Yes, but you have to be more precise than what any link on Google tell you and really dig deeper into how this functions:
The signal at reception after recovery and interpretation from the voltage levels after comparing with threshold voltage levels becomes digital as we know it (i.e. two discrete values, binary).
The signal which is traveling on the USB cable is totally analogue, and made up of voltage levels and ramp up and down times.
The use of 'digital' when speaking of a normal USB cable is a rather facile and misleading shortcut.
The analogue nature of the signal on a normal USB cable makes any of the se cables subject to signal integrity issues.
Contrast this to a Corning optical USB, or contrast it to how a signal travels in optical fibre, which is pulses of light, i.e. here the signal truly travels in a discrete form, less subject to SI issues.
Hi Jan,
There's no conventional error per se in asynch USB (as opposed to the CRC + re-send in sync).
However, at the USB receiver, there is additional processing called into action when signal integrity issues exist so as to recover the bits (or define the bits).
This extra processing affects the DAC chips. At least that's how John Swenson (who has also done in the past new linear power supplies for better SQ as I believe you have too) explains his findings.
So, you may get the bits exactly as they were sent after the USB interface has done its job, but the noise profile of one transmission with high signal integrity issues isn't the same where there is no or much lower signal integrity issues.
Now, one could opt for a Corning Optical USB cable, because then we do know that the SI is less of an issue theoretically and in practice.
On the other hand, one needs to hear or measure whether the inherent noise profile of the optical decoder/receptor makes a difference as well.
Others prefer regeneration, re-clocking, clean power at the reception.
Mileages may vary...
What's even funnier is that some of the effects are rather obvious to hear and very easy as well to test oneself.
How difficult is it to add a switch on the GND line and toggle it on and off?
Or how difficult is it to put a piece of tape over it in the USB connector DAC-side?
Maybe "USB is digital. It's all perfect as is." is even easier for these people?
Totally agree, and I did post that not every issue is related to jitter.
The fact that Gino posted here with the 'jitter' term just fosters the contradictory, adds more confusion to his original questions (over at CA). It also gives an opportunity for the trolls to come out the woodwork so we can spot them too.
This makes it confusing for someone who just wants some guidelines as to where to focus tweaks, equipment purchase or even DIY, someone like Gino, or someone else.
Normally, the certified ones are the only ones which are allowed to use the USB logo. You can refer to the official USB website for how a manufacturers gets the certification to get more details on this. This certification is done by an approved third-party.
Now, of course, someone could also put the logo in there without proper certification, or use a fake logo.
I think you can get a certified USB cable from a known good manufacturer like say Philips and know it was well built.
However, you really need to understand that the logo or the certification does not guarantee that a USB cable will not be subject to signal integrity issues.
SI issues are inherent to the normal USB cable...
What does that mean in practice:
- You could get a no-logo, not certified USB cable from a third party which could be better than a logo-bearing, certified USB cable
- You could buy a known manufacturer logo-bearing, certified USB cable thinking that this is perfect. All this says is that it's certified to follow the USB specs to a certain tolerance. You may still get SI-related issues and then SQ differences if compared to a better built one.
That's totally and verifiably wrong though... The product is much more than that: the USB PHY power network noise profile being one, then there's clean power as well, and additionally, proper impedance matching.
As Swenson was mentioned I present the following of an example of him trying too hard and creating almost total nonsense. Q&A with John Swenson. Part 1: What is Digital? | AudioStream . OK it was an impossible task but still
!!!
Generally I've found that higher end systems from every manufacturer tend to be closer in cost than further, Apple included. It's lower revs products where the Apple tax seems greatest.
Wrt USB cables making an audible difference: only if there's something pathologically wrong. At least for USB 1 and 2, get a certified cable and move on.
I think a good question to ask is "is the analogue coding like on a USB cable" the best way to transmit a very high frequency signal for real-time audio?
To use the logo costs big money and you also have to get samples certified. I'm sure a lot of cheap cables are made by manuafacturers who never paid the money and never paid for certification.
You will also find very few cables that actually meet every one of the requirements in the certification document due to the extra labor costs required to solder the connector to the cable. A lot of this is due to the very poor mechanical design of the connector to cable interface. For USB 3.x, I defy anyone to solder the relatively large gauge wires to the tiny tabs one is given to solder to and keep the multiple twisted pairs orientated to meet differntial impedance, skew and crosstalk. It's like the cable connector designers never saw the bulk cable. The cables sent for certification have a lot more labor put into them versus a production cable.
Last edited:
Do you mean that you have finally got it that if the data is transported by the cable bit perfect then USB cables have to sound the same by definition.... Of course if you have any information that disproves this then it would be interesting....
- Status
- Not open for further replies.