[2222440]

   
I am surprised you can notice the difference between USB ports, especially considering you are applying digital volume control which means you are dithering down the original source signal... What are you using as a reference? Guess, I really have to start experimenting with my photo printer to see which USB ports gives better prints? 

Not just me has noticed the differences of USB ports, some of my friends who use computer audio found the same.
 
As the name USB implies, it is SERIAL, both inside and outside the computer, the first port in the queue has the shortest distance to the processing center. The difference is rather obvious on my 6 years old IBM T60 laptop, it is due to another factor, the unequal oxidized levels of the metal surface, the disparity is less noticeable on my new ASUS laptop.
 
Regarding to data transferring rate, individual USB port on my MacBook can vary from constant 60m/s to constant 80m/s copying the same thing. The ports seem to have directions, some good at data in, some good at data out. 
 
The gears I am using:
 
i7  ASUS laptop, Win8.1(OS setting optimized), foobar/HQplayer/Jplay/JRiver
---> Supra USB cable/ Vertere D Fi USB cable/  some other printer USB cables
---> Ciunas DAC
[size=small]--->Cardas Golden Reference RCA interconnect cable[/size]
[size=small]--->Sim Audio Moon W5.3RS power amp  +  handmade Supra conductor wire power cable[/size]
[size=small]--->Bi-wire speaker cable. [/size]
[size=small]      Bass:  Supra PLY 3.4, [/size]
[size=small]      High-Mid: brandless cable[/size]
[size=small]--->B&W Matrix 801SE[/size]

 

[daerron]

  Not just me has noticed the differences of USB ports, some of my friends who use computer audio found the same.
 
As the name USB implies, it is SERIAL, both inside and outside the computer, the first port in the queue has the shortest distance to the processing center. The difference is rather obvious on my 6 years old IBM T60 laptop, it is due to another factor, the unequal oxidized levels of the metal surface, the disparity is less noticeable on my new ASUS laptop.
 
Regarding to data transferring rate, individual USB port on my MacBook can vary from constant 60m/s to constant 80m/s copying the same thing. The ports seem to have directions, some good at data in, some good at data out. 
 
The gears I am using:
 
i7  ASUS laptop, Win8.1(OS setting optimized), foobar/HQplayer/Jplay/JRiver
---> Supra USB cable/ Vertere D Fi USB cable/  some other printer USB cables
---> Ciunas DAC
[size=small]--->Cardas Golden Reference RCA interconnect cable[/size]
[size=small]--->Sim Audio Moon W5.3RS power amp  +  handmade Supra conductor wire power cable[/size]
[size=small]--->Bi-wire speaker cable. [/size]
[size=small]      Bass:  Supra PLY 3.4, [/size]
[size=small]      High-Mid: brandless cable[/size]
[size=small]--->B&W Matrix 801SE[/size]

 
Speed has nothing to do with the quality of the sound. In either event the speed would be largely irrelevant as audio is pre-buffered before being streamed as will be buffered on the receiving end as well, typically using a FIFO buffer. Most DACs have their own separate clock and will be used to read out the received data out of the FIFO. This is what asynchronous data transfer and clocking implies. The transfer of USB data is also digital and not affected by wear and tear e.g. the USB port will either work or it won't.

 

[highstream]

The guy who developed the beta dac and two-headed USB cable I mentioned several posts back told me to experiment to find the best sounding USB ports (hub) on my desktop. There definitely was an immediately noticeable difference between USB 2.0 pairs, including at that time a cheap aftermarket add-on card.  Why there'd be a difference, I don't know and didn't ask. Then I found a used SOtM (PCI) USB card, one of the three better ones developed for audio, and there was no question which to use (I also tried the supposedly best JCAT card, but my computer wouldn't boot - apparently a BIOS issue).

 

[2222440]

   
Speed has nothing to do with the quality of the sound. In either event the speed would be largely irrelevant as audio is pre-buffered before being streamed as will be buffered on the receiving end as well, typically using a FIFO buffer. Most DACs have their own separate clock and will be used to read out the received data out of the FIFO. This is what asynchronous data transfer and clocking implies. The transfer of USB data is also digital and not affected by wear and tear e.g. the USB port will either work or it won't.

I am not saying the speed relates to sound quality, just another example of the disparity of USB ports.
USB 2.0 's bandwidth is far more than enough for audio data.

 

[2222440]

  The guy who developed the beta dac and two-headed USB cable I mentioned several posts back told me to experiment to find the best sounding USB ports (hub) on my desktop. There definitely was an immediately noticeable difference between USB 2.0 pairs, including at that time a cheap aftermarket add-on card.  Why there'd be a difference, I don't know and didn't ask. Then I found a used SOtM (PCI) USB card, one of the three better ones developed for audio, and there was no question which to use (I also tried the supposedly best JCAT card, but my computer wouldn't boot - apparently a BIOS issue).

I guess the difference could be caused by the following reasons
1. All the USB ports are redirected from the main USB port inside a computer, especially desktop PC, the linking cable can be very long, In laptop, the cables are shorter. Usually the one closest to the motherboard has the least loss and interference.
 
2.The physical characteristics of each port let them sound different, typically the oxidized level of them. This often happens to RCA jacks. Different RCA inputs to the same amp circuit sound a lot different.

 

[2222440]

   
 The transfer of USB data is also digital and not affected by wear and tear e.g. the USB port will either work or it won't.

 
 
This is the digital theory.  However,  there are still lots unknown areas to the current theory. Some 30 years ago when CD was invented, the 16/44.1 format was regarded more than enough for human ear, but now comes 32bit/384khz.
 
