Head-Fi.org › Forums › Equipment Forums › Sound Science › USB cable and Sound Quality
New Posts  All Forums:Forum Nav:

USB cable and Sound Quality - Page 11  

post #151 of 783

Clearly, a lot of people know more than you do .

No, the cable does not have any effect on how warm your zeros and ones sound .

Your best argument is 'IF' ..

Gear mentioned in this thread:

post #152 of 783
Quote:
Originally Posted by hifuguy View Post

Ya know... I really like you guys, but I think I am going to need some therapy pretty soon.  :-0

 

Just for fun.... here's a nice hypothetical question for you guys that believe in your hearts-of-hearts that USB cables don't sound better. Please humor me for just a second here... I'm really curious.

 

What would you say "IF" (note the big, and bolded IF).

 

1 - We all got together at my house... and drank a couple of beers (always a good plan... and maybe a pizza too... I like pepperoni)

2 - We ran a 100% fully legitimate DBT to you full specifications.

3 - The result of said DBT was that, "Yes" the high-end USB cable was identified as significantly better every single time. DBT passed with flying colors (remember the IF).

4 - The beer in step 1 had no influence whatsoever in the results of step 3.

 

What conclusions and thoughts would you have to say, if this were indeed the result?

 

And yes, I really do believe in my heart-of-hearts, that on my system, in my living room, on my music, that I really could deliver this result.

 

But coming back to our individual realities in the real world... I again suggest that unless you have heard a really good cable (Locus Design Polestar and Axis in my case), then I don't think that there is very much that you can tangibly contribute to threads such as this. The same goes for interconnects, speaker cables and power cables. On the other hand, if you really have indeed listened to a specific cable change (USB Cables in the case of this thread), then your experiences (confirming that you can hear improvements or conversely, that you can't) truly are invaluable to the our community and to our understanding of what is worthwhile and what isn't.

 

If a cable does deliver an improvement, then it becomes a simple question of whether that amount of improvement is worth it to you, as you prioritize your short and long term financial situation and goals. If yes, then buy it; if no, then send it back. It's pretty simple really. Therory is great, but in a thread like this where all of us are really pretty much amatures at best, the back and forth theory debates fail to bring much real value to the table in terms of a specif cable being considered is worthwhile (or not). And with that... I'll be at the therapist if you need me any further :-)

 

Take Care... Happy Listening!

I like beer and pizza biggrin.gif.

 

The nice thing about USB cable transfers is that due to the discrete nature of the signal, all you need is to verify error rate (since the analog conversion is done down the chain inside the DAC). Do you think running the USB cables in loopback and measuring bit error rate would suffice? Or is that not fair game here?

 

Say something like this:

http://www.protechdiagnostics.com/usb-loop-back-plugs.htm#

or this:

http://www.passmark.com/products/usb2loopback.htm


Edited by ultrabike - 8/15/12 at 4:36pm
post #153 of 783
Quote:
Originally Posted by hifuguy View Post

What conclusions and thoughts would you have to say, if this were indeed the result?

I'd run out immediately to see if the design of your cable is patented, because it would probably be even better for magically transmitting data than music.

As for not being qualified to comment on a cable without hearing it yourself, how do you feel about this one...

People are not qualified to speak on ghosts unless they've seen a ghost themself.
Edited by bigshot - 8/15/12 at 4:43pm
post #154 of 783
Quote:
Originally Posted by ultrabike View Post

I like beer and pizza biggrin.gif.

 

The nice thing about USB cable transfers is that due to the discrete nature of the signal, all you need is to verify error rate (since the analog conversion is done down the chain inside the DAC). Do you think running the USB cables in loopback and measuring bit error rate would suffice? Or is that not fair game here?

 

Say something like this:

http://www.protechdiagnostics.com/usb-loop-back-plugs.htm#

or this:

http://www.passmark.com/products/usb2loopback.htm

 

I don't think you can reduce it to BER (well, bits flipped over the channel, prior to coding / retransmission / whatever it does).  People are always going to make a big deal out of the small stuff.

 

 

If the device is USB bus powered, then surely the cable can be considered part of the power supply in a sense, and it may make some small difference—particularly if the "cable" in question is doing things like including a ferrite bead, rather than just being conductors in a sheath; or if you have a massive source of EMI nearby and you're touting a well-shielded cable.  If the ground, power, or signal lines are not isolated, then some kind of very small effect could couple into the analog parts of the DAC... or even if there is galvanic isolation maybe there could be some effect, and so on.

 

Note that "some small difference" doesn't necessarily imply a difference that people will pick up.

