USB audio transmission isn’t bit true

[beerandmusic]

Great, glad to hear everyone agrees that the DAC does not receive music data from USB with 100% accuracy!

Something both sides can agree on.

 

Myth dispelled.

 

 

[esldude]

So dacs get the bits in completely accurate order more than 99.999999999999% of the time. Apparently more often than that some human has the mal conceived idea this rate of reliable performance is a problem. 

[sandyk]

58 minutes ago, barrows said:

 When done well, there is nothing wrong with USB audio, as with any interface, analog or digital, it does need to be done well if the best sound quality is desired.

 

Your typical SMPS Laptop without additional help (and additional expense !) between it's USB port and the DAC is FAR removed from this pre-requisite , despite what many from the Objective crowd will insist !

[beerandmusic]

51 minutes ago, esldude said:

So dacs get the bits in completely accurate order more than 99.999999999999% of the time. Apparently more often than that some human has the mal conceived idea this rate of reliable performance is a problem. 

 

Now that everyone agrees that usb does not deliver the bits with 100% accuracy, let's ponder the following:

 

Why does killing processes and changing buffer size improve SQ?

 

If all that exists is music and noise, how exactly does that translate into noise and influence the processing?

 

Is it that the dac can't handle the noise or is it that the noise causes an error, and that some dacs having to process for that error causes things like (more detail, more soundstage, ...etc..)?

 

RANKIN:

  Isosynchronous frames are afforded high priority in the system, but interrupts can affect priority.... because system stuff has really high priority over music playback software and again the music software can fault and bad sound will result. When a music app faults it becomes NON-bit true.

 

You may believe or not believe what Rankin says...but i think he knows USB more than most people here...perhaps not all, but most.

 

Curious why people believe one enet solution sounds different than a usb solution at all...shouldn't all enet and usb solutions sound the same given same os, same power, same everything else?  So no matter what the os settings are, no matter what power supply you use, no matter what OS or how many processes you have, the music is always recieved 99.99999999999999999999% accurate...and this mysterious noise signature that makes every usb and every enet solution unique can affect the DAC processing, but cannot affect the music...yet it is consistent...like more bass, more detail, more soundstage, not in a malfunctioning way, but all the time...I guess that is logical to some people.

 

Anyway, I made my point, the audio engineers can discuss as to why there are consistent differences by things like killing processes, changing buffer settings, etc...

I am not going to suggest that it is a problem or what the real issues are, because it is not my expertise...my only point of this thread is that I know now that DACS do not always receive data with 100% accuracy...and that different things can affect that accuracy. 

 

Read and understand the article and point out where Rankin is wrong and why...after all who cares what I say, i don't know crap, and am only quoting Rankin who likely knows more than "most" people here inre Audio....instead of being offended by my word "most", show your knowledge is greater than Rankin and disprove him....By ridiculing or responding to me, you are only showing your own ignorance, since i will be the first to admit I don't know anything about audio engineering.

 

[JohnSwenson]

I have done same tests that Gordon did and get similar results. Under some conditions you can get lots of data errors (many per second) with USB audio, under other conditions you can get no errors for days.

 

The biggest correlation I got was with cable length. With a cable greater than 3m you have pretty good chance of getting a large number of errors. Under 2m errors are few and far between (days between errors).

 

Between 2m and 3m is where the fun happens, you can trade off cable length and cable quality. For example with Supra cables you can go with significantly longer cables than cheap cables. In this range hardware has significant impact. For example my SuperMicro motherboard is error free with longer cheap cables than my cheap laptop.

 

I also did a lot of listening to the audio with different error rates on the USB. At low error rates the sound is identical. As error rates increase you start hearing the infamous clicks and pops. If you are getting say a click per minute, the sound quality in between clicks doesn't change. As the clicks come more often they get so annoying its impossible to tell if the sound quality is changing or not. At some point the errors come so often the system just shuts down, it can't handle that many errors. (a 5m cheap cable has a high probability of doing this).

 

So my conclusion was that if you stay with cables less than 2m you can be pretty sure you are essentially error free.

 

These rare bit errors do not seem to cause any sound change (other than a possible click).

 

John S.

 

[Arpiben]

6 hours ago, mansr said:

That's for USB 3. The USB 2.0 spec doesn't mention any particular number. Of course, given the much lower data rate, achieving the same level there should not be a problem.

