Someone in the PC Audio forum in the Audio Asylum was bragging about how well his $550 USB cable sounded. He posted the review like a sales pitch for Synergy Cables.
While I repect other people's feelings I couldn't resist explaining that the cable conducts packets of data in the form of hi-speed pulses. I also explained that I couldn't understand how that $550 USB cable could have that kind of impact on the sound.
Holy sh*t!!! A Nuclear Word War 3 flame war errupted. Those people in that group take their cables seriously. I got to stay out of that forum!
While I repect other people's feelings I couldn't resist explaining that the cable conducts packets of data in the form of hi-speed pulses. I also explained that I couldn't understand how that $550 USB cable could have that kind of impact on the sound.
Holy sh*t!!! A Nuclear Word War 3 flame war errupted. Those people in that group take their cables seriously. I got to stay out of that forum!
Could digital cables have an effect on jitter or S/N that would affect the sound quality?
http://en.wikipedia.org/wiki/Jitter
http://en.wikipedia.org/wiki/Jitter
SY said:
It's a matter of whether the digital cable in question carries timing information or not. For an ethernet cable connecting a Squeezebox, there's no timing information and so the ethernet cable should have no effect if it's not faulty. Data transfer is asynchronous at that point.
USB is a mixed bag, though. I have the E-MU 0404 USB which uses asynchronous USB mode, so I'd expect the USB cable to have no effect. But most USB audio interfaces don't use asynchronous mode. The poster talking about the expensive USB cable was using one of Gordon's USB DACs, and I don't know what USB mode he's using. I didn't have the stomach to go through that thread to find out. That used to be a great forum, but it's been overrun by the usual audiophile suspects.
Oooh, quantum. 
JimOfOakCreek said:
You can probably alleviate some of your ignorance with some reading and experimenting.
99% of USB audio implementations rely on an embedded clock and consequently PLL clock extraction. Cable quality is sadly an issue, as it is in spdif transfers relying on the embedded clock.
What exactly is your problem? The specific figure of $550?
Re: Re: I Just Got Flamed In The AudioAsylum!
It's possible I may be ignorant but then it's possible I may have it right.
At the end of the USB cable we have the USB DAC with its clock and a buffer. It would seem to me any timing errors will get corrected in the end.
Even if the PPL is not 100% accurate we have a buffer at the end of the cable. Doesn't the DAC draw the data off of the buffer? Doesn't the DAC simply read the headers on the packets and adjust accordingly? Correct me if I'm wrong.
So how does the $550 cable impact these errors? I'm open minded and willing to learn.
I don't have a problem with someone buying a $550 USB cable. After all it's their money! However, I should be able to post my opinion without a hostile reaction.
analog_sa said:
It's possible I may be ignorant but then it's possible I may have it right.
At the end of the USB cable we have the USB DAC with its clock and a buffer. It would seem to me any timing errors will get corrected in the end.
Even if the PPL is not 100% accurate we have a buffer at the end of the cable. Doesn't the DAC draw the data off of the buffer? Doesn't the DAC simply read the headers on the packets and adjust accordingly? Correct me if I'm wrong.
So how does the $550 cable impact these errors? I'm open minded and willing to learn.
I don't have a problem with someone buying a $550 USB cable. After all it's their money! However, I should be able to post my opinion without a hostile reaction.
Re: Re: Re: I Just Got Flamed In The AudioAsylum!
I am certainly not an expert in the intricacies of USB audio. One thing is certain though: there is no "own clock" - it is PLL reconstructed clock. And the buffer is not anywhere deep enough.
From what i understand an asynchronous USB implementation works a lot better but most commercial and diy dacs do not use it as the chips require involved custom programming. Not to mention writing a device driver.
I have no experience with $550 cables either. What i have tried are cheap generic cables and yes, they do sound different. As do different USB motherboard chips. What makes me think that the USB audio standard is not that sturdy.
And neither is USB data transfer. Google "USB delayed write failure" and you'll find lots of external usb hard drives do not work as advertised.
JimOfOakCreek said:
I am certainly not an expert in the intricacies of USB audio. One thing is certain though: there is no "own clock" - it is PLL reconstructed clock. And the buffer is not anywhere deep enough.
From what i understand an asynchronous USB implementation works a lot better but most commercial and diy dacs do not use it as the chips require involved custom programming. Not to mention writing a device driver.
I have no experience with $550 cables either. What i have tried are cheap generic cables and yes, they do sound different. As do different USB motherboard chips. What makes me think that the USB audio standard is not that sturdy.
And neither is USB data transfer. Google "USB delayed write failure" and you'll find lots of external usb hard drives do not work as advertised.
Re: Re: Re: Re: I Just Got Flamed In The AudioAsylum!
That response has always puzzled me. How deep should the buffer be? At what point do errors become irrelevant because all the errors get buffered?
I understand that there is a latency issue with larger buffers. But...if the buffer was large enough to contain 3 seconds of audio that certainly should be enough to buffer ALL timing errors. After all we are talking miliseconds.
A 3 second latency isn't all that annoying is it?
So why don't manufacturers design a 3 second buffer with all USB DACs and the jitter errors would completely disappear, right?
analog_sa said:
That response has always puzzled me. How deep should the buffer be? At what point do errors become irrelevant because all the errors get buffered?
I understand that there is a latency issue with larger buffers. But...if the buffer was large enough to contain 3 seconds of audio that certainly should be enough to buffer ALL timing errors. After all we are talking miliseconds.
A 3 second latency isn't all that annoying is it?
So why don't manufacturers design a 3 second buffer with all USB DACs and the jitter errors would completely disappear, right?
I come at this with 18 years in the computer biz.
I think $550 is a tad extreme to say the least but it is no more extreme then using a $5 dollar cable. As is usually the case with most of these issue where we have extremes the truth lies somewhere in the middle.
Data transfer has little to do with timing issues as even a modest buffer is plenty to handle that except in the most extreme cases. The problem is packet corruption and even the largest buffer will not necessarily solve that.
The error correction has improved substantially but the vast majority of USB devices still use correction routines aimed at packets of data, not audio. The difference is obvious to anyone with the training and almost impossible to explain to anyone without.
The bottom line is buy decent cables. I wouldn't use a $5 cable to sync my pda much less play audio. But $550 and quantum tunneling ..damn, thats some funny crap. I guess everyone needs a hobby
edit.........................
I just thought I should come back to this and say a decent good USB cable can be had $30 and I personally would be comfortable using that for data or audio.
I think $550 is a tad extreme to say the least but it is no more extreme then using a $5 dollar cable. As is usually the case with most of these issue where we have extremes the truth lies somewhere in the middle.
Data transfer has little to do with timing issues as even a modest buffer is plenty to handle that except in the most extreme cases. The problem is packet corruption and even the largest buffer will not necessarily solve that.
The error correction has improved substantially but the vast majority of USB devices still use correction routines aimed at packets of data, not audio. The difference is obvious to anyone with the training and almost impossible to explain to anyone without.
The bottom line is buy decent cables. I wouldn't use a $5 cable to sync my pda much less play audio. But $550 and quantum tunneling ..damn, thats some funny crap. I guess everyone needs a hobby
edit.........................
I just thought I should come back to this and say a decent good USB cable can be had $30 and I personally would be comfortable using that for data or audio.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.