Longest USB cable length to DAC for audio?

[esldude]

5 hours ago, 57gold said:

Not an audiophile.  A musician since 1975 when I joined the union.

 

Ted, one of the moderators here, recommended the Granite Audio cable.  I believed the specs that up to 15 ft would be cool.

 

Sounded like crap.

Many audiophile brand USB cables, even expensive ones, don't meet USB 2.0 spec.  botrytis gave a link that shows what will be on it if it is certified.  The 2.0 spec is for up to 5 meters (16.5 feet).  Functioning properly USB cables don't have a sound.  

[esldude]

3 hours ago, sandyk said:

https://audiophilestyle.com/forums/topic/56912-usb-cables/page/4/?tab=comments#comment-978036    #84

 

 

 Did your measurements tell you that before you actually listened to them ?

That is NOT the experience of a vast number of forum members, and is why many prefer USB cables such as the Phasure Lush cables.

Measurements confirmed what I had heard.  If a USB cable works without dropouts, it sounds the same.  Of course I've not been listening to devices powered via USB.  They were merely transferring data.  I've rarely used the longer lengths of USB for audio (use them for printers sometimes).  I used the active USB extenders for really long USB connections.  Up to  25 meters (82.5 ft) without any error problems.  No sound problems either.  I wouldn't use them for a USB powered device.  For that I'd use the active extender and put a powered USB hub at the far end.  

[esldude]

19 minutes ago, sandyk said:

 That is exactly my point. Some DACs DO require a current draw from the USB port, the amount of which will depend on the design of the DAC . I also have Sandisk USB memory sticks that fail when plugged into a >3M long cable intermittently, sometimes not even being recognised. The faster the USB stick the more chance of this happening.

 

If this was the case there would be no market for boutique USB cables such as the Phasure " Lush"  and many others, but there clearly is, going by the threads about the Lush cable in this forum alone, and several other forums.

And of course, you performed the measurements before actually listening , didn't you ? 

 Why not post the actual measurements of the various USB 2.0 cables then   ?

No, I listened first. 

 

I've probably listed measurements at various times.  I don't have them now, as having figured out nothing is going on with them I've not had much interest.  I could do a couple again if there is any interest.  Of course I don't currently have any boutique USB cables.  

 

As for what there is an isn't a market for,  it isn't proof of anything other than what people will buy.  

 

Maybe these eggs aren't good for hormones, but work great for improved sound quality.  They are sold out at least. 

 

https://www.businessinsider.com/what-is-a-vaginal-egg-goop-lawsuit-2018-9

 

[esldude]

1 minute ago, sandyk said:

 

 Again, I will point out that just because you are unable to hear differences doesn't necessarily mean that others are unable to.

I guess that this kind of argument will never be conclusively settled, just like most other cable discussions can drag on for years ,even in the same forum, and get regularly revived.

You can't actually hear a difference that isn't in the signal.  You can perceive one sometimes.  Do you understand the difference?

[esldude]

8 minutes ago, sandyk said:

 

 Please do not insult a large number of other members by suggesting that they are delusional.

 Not all DACs fully reject RF/EMI that travels along with the Binary Data either.

 

 This issue is not likely to be fully resolved any time soon, no matter who posts what.

I'm proclaiming other members are human.  Maybe you wish to dispute that?

[esldude]

On 8/31/2019 at 11:42 AM, Stevan Mirkovic said:

 

Hello, this is my first post on this forum, which I have been following for some time.

I find many of the things connected to computer audio very interesting and perplexing.

English is not my native language, so I'm sorry in advance for the mistakes I will make.

 

Some time ago I have stumbled on a thread on some local forum (not in english, so I guess that the link will be of little use here), where a member conducted an interesting test in trying to determine if there is any difference in SQ between software audio players and operating systems versions.

The tests consisted of recording the S/PDIF signal coming from an USB to S/PDIF converter, using the coax S/PDIF input of a professional  PCM recorder.

 

He played a couple of test songs using combinations of different Mac OS versions and different players (if I remember correctly Amarra and Audirvana+ among others), and he recorded the results over the coax S/PDIF connection.

Later he precisely trimmed the recordings in order tho have them fully aligned, so that he can compare them using the null test technique.

For this test he disabled all of the processing options that were available in the players (various forms of down/upsampling, etc.) in order to have as "bit perfect" output as possible.

