. I did a quick comparison to a short cabe and heard no significant difference.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: