Do USB Cables and Protocol really matter?

[HiRez_Lover]

I am not of the belief that audio cables can't make a difference (just ask the folks at The Cable Co),

I have tested dozens of interconnects and a/c cords and there are definite differences.

 

However, for USB, would shielding and construction really have an effect on the bits being sent and received?

 

If so: What cables would you point at to try?

Also, I see WireWorld now selling USB 3.0 cables - does the USB protocol have and effect

on overall sound?

 

TIA

Jay

[YashN]

I am not of the belief that audio cables can't make a difference (just ask the folks at The Cable Co),

I have tested dozens of interconnects and a/c cords and there are definite differences.

 

However, for USB, would shielding and construction really have an effect on the bits being sent and received?

 

If so: What cables would you point at to try?

Also, I see WireWorld now selling USB 3.0 cables - does the USB protocol have and effect

on overall sound?

 

TIA

Jay

 

The USB protocol for file transfer doesn't apply for Async USB audio applications is my understanding.

 

As for USB cables, I have posted several times about my own experience building one from basic USB cable and hearing vast improvements, and I am not the only one who heard the difference immediately.

 

I am not sure how much of a jump there is from that enhanced DIY USB cable and commercial ones simply because I haven't bought a commercial one, but there should definitely be a difference especially if signal lines and power lines are separated AND your DAC or equipment is susceptible to some USB jitter/noise (so not for all systems, but I think most of them).

[mrvco]

Based on my hit or miss experience with USB over the years, it is not hard to believe that there would be room for improvement.

[Stereolab42]

The USB transfer type typically used for streaming audio (isochronous) does not allow for retries if data is corrupted. The question is how often is this likely? Probably more likely the longer the cable and the more dense the data (e.g. high-definition audio). So a better cable and so on is probably a sound investment, as long as you don't go overboard. Now if you've got an exaSound DAC, which has a custom ASIO driver with (I assume) proper error correction with retries, you've got nothing to worry about.

[Starre]

USB cables do matter, I have tried at least 5 different ones and I had no trouble hearing the change of cables. It is not a night and day kind of difference but enough to make it worth trying. Of course, a good resolution system makes it easier to hear.

[thesurfingalien]

Now if you've got an exaSound DAC, which has a custom ASIO driver with (I assume) proper error correction with retries, you've got nothing to worry about.

 

Not completely sure, but I believe exaSound DACs still use isochronous USB, which has error-detection (it drops faulty packets). The ASIO-driver guaranties bit-perfect transport with low latency, but it does not necessarily mean a proprietary protocol is used.

 

I do not think bit-errors in USB-cables are that much of a worry if one uses a certified cable of the correct specification. If a cable causes bit-errors, it will be very audible because the dropped packets will be played as silences on most DACs.

 

Peter

[Jud]

The difference is not in the bits (if one is dropped you'll certainly hear it), but if at all in the timing (jitter) and noise.

 

Timing is not directly affected by the cable with async USB, but could possibly be indirectly affected by noise entering the clock or DAC chip.

 

All this is rather speculative and anecdotal, and there is a lot of work needed to satisfactorily confirm (or refute) that it is at least in part responsible for reported audible differences.

[thesurfingalien]

The difference is not in the bits (if one is dropped you'll certainly hear it), but if at all in the timing (jitter) and noise.

 

Timing is not directly affected by the cable with async USB, but could possibly be indirectly affected by noise entering the clock or DAC chip.

 

All this is rather speculative and anecdotal, and there is a lot of work needed to satisfactorily confirm (or refute) that it is at least in part responsible for reported audible differences.

 

Hi Jud,

 

That's why I limited my response to "bits"

 

There are plenty of topics here on CA that go into "differences between USB cables" for the OP to check out.

 

Peter

[mrvco]

Considering the amount of anecdotal evidence saying that USB cables do matter, I think the real question is "Will this USB cable matter in my specific system?". Many people here have more invested in their current DAC than I've spent on audio equipment during my entire lifetime, so I try to maintain some rational perspective before getting excited about a game-changing USB cable (or whatever) that costs as much as I paid for my DAC.

[thesurfingalien]

Considering the amount of anecdotal evidence saying that USB cables do matter, I think the real question is "Will this USB cable matter in my specific system?". Many people here have more invested in their current DAC than I've spent on audio equipment during my entire lifetime, so I try to maintain some rational perspective before getting excited about a game-changing USB cable (or whatever) that costs as much as I paid for my DAC.

 

That sounds like a very reasonable decision to me.

[Stereolab42]

Not completely sure, but I believe exaSound DACs still use isochronous USB, which has error-detection (it drops faulty packets). The ASIO-driver guaranties bit-perfect transport with low latency, but it does not necessarily mean a proprietary protocol is used.

 

I do not think bit-errors in USB-cables are that much of a worry if one uses a certified cable of the correct specification. If a cable causes bit-errors, it will be very audible because the dropped packets will be played as silences on most DACs.

 

Peter

 

George has not said precisely that they are using a proprietary protocol, but if you read between the lines it's hard to assume otherwise. They've said their DAC is completely immune to induced errors/jitter, which means they must be doing error correction... which rules out isochronous transfers (unless there's some way to abuse the protocol I'm not aware of).

[spdif-usb]

It depends. Some people prefer a warm sounding power amp to match with bright sounding speakers, or vice versa. Personally, I, prefer a neutral sounding power amp to match with fairly neutral sounding speakers. Similarly, some people prefer a DAC that sounds poor if used without a supercable, whereas I prefer a DAC that sounds exactly the same with a supercable as it sounds with a cheaper than cheap generic USB cable (i.e., one that I have, as a matter of true fact, plugged into an even cheaper USB extension cord, just because I can).

[JohnSwenson]

Also, I see WireWorld now selling USB 3.0 cables - does the USB protocol have and effect

on overall sound?

 

TIA

Jay

 

USB 3.0 supports all the modes of USB 2.0, plus adds additional ones. In particular it adds "super speed" which is quite a bit faster than "high speed" which is the fastest in USB2.0. Super speed requires a completely separate set of wires, they figured out how to add contacts for these in the "A" connector, but they couldn't do it for the "B" connector. The upshot is that a cable that supports all modes of USB3.0 will have a B connector that will not fit the jack on your DAC (unless it has a 3.0 B jack, which at the moment is extremely rare on DACs).

 

So it doesn't matter whether a 3.0 cable will sound better, it won't plug into the DAC, so I guess it sounds really lousy, no sound at all!

 

John S.

[thesurfingalien]

George has not said precisely that they are using a proprietary protocol, but if you read between the lines it's hard to assume otherwise. They've said their DAC is completely immune to induced errors/jitter, which means they must be doing error correction... which rules out isochronous transfers (unless there's some way to abuse the protocol I'm not aware of).

 

Like I said, I am not sure... I know their DACs run as master, so the DAC demands data from the PC. They also say their DACs are galvanically isolated from the USB-side. The data is buffered and re-clocked into the DAC. Could be that isochronous does not support that, but asynchronous does. However, I do not know (from the top of my head) if asynchronous supports error-correction.

 

What I do know is that USB data-errors are pretty rare to begin with; I can transfer data at full speed to a USB hard disk for days on end and not have a single error. The only times I detected errors was with defective or below specification cables.

 

BTW, who is George?

 

Peter

[sandyk]

What I do know is that USB data-errors are pretty rare to begin with; I can transfer data at full speed to a USB hard disk for days on end and not have a single error. The only times I detected errors was with defective or below specification cables.

 

BTW, who is George?

 

Peter

 

Hi Peter

Nobody is disputing that USB cables that meet the relevant specifications will transfer data with vanishingly few errors. What many are saying is that the higher performing USB cables have better RF/EMI rejection, which does affect the operation of the DAC despite correct binary data being received.

Is there a George in this thread ?

 

Regards

Alex

[Superdad]

I know their DACs run as master, so the DAC demands data from the PC. The data is buffered and re-clocked into the DAC.

 

That is standard stuff for any async-USB set up.

 

Could be that isochronous does not support that, but asynchronous does. However, I do not know (from the top of my head) if asynchronous supports error-correction.

 

Not correct. Synchronous, adaptive, and asynchronous are all "isochronous"--which just means the bus has reserved bandwidth for the data. And none of those modes supports error correction. That is not part of the USB protocol standard.

 

BTW, who is George?

 

George Klissarov is the engineer at exaSound. very sharp guy.

 

They also say their DACs are galvanically isolated from the USB-side.

 

Yes, they are one of the VERY few actually doing galvanic isolation without using lousy ADuM or optical isolators, and without going to some crazy expensive extremes. I think I just discovered how they are doing it, but it would not be nice of me to broadcast such information publicly.

[Superdad]

George has not said precisely that they are using a proprietary protocol, but if you read between the lines it's hard to assume otherwise. They've said their DAC is completely immune to induced errors/jitter, which means they must be doing error correction... which rules out isochronous transfers (unless there's some way to abuse the protocol I'm not aware of).

 

 

No, no proprietary protocol involved. And no error correction. Just really good isolation and smart engineering.

[thesurfingalien]

Hi Peter

Nobody is disputing that USB cables that meet the relevant specifications will transfer data with vanishingly few errors. What many are saying is that the higher performing USB cables have better RF/EMI rejection, which does affect the operation of the DAC despite correct binary data being received.

Is there a George in this thread ?

 

Regards

Alex

 

Hi Alex,

 

I was just responding to Stereolab42's post. He mentioned a George, BTW.

 

I am not going to open a can of worms again by arguing about possible beneficial effects of higher performance USB-cables. Like I said, there are plenty of topics dealing with that subject here on CA already. Besides that, I think the OP should make up his own mind by listening for himself.

 

Peter

[thesurfingalien]

No, no proprietary protocol involved. And no error correction. Just really good isolation and smart engineering.

 

I just checked, and for their e22 it says:

 

Proprietary Asynchronous USB Zero-Jitter Interface with error correction.

 

I am fairly sure you are correct and the interface is proprietary, but uses the regular asynchronous protocol.

 

 

I must confess I am rather impressed by their products; I would buy their e28 to use with a software-based x-over / phase-correction active loudspeaker-system in the blink of eye if I had the money for it...

 

Peter

[Jud]

No, no proprietary protocol involved. And no error correction. Just really good isolation and smart engineering.

 

Yes, I figured, because the only "proprietary protocol" giving complete immunity to jitter is marketing-speak. In other words, I know that the ExaSound equipment represents good engineering, but words like complete or perfect belong to the advertising rather than the engineering world.

[sandyk]

Besides that, I think the OP should make up his own mind by listening for himself.

 

Peter

 

Hi Peter

The OP needs to be aware of those other threads if he is to be aware of suitable USB cables for comparison purposes.

I don't need boutique USB cables because I have virtually eliminated the need for them by other technical means, such as keeping noisy +5v SMPS from the PC out of the USB cable and attending to potential ground loops.

Regards

Alex

[thesurfingalien]

Hi Peter

The OP needs to be aware of those other threads if he is to be aware of suitable USB cables for comparison purposes.

I don't need boutique USB cables because I have virtually eliminated the need for them by other technical means, such as keeping noisy +5v SMPS from the PC out of the USB cable and attending to potential ground loops.

Regards

Alex

 

Hi Alex,

 

I mentioned the existence of other threads twice, and the OP can search for these.

 

Your approach to use a high-quality PSU for the 5V supply and attending to potential ground-loops makes way more sense to me than investing in, as you call them, boutique cables!

 

Regards,

Peter

[sandyk]

Hi Alex,

 

I mentioned the existence of other threads twice, and the OP can search for these.

 

Your approach to use a high-quality PSU for the 5V supply and attending to potential ground-loops makes way more sense to me than investing in, as you call them, boutique cables!

 

Regards,

Peter

 

Hi Peter

I can't believe that some members are prepared to pay over $1,000 for a freaking USB cable. Some spend more on USB cables than the cost of the DAC itself!

 

Regards

Alex

[Stereolab42]

No, no proprietary protocol involved. And no error correction. Just really good isolation and smart engineering.

 

As alien points out, by their own claim the e22 uses a "proprietary Asynchronous USB Zero-Jitter Interface with error correction." I had forgotten I read that, embarrassingly enough. It's probably unwise for exaSound to elaborate more on that due to competitive reasons.