The first batch of tests surprised him a lot, since all the test files nulled each other completely! All the recordings were identical regarding their musical content down to the last bit.

 

In order to check if there was something wrong with his approach, he conducted a second test with upsampling filters turned on, and the differences between different players were there and they were very visible in the null test. So the test was working.

Also, as expected, there was a difference between the original files and the recordings but it was very small, if I remember correctly something under -120dB or so.

 

So the results of null testing were: there were differences between players with their filters on (as expected), there were differences between original files and their equivalent recordings (also expected), but there were ZERO differences between different players and OS-es with the filters off (for some people expected, but for some VERY unexpected!).

He posted the results on the forum and he was flamed since many members said that this was not possible since they can clearly hear vast differences between different players (but some of them were not even aware of different filter settings available in their software, or were they on or off).

After the (too) long intro  , the questions are:

 

1 - Is this test really valid, or is there some mechanism in professional PCM recorders that can get rid of possible errors (jitter or something) on its S/PDIF input, so that it always receives a bit-perfect signal regardless of the differences in software used on the computer side (provided that all the filters are turned off of course)?

 

2 - If it is valid, can this test be used to determine the differences between the USB cables!?

One could swap different USB cables between a computer and some USB to S/PDIF interface, make a recording from the S/PDIF output, and then compare the recordings using the null test? The null test will also provide the possibility of hearing only the differences if they are present.

 

For now, I can't think of any other test that is simple and "scientific" enough, and that does not rely on our subjective hearing abilities.

 

I'm usually restraining myself from writing on forums, this is the reason that this is my first post here. 

Many things in audio make no sense from "scientific" point of view and there is a constant war between believers and non believers, something that I would prefer not to be caught in, but there are examples like this and I can't stop myself from wondering...

 

 

Your English is very good. 

 

This looks like your first post, so welcome to Audiophile Style forums. 

 

I don't see anything wrong with the tests you describe.  It indicates the software gets the bits right, and that is all they have to do. The clocking is downstream in the USB to SPDIF converter.  

 

Any claim the bits can be perfect and various playback software has an inherent sound or coloration is highly unlikely bordering on impossible.   

[esldude]

3 hours ago, Stevan Mirkovic said:

 

Thank you for the welcome!  

 

So does this mean that this test could be accepted as a part of a proper test (not perfect of course) that @JohnSwenson asked for in one of his previous posts?
As some form of a “scientific” and unbiased test that does not rely on our subjective hearing abilities?

 

One could just swap the USB cables between the computer and the USB to S/PDIF interface,

record the S/PDIF output to PCM WAV files using a PCM recorder with the S/PDIF input capability,

and then compare the recorded files using the null test technique.

 

Of course that the results may be influenced by the hardware - some USB to S/PDIF converters may be more immune or sensitive than others to possible imperfections in the USB chain, etc, etc.

 

But if we keep the hardware constant and the only variable are the USB cables, we should be able to record/detect at least some of the differences between the cables?

Hopefully ones that have actual influence on the sound quality, not only dropouts in case of poorly constructed or too long cables.  

It would tell you the data is not changed.  With most DACs that is enough to say the sound is the same.  The data stream is not why USB cables sound different (if they do). 

 

Noise can be carried along the USB connection from source to DAC, and have an effect on the analog output when digits are converted to an analog signal.  There are DACs that are susceptible to this though not very many.  Some Schiit DACs being the main ones I know about.  The great majority of DACs isolate themselves from USB noise well enough it has no audible consequences.  

 

You could record and null out the analog output with different USB cables to see if anything changes.  It isn't as clean as doing this digitally.  You'll get very low level differences even if the USB cable has had no influence whatsoever.  I've done this and gotten nulls of less than -100 db using pkane's deltawave software.( https://deltaw.org/ )  If you record the analog output and do it again changing nothing you also get results less than -100 db.  So whatever effect the USB cable has is lower than the intrinsic noise floors of the gear involved.  

 

What you can see sometimes is low levels of 8 khz tones that are related to the USB format.  It may vary a bit from cable to cable and at different lengths.  I don't know where John Swenson's latest measurements lie on such things.  I don't take part in manufacturer curated parts of the forum.  

 

For the most part the USB cable matters somewhere between not at all, to an amount you might measure, but will never hear.  (as long as don't use cheaper Schiit DACs).