USB 2.0 do have Jitter and Eye aperture specifications. Therefore this translates into BER even if probably not specified ( I need to check)

A certified USB cable will be more than sufficient at audio rates.

 

 

[beerandmusic]

5 minutes ago, JohnSwenson said:

I have done same tests that Gordon did and get similar results. Under some conditions you can get lots of data errors (many per second) with USB audio, under other conditions you can get no errors for days.

 

The biggest correlation I got was with cable length. With a cable greater than 3m you have pretty good chance of getting a large number of errors. Under 2m errors are few and far between (days between errors).

 

Between 2m and 3m is where the fun happens, you can trade off cable length and cable quality. For example with Supra cables you can go with significantly longer cables than cheap cables. In this range hardware has significant impact. For example my SuperMicro motherboard is error free with longer cheap cables than my cheap laptop.

 

I also did a lot of listening to the audio with different error rates on the USB. At low error rates the sound is identical. As error rates increase you start hearing the infamous clicks and pops. If you are getting say a click per minute, the sound quality in between clicks doesn't change. As the clicks come more often they get so annoying its impossible to tell if the sound quality is changing or not. At some point the errors come so often the system just shuts down, it can't handle that many errors. (a 5m cheap cable has a high probability of doing this).

 

So my conclusion was that if you stay with cables less than 2m you can be pretty sure you are essentially error free.

 

These rare bit errors do not seem to cause any sound change (other than a possible click).

 

John S.

 

thank you sir....one smart cookie in the bunch. 

Any testing inre quad dsd upsampling?

[beerandmusic]

28 minutes ago, JohnSwenson said:

I have done same tests that Gordon did and get similar results. Under some conditions you can get lots of data errors (many per second) with USB audio, under other conditions you can get no errors for days.

 

The biggest correlation I got was with cable length. With a cable greater than 3m you have pretty good chance of getting a large number of errors. Under 2m errors are few and far between (days between errors).

 

Between 2m and 3m is where the fun happens, you can trade off cable length and cable quality. For example with Supra cables you can go with significantly longer cables than cheap cables. In this range hardware has significant impact. For example my SuperMicro motherboard is error free with longer cheap cables than my cheap laptop.

 

I also did a lot of listening to the audio with different error rates on the USB. At low error rates the sound is identical. As error rates increase you start hearing the infamous clicks and pops. If you are getting say a click per minute, the sound quality in between clicks doesn't change. As the clicks come more often they get so annoying its impossible to tell if the sound quality is changing or not. At some point the errors come so often the system just shuts down, it can't handle that many errors. (a 5m cheap cable has a high probability of doing this).

 

So my conclusion was that if you stay with cables less than 2m you can be pretty sure you are essentially error free.

 

These rare bit errors do not seem to cause any sound change (other than a possible click).

 

John S.

 

 

What in your belief causes things like more detail and soundstage using same dac?

What is your comment about Rankins statement:

Bringing it all back home, the iFi iPurifier 2 likely improves the sound of the Sonicorbiter SE because it minimises transmission errors by making lighter work for the Mytek Brooklyn’s USB receiver chip....that doesn't sound like "clicks" to me.  I think excessive errrors can cause clicks where others will affect in other ways...(non-engineer opinion-smile).

 

I mean if you only have 1 dropped bit that I highly doubt you would hear a click....i mean how many bits in quad dsd are being processed in one second?..clearly you would not hear that as a click?....

 

You aren't the first one to suggest dropped bits only make clicks, but i think everybody and their brother jump on that theory to disprove that dacs recieve music with 100% accuracy....it's easy to say, well i didnt hear a click so i got 100% accuracy, but i am not so convinced.....in your testing did you ever see dropped bits but not hear clicks

[beerandmusic]

52 minutes ago, JohnSwenson said:

The biggest correlation I got was with cable length. With a cable greater than 3m you have pretty good chance of getting a large number of errors....

.....For example my SuperMicro motherboard is error free with longer cheap cables than my cheap laptop.

 

Along the same lines as that bolded, in all likelihood, if you had a slow processor with very little memory, and lots of processes running on a cheap mb and noisy ps, it is feasible that you would see even more errors per second even with a 1 foot cable....just for example..not saying you would or would not, just that it is feasible in theory, and that other things such as os, software, interrupts etc, also could cause more errors.

[beerandmusic]

Relevant

 

someone shared this link on one of my other threads....interesting listen....

 

https://darko.audio/2018/11/why-isnt-digital-audio-just-ones-and-zeroes/

 

A main point which i made earlier...it is best to get the signal right first rather than to try and fix it. (about 10:15 into listen).

[beerandmusic]

Everything i am reading reinforces my belief that ideally you would have a "network player" (e.g. streamer dac in one box) so that you have single manufacturer that can measure from input to output and know the entire environment (within reason) with less unknowns.  i am sure this concept will be pushed back by boutique shops or those afiliated and ignorant followers...but it is the best design....again, think SACD player...it can exist in one box.

[esldude]

1 hour ago, beerandmusic said:

snippage.............

 

Curious why people believe one enet solution sounds different than a usb solution at all...shouldn't all enet and usb solutions sound the same given same os, same power, same everything else?  So no matter what the os settings are, no matter what power supply you use, no matter what OS or how many processes you have, the music is always recieved 99.99999999999999999999% accurate...and this mysterious noise signature that makes every usb and every enet solution unique can affect the DAC processing, but cannot affect the music...yet it is consistent...like more bass, more detail, more soundstage, not in a malfunctioning way, but all the time...I guess that is logical to some people.

snippage..........

Everything, every idea of how something might be better with audiophiles has this consistent description of more bass, firmer bass, more soundstage, more space, more there there etc. etc. etc. etc. 

 

Analog cables, power supplies, digital improvements like oversampling, like Non oversampling, like sigma delta, like multi-bit and on and on and on.  It is all in everyone's head.  So yeah, it is logical to me. 

 

Your random and rarely encountered bit errors, your killed processes, your OS differences, all of that aren't the explanation.  The human in the loop is the explanation.  

[firedog]

5 hours ago, tmtomh said:

 

This is what this thread - and most of the other threads you start - really is about: You've found information that you think confirms what you want to believe, and you've no interest in any new or additional information that would complicate or negate what you think you know - all you care bout is "piece in your own mind." That's fine - but it's got nothing to do with a discussion forum. Just say it into a mirror and you'll achieve the same effect as posting it here.

Yep. Just looking for "evidence" to confirm what he already thinks. Then even if it is irrelevant, he latches onto it as if it is significant. One error every million, or  one error in 12 hours - and suddenly he has "confirmed" his preconceived idea. The fact that he has actually proved the opposite doesn't ever occur to him, apparently. 

[beerandmusic]

14 minutes ago, esldude said:

Your random and rarely encountered bit errors, your killed processes, your OS differences, all of that aren't the explanation.  The human in the loop is the explanation.  

 

I agree to some degree, but it really is a lot of everything, including personal belief and bias.  Many things do make a difference...most people doing these so called tests, do not have perfect environment for switching between kits, so i do grant you that.... but i also know i have heard differences, so regardless whether i can say it is better or worse, or i like one more than the other, regardless which is more accurate, or if i can even articulate how they sound different....there are differences even just by switching inputs on my ND8006....i dont think anything is OMG worthy, that we "can" be relatively close to a plateau, provided we have applied known lessons no matter whether we go enet or usb...but in my humble opinion, a network player is the ideal design with an optional LPS dependent on budget.  I also would like to revisit SD-card players noise measurements...that also could be an option for the "ideal" network player....beware every boutique dealer will push hard against that concept along with all their followers, but that is where I will continue to put my focus.

[beerandmusic]

13 minutes ago, firedog said:

Yep. Just looking for "evidence" to confirm what he already thinks. Then even if it is irrelevant, he latches onto it as if it is significant. One error every million, or  one error in 12 hours - and suddenly he has "confirmed" his preconceived idea. The fact that he has actually proved the opposite doesn't ever occur to him, apparently. 

 

THe darko audio in link above is very close to my beliefs...but i wasn't trying to get anything out of this thread other than for people to stop saying the data is always received at the DAC with 100% accuracy....Never bought it, never will....you guys may not understand why it was important to me..it was just i had difficult time that everyone seemed to accept that as fact, when it wasn't true.  Not that these "bits" make a dramatic difference, but the fact that there is any difference whatsoever reinforces my other thoughts....example...if i have to accept that 2 pennies = 3 cents then i can't accept that 3 pennies = 3 cents....I don't care since it's only one penny, but it makes all other logic fail....too hard to explain, let's just say it was mind settling for me to come to truth.