 

When the DAC is being clocked by a crystal oscillator, DAC is getting info through I2S or something from the USB receiver (so jitter or other small irregularities on the USB transmission itself are quite isolated from the operation of the DAC), power supply rails are all filtered, some reasonable design standards are applied, it really seems like there shouldn't be much effect from any part of the USB transmission itself, so long as all the data gets there on time.  That's regardless of asynchronous mode or whatever people want to use.  And if there is some effect from the USB transmission, it seems like the active parts of the system, rather than the cable connecting those elements, should have a much larger impact.

 

 

 

 

Quote:
Originally Posted by hifuguy View Post

What would you say "IF" (note the big, and bolded IF).

 

1 - We all got together at my house... and drank a couple of beers (always a good plan... and maybe a pizza too... I like pepperoni)

2 - We ran a 100% fully legitimate DBT to you full specifications.

3 - The result of said DBT was that, "Yes" the high-end USB cable was identified as significantly better every single time. DBT passed with flying colors (remember the IF).

4 - The beer in step 1 had no influence whatsoever in the results of step 3.

 

What conclusions and thoughts would you have to say, if this were indeed the result?

 

DBT can pass with a whole lot less than 100% identification rate, by the way.

 

Anyway, I'd want to try it again with other USB cables and analyze what's going on with the two that were tested.  I would think that if it actually happened, the most likely explanation would be that one of the two cables is faulty.  Or in the case of one cable doing things like lifting power/ground while the other does not, try again with a lower-end cable of similar functionality, and I'd check the DAC to see if there's any particularly bad susceptibility to whatever.

 

If everything everybody comes up can't find an explanation (i.e. no faulty cable, checks out in all the other ways) and particularly if these results can be repeated by others, then I would believe it.

post #155 of 783
Quote:
Originally Posted by bigshot View Post

What would I say to:

People are not qualified to speak on ghosts unless they've seen a ghost themself.

 

Interesting question!

 

In the context of a forum dedicated to gathering the real-world implications and impact of ghosts.... I'd most 100% agree with the premise :-)

post #156 of 783
Quote:
Originally Posted by mikeaj View Post

 

I don't think you can reduce it to BER (well, bits flipped over the channel, prior to coding / retransmission / whatever it does).  People are always going to make a big deal out of the small stuff.

 

 

If the device is USB bus powered, then surely the cable can be considered part of the power supply in a sense, and it may make some small difference—particularly if the "cable" in question is doing things like including a ferrite bead, rather than just being conductors in a sheath; or if you have a massive source of EMI nearby and you're touting a well-shielded cable.  If the ground, power, or signal lines are not isolated, then some kind of very small effect could couple into the analog parts of the DAC... or even if there is galvanic isolation maybe there could be some effect, and so on.

 

Note that "some small difference" doesn't necessarily imply a difference that people will pick up.

 

When the DAC is being clocked by a crystal oscillator, DAC is getting info through I2S or something from the USB receiver (so jitter or other small irregularities on the USB transmission itself are quite isolated from the operation of the DAC), power supply rails are all filtered, some reasonable design standards are applied, it really seems like there shouldn't be much effect from any part of the USB transmission itself, so long as all the data gets there on time.  That's regardless of asynchronous mode or whatever people want to use.  And if there is some effect from the USB transmission, it seems like the active parts of the system, rather than the cable connecting those elements, should have a much larger impact.

 

DBT can pass with a whole lot less than 100% identification rate, by the way.

 

Anyway, I'd want to try it again with other USB cables and analyze what's going on with the two that were tested.  I would think that if it actually happened, the most likely explanation would be that one of the two cables is faulty.  Or in the case of one cable doing things like lifting power/ground while the other does not, try again with a lower-end cable of similar functionality, and I'd check the DAC to see if there's any particularly bad susceptibility to whatever.

 

If everything everybody comes up can't find an explanation (i.e. no faulty cable, checks out in all the other ways) and particularly if these results can be repeated by others, then I would believe it.

 

Thanks for your comments... I really liked the last 13 words of your last post. You should give it a try. Pizza arrives in 30 minutes!  :-) 

post #157 of 783

everyone has a right to defend his wallet....it really is a pain to touch/buy any branded cable.

i defintely agree that usb cables wont do any good to your system. pls do not participate in any DBT...

or sneak a usb home to test.. u will hate yourself. i rather u have peace n bli$$.

wink.gif

 

but being an addict...i am just sad that the mcbookAir hasnt got a usb3 socket...sobsob.

 

that wireworld usb3 cable is so gorgeously made...how flat...how intricate inside..oooo...i wonder how it would sound like..

and it IS less than usd100bucks.oooooooooo...somebody SHOOT me now...save me.

atsmile.gif

 

=======================

 

hey hifiguy....

i had a REL subwoofer before...quite musical when properly matched to my bookshelf speakers...but my neighbours hated it when the dinosours stomped thru my living room.

 

But after a while...i still went back to my floorstanders...i cant get away from the paradigm of a tweeter/mid/bass driver set.

If u dun need the sub...power to u man!!

post #158 of 783

The nice thing about USB cable transfers is that due to the discrete nature of the signal, all you need is to verify error rate (since the analog conversion is done down the chain inside the DAC). Do you think running the USB cables in loopback and measuring bit error rate would suffice? Or is that not fair game here?

 

Say something like this:

http://www.protechdiagnostics.com/usb-loop-back-plugs.htm#

or this:

http://www.passmark.com/products/usb2loopback.htm

 

I don't think the sound quality that we have been debating here about USB cables is the result of data errors. I think the tester gizmos in the links above would give a pass to most any USB cable. I'm the first to admit that the underlying reasons as to why USB cables sound so different is challenging to understand. Audio, in the rarefied context that all of us talk about and love so much, is incredibly fragile. If I understood all of this, I'd patent the intellectual property aspects of it and start a company too.

post #159 of 783

 

hey hifiguy....

i had a REL subwoofer before...quite musical when properly matched to my bookshelf speakers...but my neighbours hated it when the dinosours stomped thru my living room.

 

But after a while...i still went back to my floorstanders...i cant get away from the paradigm of a tweeter/mid/bass driver set.

If u dun need the sub...power to u man!!

 

The Rel Subs are very nice.... cool for you. The sub I tried briefly wasn't in that league. You sound happy and that's what this is all about. My room isn't treated accoustically. It's my family room and the chances of my wife going for that idea are a big fat zero. Kind of a bummer, but that's my reality. In addition to the challenge of getting the "right" sub in the "right" place and adjusted just "right," I'm pretty sure my room might be best left 'unexcitted" at the real low (yummy) frequencies. I sometimes think a sin of omission is better than a sin of commission. As you say, "If u dun need the sub"...   Thanks so much for your thoughts!


Edited by hifuguy - 8/15/12 at 5:52pm
post #160 of 783

tot i read somewhere that usb data of 0s and 1s are not really 0s n 1s....

but a series of valleys( counted as 0s...) ...and peaks (counted as 1s )...

and below that waveform...there are aload of other signals that got transmittd as sound..luxurious sound or noise otherwise.

 

much like fuzzy logic in a washing machine...

4.5 / 4.9kg..5kg....its a full load..

2.3/ 3 kg...its a half load.

 

i am no scientist..or electrical engineer..

 

if i dun get at least a half load on my truck..its a NO GO...tats a  "0

if i get around 3/4 load..thats a GO...its a "1"

 

beerchug.gif

post #161 of 783
Quote:
Originally Posted by mikeaj View Post

I don't think you can reduce it to BER (well, bits flipped over the channel, prior to coding / retransmission / whatever it does).  People are always going to make a big deal out of the small stuff.

 

If the device is USB bus powered, then surely the cable can be considered part of the power supply in a sense, and it may make some small difference—particularly if the "cable" in question is doing things like including a ferrite bead, rather than just being conductors in a sheath; or if you have a massive source of EMI nearby and you're touting a well-shielded cable.  If the ground, power, or signal lines are not isolated, then some kind of very small effect could couple into the analog parts of the DAC... or even if there is galvanic isolation maybe there could be some effect, and so on.

 

If the device is USB powered then the problem will couple through the interconnects and the cable will matter little as far as that goes. I would make sense to me that a better source and/or DAC would make the difference.

 

The USB cable can have an affect on the signal due to a ferrite bead, EMI pick up, poor contact, out of spec length and/or gauge. This will translate into a higher noise floor, jitter, and affect the cable transfer function (ferrite bids, cable gauge and length.) All these parameters will affect bit error rate which is a direct function of SNR. FWIW, the test "gizmos" have the following disclaimer about cables:

 

"While just about any USB cable will work, it is better to select a quality high speed, shielded, USB cable when testing. Note that some of the cheaper USB cables are not shielded and thus are not recommended. The cable must be shorter than the 5m allowed in the USB standard. If a USB connection is required beyond 5m one or more hubs are required to extend a USB connection. We have noted during our testing that longer cables are more likely to have higher error rates."

 

So crappy cable should have crappy BER performance, in their experience.

 

Quote:

Originally Posted by mikeaj View Post


Note that "some small difference" doesn't necessarily imply a difference that people will pick up.