In practical use, I find there are lots analogue characteristics on digital devices, including USB cables.
Generally, a thicker USB cable sounds more lush than a thin USB cable if they use same conductor, a soft USB cable sounds more laid back than a hard one. This rule applies to both digital and analog cables. 
 
I can roughly judge a cable's overall sound characteristic by pinching it and observing its conductor material, as you don't always have a chance to listen to it before paying. This skill has been proved effective most of the time.
 
One more thing worth to mention, if you play the music stored in an external hard drive, the USB cable linking to the computer plays a big role to the sound.

 

[i019791]

   
 
This is the digital theory.  However,  there are still lots unknown areas to the current theory. Some 30 years ago when CD was invented, the 16/44.1 format was regarded more than enough for human ear, but now comes 32bit/384khz.
 
In practical use, I find there are lots analogue characteristics on digital devices, including USB cables.
Generally, a thicker USB cable sounds more lush than a thin USB cable if they use same conductor, a soft USB cable sounds more laid back than a hard one. This rule applies to both digital and analog cables. 
 
I can roughly judge a cable's overall sound characteristic by pinching it and observing its conductor material, as you don't always have a chance to listen to it before paying. This skill has been proved effective most of the time.
 
One more thing worth to mention, if you play the music stored in an external hard drive, the USB cable linking to the computer plays a big role to the sound.

Thicker = lusher was the most funny for me

 

[daerron]

I am already considering dipping the interconnects in water for a more liquid sound. Perhaps they need some fridge treatment to conduct signals better? 

 

 

[Argo Duck]

^ Just don't dip the whole dac in water. You can have too much of a good thing

 

[2222440]

I am already considering dipping the interconnects in water for a more liquid sound. Perhaps they need some fridge treatment to conduct signals better?   

you will never know until you try it. let me know your result. if you cant hear the difference then you might need a more revealing system.

 

[2222440]

I am already considering dipping the interconnects in water for a more liquid sound. Perhaps they need some fridge treatment to conduct signals better?   

 

[2222440]

  Thicker = lusher was the most funny for me

 

 

[size=small]It is just a comparison to thinner ones, if other conditions are the same, thicker USUALLY lusher, of course not always. [/size]

[size=small]Simply being thick can not produce lush sound, I find those cables with groups of hundreds of very thin and soft conductor wire cores sound lusher and warmer than those with less amount of conductor wire cores.[/size]

 

[2222440]

   
The transfer of USB data is also digital and not affected by wear and tear e.g. the USB port will either work or it won't.

 
Both usb ports and cables are WEAR and TEAR devices, even CDs are of WEAR and TEAR.
 
Some of the guys here might think this is funny or redicolous, but don't forget, all logical digital data are implemented by physical analog devices. 
The 0s and 1s are implemented through high-low voltages of the current.
 
I am curious whether you use a printer usb cable for your hifi.  Do you run in new cables? 
If it is just a work/won't-work matter, do you think all those expensive HIFI usb cables in the market are all hypes? 
 
Even CDs need to be run in. When you play a brand new CD the first time, it usually sound slightly rough and sharp, after a fews times repeats, the CD will sound slightly different, more mellow, comparing to itself.  I have done this test for hundreds of new CDs. If you can not hear the difference, that means your hifi or your ears are not good enough. CDs wear out much slower than LPs, but they do wear out, unnoticeably.

 

[daerron]

Look, I am having a good chuckle with this discussion, frankly you are taking the mickey out of what electronic/computer engineers do for a living, but I'll have one more bite and then I'm done, no use further arguing the cause (I might just give up on the hobby altogether).

 
The physical interface of digital protocols forms a small part in the underlying architecture that determines ultimate performance and reliability. Good protocol design & implementation, drivers and software are hugely influential on the eventual outcome. This is why it took so much effort for audio designers to get USB correct for audio, because the protocol itself is tricky to master and not ideal for audio transfer and was more chosen because of convenience and availability. My first design experience with USB was frankly a nightmare too. For hardware EMI considerations, track routing and clocking are far more important than the state of the connectors and their distance from the PHY controller not because it will affect whatever is transferred over the interface, but for reliability.
 
Go study up the OSI model to learn about digital layers and also read up on technologies such as ECC and CRC. A CD can suffer wear and ports do suffer oxidation, but it is still possible to perfectly recreate the data from a disc, unless it becomes very badly damaged where the laser can no longer detect individual 0/1s and where this happen it will either halt playback or create hiss. There are hundreds or read/write errors on physical hard disks, yet these can be corrected (to a point though) by using error checking. These will show you how extremely robust digital interfaces are and why they have been created. They do not behave in unpredictable ways unless they have been poorly implemented. I would not get into a car or an aircraft for that matter if that was the case. Digital systems deals with 0s and 1s and synchronous clocking. This is why most effort is spent on reducing and eradicating clock jitter.
 
And yes, unfortunately my hearing ability is quite crap with a significant loss in the upper frequencies, no arguments about that. I do use quality analogue interconnects and speaker cables (where I think they could make a difference, check my profile for my cable inventory), but definitely won't spend ridiculous amounts of money on digital cables as long as they meet the digital interface specification. If you can hear a difference then good for you, but you should consider yourself and your friends experience as an exception and not the rule.

 

[Clemmaster]

FYI, USB audio does NOT feature error correction. It is not like your printer that will re-read the corrupted data (USB bulk mode).
Does it matter? Probably not, unless there's a heavy activity on the computer, there should never be any data dropout (I read a paper that said it would be in the order of one audio sample every few hours / days).

If you think analog cable make a different, consider how much more bandwidth a digital signal requires. Theoretically, it is infinite (square signal). We're dealing with transmission lines at this point: much more critical than a 5-30kHz audio signal.