[esldude]

1 hour ago, beerandmusic said:

 

THe darko audio in link above is very close to my beliefs...but i wasn't trying to get anything out of this thread other than for people to stop saying the data is always received at the DAC with 100% accuracy....Never bought it, never will....you guys may not understand why it was important to me..it was just i had difficult time that everyone seemed to accept that as fact, when it wasn't true.  Not that these "bits" make a dramatic difference, but the fact that there is any difference whatsoever reinforces my other thoughts....example...if i have to accept that 2 pennies = 3 cents then i can't accept that 3 pennies = 3 cents....I don't care since it's only one penny, but it makes all other logic fail....too hard to explain, let's just say it was mind settling for me to come to truth.

You are not even aware of how much you are embarrassing yourself.  That makes it all the more embarrassing because you'll not understand this.  

 

My apologies.  

[firedog]

4 hours ago, beerandmusic said:

 

THe darko audio in link above is very close to my beliefs...but i wasn't trying to get anything out of this thread other than for people to stop saying the data is always received at the DAC with 100% accuracy....Never bought it, never will....you guys may not understand why it was important to me..it was just i had difficult time that everyone seemed to accept that as fact, when it wasn't true.  Not that these "bits" make a dramatic difference, but the fact that there is any difference whatsoever reinforces my other thoughts....example...if i have to accept that 2 pennies = 3 cents then i can't accept that 3 pennies = 3 cents....I don't care since it's only one penny, but it makes all other logic fail....too hard to explain, let's just say it was mind settling for me to come to truth.

Question: if you were told you had a choice between a gift of 1,000,000 dollars and 999,999.99 dollars, would you think it was a significant difference? That's what we are talking about here, not the difference between 2 cents and 3 cents. So these bits not only don't make a "dramatic difference", they in all practical effect make NO difference. Holding on to any other conclusion puts you in the obsessive compulsive camp where one might think the difference in the two sums I mentioned above is significant. 

As far as siginificant digits: you do realize that the Apollo program was devised by people working with slide rules?  And everything worked fine b/c their calculations were accurate in terms of significant digits, even if not "absolutely" accurate. You are doing the irrational opposite - essentially insisting on the significance of an insignigant digit. 

[yamamoto2002]

13 hours ago, beerandmusic said:

What about rankins testing in initial post...

 

 

I read the article

 

As Rankin-san said, it seems USB reliability becomes lower when many USB devices connected to one USB host controller. So maybe the error he experienced is caused by interference with other USB devices? (It should not be happened but in the real world, it happens)

 

You may have a interest to the following experiment, Challenge to connect theoretical max number of USB devices to one Windows 7 PC. This 3 page document is full of nightmare experience with unreliable computer. He has a great patience. From 80 USB devices, computer becomes really unstable.
 

http://ascii.jp/elem/000/000/720/720038/

 

This may be caused by CRC error (I don't have a equipment to confirm this so this is a guess)

 

I have similar experience to the document with my PC. When a USB hub with several USB devices is connected to the PC, often the hub and all the subsequent USB devices is not recognized and reconnecting cures the problem.

 

 

 

[mansr]

7 hours ago, JohnSwenson said:

I have done same tests that Gordon did and get similar results. Under some conditions you can get lots of data errors (many per second) with USB audio, under other conditions you can get no errors for days.

 

The biggest correlation I got was with cable length. With a cable greater than 3m you have pretty good chance of getting a large number of errors. Under 2m errors are few and far between (days between errors).

 

Between 2m and 3m is where the fun happens, you can trade off cable length and cable quality. For example with Supra cables you can go with significantly longer cables than cheap cables. In this range hardware has significant impact. For example my SuperMicro motherboard is error free with longer cheap cables than my cheap laptop.

 

I also did a lot of listening to the audio with different error rates on the USB. At low error rates the sound is identical. As error rates increase you start hearing the infamous clicks and pops. If you are getting say a click per minute, the sound quality in between clicks doesn't change. As the clicks come more often they get so annoying its impossible to tell if the sound quality is changing or not. At some point the errors come so often the system just shuts down, it can't handle that many errors. (a 5m cheap cable has a high probability of doing this).