 

When the DAC is being clocked by a crystal oscillator, DAC is getting info through I2S or something from the USB receiver (so jitter or other small irregularities on the USB transmission itself are quite isolated from the operation of the DAC), power supply rails are all filtered, some reasonable design standards are applied, it really seems like there shouldn't be much effect from any part of the USB transmission itself, so long as all the data gets there on time.  That's regardless of asynchronous mode or whatever people want to use.  And if there is some effect from the USB transmission, it seems like the active parts of the system, rather than the cable connecting those elements, should have a much larger impact.

 

Yup! beerchug.gif

post #162 of 783
Quote:
Originally Posted by Lorspeaker View Post

tot i read somewhere that usb data of 0s and 1s are not really 0s n 1s....

but a series of valleys( counted as 0s...) ...and peaks (counted as 1s )...

and below that waveform...there are aload of other signals that got transmittd as sound..luxurious sound or noise otherwise.

 

much like fuzzy logic in a washing machine...

4.5 / 4.9kg..5kg....its a full load..

2.3/ 3 kg...its a half load.

 

i am no scientist..or electrical engineer..

 

if i dun get at least a half load on my truck..its a NO GO...tats a  "0

if i get around 3/4 load..thats a GO...its a "1"

 

beerchug.gif

 

For digital communications (just looking at the data transmission, not anything else now), there is some digitally-encoded information that gets sent from a transmitter to a receiver.  To actually send information somewhere in the real world, you need to transmit a (analog) signal, with some kind of predetermined agreement between the transmitter and receiver of what analog signal (symbol) corresponds to what information.  There are lots of different ways to map the digital information to analog signals, with different advantages and disadvantages, and often times there are a lot of intermediate steps and mappings between the original digital information and the symbols that are transmitted, not to mention all kinds of higher-level book-keeping / management / framing aspects to keep things going smoothly.  Something like USB is fairly simple.  It's the difference in voltage between two lines that is used for signalling.

 

As you say, there are a range of analog values that when received will be interpreted by the receiver as a '0' symbol and another range corresponding to a '1' symbol.  If a voltage corresponding to '1' is sent and the received voltage is so far off it's in the '0' range, then it will be incorrectly interpreted as a '0', resulting in a bit error.

 

So long as all the information (encoded as those '0' and '1' symbols) gets across, there's no data loss or change in anything.  End of story.  There's no hidden other data that corresponds to the music.  The music data is represented wholly by the digital information that's sent, that gets interpreted as those '1's and '0's.  For some systems, there is built-in redundancy and/or automatic retransmissions if something goes wrong.

 

 

So people are either implying a difference in bit error rate (probability of '1' sent and '0' being interpreted and vice versa, to simplify it), or something other than the operation of the digital communications link, that's different.  See my previous post about some of the potential "something else" issues—or non-issues, depending on your perspective.

 

 

 

 

Quote:
Originally Posted by ultrabike View Post

The USB cable can have an affect on the signal due to a ferrite bead, EMI pick up, poor contact, out of spec length and/or gauge. This will translate into a higher noise floor, jitter, and affect the cable transfer function (ferrite bids, cable gauge and length.) All these parameters will affect bit error rate which is a direct function of SNR. FWIW, the test "gizmos" have the following disclaimer about cables:

 

 

That's true: those things will affect bit error rate too.  I tend to think of small BER as pretty much 0 (comparing 0 vs. 0 in my head...), since pretty much every system I look at has FEC, so it doesn't matter.  How long do they need to run the testers to get a significant number of errors?  Seems like it'd be a snorefest waiting for errors to accumulate.  

popcorn.gif

 

fake edit:

 

I found something.

 

 

 

 

10 minutes per test to ensure no errors in 2.996 * 10^12 bits... (USB3 draft)


Edited by mikeaj - 8/15/12 at 6:39pm
post #163 of 783
Quote:
Originally Posted by mikeaj View Post

10 minutes per test to ensure no errors in 2.996 * 10^12 bits... (USB3 draft)

Given a decent source and quality DAC, it sounds like a lot of flawless music through standard USB cables cool.gif

 

BTW, by better source and quality DAC, I mean by implication better USB transceivers.


Edited by ultrabike - 8/15/12 at 6:50pm
post #164 of 783
Quote:
Originally Posted by Lorspeaker View Post

.i am just sad that the mcbookAir hasnt got a usb3 socket...sobsob.

We already gave you the tip on that. HDMI is your friend.
post #165 of 783
Quote:
Originally Posted by hifuguy View Post

I don't think the sound quality that we have been debating here about USB cables is the result of data errors.

That's all a cable can do. It can't improve the sound. It either passes the data across without changing it, or it distorts it. Wires don't have magical properties to make music more musical or anything like that.
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
This thread is locked  

Gear mentioned in this thread:

Head-Fi.org › Forums › Equipment Forums › Sound Science › USB cable and Sound Quality