USB 1/2 is pretty forgiving as far as cable quality. A broken wire is probably the real issue.
I think it was mainly the connectors that didn't make a good reliable connection. I could wiggle the cable and cause dropouts and such.
Jan
Occam's Razor and information theory says that the 95% which sound the same are the 95% which can deliver an uncorrupted signal. The other 5% are either cheap rubbish or very expensive rubbish or severely incompetent DIY rubbish.Extreme_Boky said:
Yes, termination of highish speed data transfer cables is important. If close enough then the data arrives intact. If sufficiently wrong that the data is corrupted then it all sounds horrible (or gets muted). This would be noticeable. I have a horrible feeling that this is not quite what you meant.
Avoid cheap rubbish cables. Even more, avoid expensive rubbish cables. Even more than that, avoid severely incompetent DIY rubbish cables. Just buy cheap USB-compliant cables, as used to connect your printer/keyboard/etc.
They are just USB hubs that's all surrounded by a load of marketing ....
As for USB galvanic isolators.... this is not a guarantee of isolation for noise, galvanic isolators are primarily for voltage isolation, noise can be tricky.... the best example con product for these is the jitterbug...
http://www.diyaudio.com/forums/pc-based/278547-quick-audioquest-jitterbug-review-9.html#post4440782
Interesting that those who appear to prescribe the scientific method do not respect those who actually use it. Calling people ignorant who have actually made observations regarding sonic differences in USB transport implementations is insulting out of hand, more so when you have not made the same observations.
Observations (in this case listening tests) are scientific, after all the scientific method proceeds as: observations (Newton and the apple, for example)-hypothesis (speculation as to the possible causes of the observation)-and proof through experimentation (in this case finding some way to measure the phenomena observed).
In no case should observations be dismissed as Non-scientific. Just because observations have not yet been proven through measurements does not make them irrelevant. Not everyone, perhaps even no one, has the tools, or the knowledge of how to measure every possible low level effects on the analog output of the DAC due to different USB implementations.
Observations (in this case listening tests) are scientific, after all the scientific method proceeds as: observations (Newton and the apple, for example)-hypothesis (speculation as to the possible causes of the observation)-and proof through experimentation (in this case finding some way to measure the phenomena observed).
In no case should observations be dismissed as Non-scientific. Just because observations have not yet been proven through measurements does not make them irrelevant. Not everyone, perhaps even no one, has the tools, or the knowledge of how to measure every possible low level effects on the analog output of the DAC due to different USB implementations.
Respectfully,
Sy, at a former position part of my job was to do listening tests, so please be assured that when I post an observation achieved through listening it has been subject to a valid listening test, level matched (by measurement).
And, to repeat myself: By no means do I believe any data corruption of any kind is going on here. As has been mentioned, data corruption will result in a dropout, or small "tic" sound, and not a sonic degradation.
Sy, at a former position part of my job was to do listening tests, so please be assured that when I post an observation achieved through listening it has been subject to a valid listening test, level matched (by measurement).
And, to repeat myself: By no means do I believe any data corruption of any kind is going on here. As has been mentioned, data corruption will result in a dropout, or small "tic" sound, and not a sonic degradation.
Independently reinvented process every time someone put an open scope probe back on the rack. 😀
So you're just agreeing with what I said back in post 2, that USB jitter was a complete non-issue with respect to sonic degradation. Any packet loss results in pops and dropouts.
yes, sort of...
I would prefer not to refer to "USB jitter" at all. What I have observed is that USB receiver (and sender for that matter) implementation is critical for best performance, including the phase noise of the USB clock (not the audio clocks). I am not sure if the original question is really about "jitter" necessarily at all.
My point is to suggest that different USB receiver/transmitter implementations perform differently, and some are better than others (even in an all asynchronous format).
Why this is, for sure,I can only speculate, although there are engineers (such as John Swenson) who have worked quite a bit at improving USB audio transport performance.
I would prefer not to refer to "USB jitter" at all. What I have observed is that USB receiver (and sender for that matter) implementation is critical for best performance, including the phase noise of the USB clock (not the audio clocks). I am not sure if the original question is really about "jitter" necessarily at all.
My point is to suggest that different USB receiver/transmitter implementations perform differently, and some are better than others (even in an all asynchronous format).
Why this is, for sure,I can only speculate, although there are engineers (such as John Swenson) who have worked quite a bit at improving USB audio transport performance.
There can be a benefit from galvanic isolation (depending on how it isolates noise). Take a computer audio source, having a safety ground to the house wiring. Connect a USB cable from it to a usb audio adaptor, then an analog cable from the audio adaptor to the sound system (which also connects a safety ground). You have a ground loop, which can superimpose noise onto the analog audio signal from noise current following the path of the analog ground wire. Ive had it happen. Supply noise in the current conduction of the USB dac can also inject into this loop. Ive even had to use very short USB cables (lower resistance in the ground wire, and less inductive loop for magnetic field pickup, I assume) to use with some USB audio adaptors for making acceptable measurements. But magic cables aren't the fix, an analog interface with some ground isolation is.
Hi ! interesting point. This is what i was reading and caused my questions in the end. I have seen a number of products like USB isolators, repeater, reclockers and people swearing to hear benefits in highly resolving system.
For me i have already got benefit changing cheap and suspicious usb cables for better ones.
I have still problems to spot certified ones. Is there a way to understand if a usb cable meets or not the specifications ?
Thanks a lot, gino
Hi and thanks a lot for the very helpful advice.
I am interested in galvanic isolators, and i bought this one to try out
USB 1 1 to USB 5KV Isolation Adapter Converter 11023 Shentek DIN Rail Wall Mount | eBay
it works and it is also handy to inject some clean DC in the usb cable downstream. This is useful for interface that are usb powered.
I tend to prefer the sound with this device in instead of direct connection between the peripheral and the pc.
I understand that usb Vbus is 5V/500mA max. And not very clean.
Thanks a lot, gino
If the data gets through then there is no problem...
John Swenson has a product to sell, his comments and statements reflect this, all he has done is put together a standard USB hub using a standard USB hub IC with circuitry that is not that different from the data sheet, nothing special just a USB hub with a load of marketing prattle behind it, oh and USB packet noise was known years before someone decided they could use it as another Audio scaremongering myth and sell people products they don't need (quite common (very) in Audio and other markets)...
If the bits get through the data is there, no magic, not bits between bits, just data going from a to b and in the case of USB (UNIVERSAL SERIAL BUS) it is uncommon for devices not to communicate, so any data to back this up or just anecdotal stories.....
The layouts in most of the audio based stuff do not allow for true isolation or filtering of noise from one system to another, I have commented on this many times and put up some illustrations from a true isolated system which surprisingly have appeared in a device...
The noise from computers, servers audiophile music servers is high frequency mainly, the main path for this noise coupling is capacitively so unless this is catered for carefully in the layout this noise will just pass through happily and does in most devices. The main use for galvanic isolation is to protect one lot of circuitry from high voltages or faults that could cause high voltages and thus fry said circuitry, isolation for noise requires more than just say a Adum device (the Adums are capacitively coupled input to output). It can as pointed out though break ground loops.🙂
Firstly the point is to suggest, hence it is an idea with no current factual evidence...
As to cables they have the USB logo on them, that is a certified cable, look at USB.org for further info.....
- Status
- Not open for further replies.