 

So my conclusion was that if you stay with cables less than 2m you can be pretty sure you are essentially error free.

 

These rare bit errors do not seem to cause any sound change (other than a possible click).

That matches my own experience exactly.

[beerandmusic]

4 hours ago, esldude said:

You are not even aware of how much you are embarrassing yourself.  That makes it all the more embarrassing because you'll not understand this.  

 

My apologies.  

no need to apologize...you must of missed my post about how ignorant someone is to even respond to me because I will be first to admit I don't know crap.  Even in the Darko audio it is clear that engineer doesn't understand some things and suggest no one does fully understand it.  I will put you much further down the knowledge line than him, and we won't even discuss where in line you are in regards to human attributes that really matter in life.

 

For the record, i was flown to washington dc for a secret virtualization system, and i found system problems delaying rollout in areas where network engineers (which i am not one), couldn't find their own problems.  I am likely a lot more intelligent than you, so you and rolfl, tomtomm and other trolls on this site.  True, you all know more than me, but knowledge is not intelligence.  I can more accurately theorize in my mind, than you can resolve with your hands....in the mean time, i will just go on embarrassing myself. I grew up with lots of freckles, very skinny, couldn't pronounce my r's and other issues where I was teased most of my life, so I am used to it, so i certainly don't mind any nonsense you can throw my way....it's time for you to grow up little boy, before it's too late...you are the one embarrassing yourself....perhaps not to your troll buddies who will give +1 Likes to your ridiculing posts....but to everyone who really matters.

[beerandmusic]

21 minutes ago, mansr said:

That matches my own experience exactly.

try injecting all the different kinds of noise at different points

[beerandmusic]

1 hour ago, yamamoto2002 said:

 

I read the article

 

As Rankin-san said, it seems USB reliability becomes lower when many USB devices connected to one USB host controller. So maybe the error he experienced is caused by interference with other USB devices? (It should not be happened but in the real world, it happens)

 

You may have a interest to the following experiment, Challenge to connect theoretical max number of USB devices to one Windows 7 PC. This 3 page document is full of nightmare experience with unreliable computer. He has a great patience. From 80 USB devices, computer becomes really unstable.
 

http://ascii.jp/elem/000/000/720/720038/

 

This may be caused by CRC error (I don't have a equipment to confirm this so this is a guess)

 

I have similar experience to the document with my PC. When a USB hub with several USB devices is connected to the PC, often the hub and all the subsequent USB devices is not recognized and reconnecting cures the problem.

 

I am not going to read your link because it is not of interest to me (no disrespect to you)....is it about bulk or isochronis transmissions...either way, i don't really care.  The reason this topic was of interest to me, was because it was a stumbling block for me regardless of significance or whether it is relative or not.  I still will not decide if it can or cannot impact SQ and in which ways.  It is accepted it can cause clicks, but no one knows the depth which is possible.

I don't know actual numbers, but how many bits are processed in one second for quad dsd? 

[pkane2001]

40 minutes ago, beerandmusic said:

I am likely a lot more intelligent than you, so you and rolfl, tomtomm and other trolls on this site.  

 

So intelligent and yet so humble! 

 

[beerandmusic]

Off topic

 

How come we didn't need multi-thouusand dollar power supplies with SACD players and audiophiles didn't need such elaborate dacs and were content with a high end SACD player with its own internal dac back in those years?

 

Can't we do same with a sd card player today?  Want more storage, a NAS to a "network player" will be more than sufficient without USB...why require a poor audio transmission circuit or expensive cables for "optimal play"?

 

It also seems we could live without USB back then, why do we need it to day especially considering ISOCHRONus transmission....the boutique shops will push back til out of business or change their business.

 

Whatever...so much waste of life.

 

 

[beerandmusic]

7 minutes ago, pkane2001 said:

 

So intelligent and yet so humble! 

 

yea, sorry i was pushed so much...i have to stand up once in awhile...once i stood up to a bully in elementary school and clocked him on the nose and he cried, and many cheered....one of the only times i threw a punch in my life...i think i do pretty well for exercising my patience with idiots.