[gregorio]

You could have saved yourself all that typing & just given us the soundbite version "bits are bits" but you would still be wrong - in mixed signal systems, the incoming noise can have an effect on the analogue part of a digital to ANALOGUE converter.

What is it about this that you don't understand?

 
The irrational reasoning part, IE. All of it!
 
If the incoming noise on the digital signal has an effect on the output of a DAC, then either the noise is so extreme as to almost obliterate the square wave or the DAC is not performing the fundamental purpose of why digital audio was invented in the first place!
 
What is it about this that YOU don't understand, it's simplified enough for a 10 year old to understand? No need to answer this question, we all already know what it is, as detailed in my last post!
 
G

 

[mmerrill99]

You could have saved yourself all that typing

The irrational reasoning part, IE. All of it!

If the incoming noise on the digital signal has an effect on the output of a DAC, then either the noise is so extreme as to almost obliterate the square wave or the DAC is not performing the fundamental purpose of why digital audio was invented in the first place!

What is it about this that YOU don't understand, it's simplified enough for a 10 year old to understand? No need to answer this question, we all already know what it is, as detailed in my last post!

G

OK, here's the problem - I understand the system as a mixed-signal system i.e a system containing digital & analogue circuitry - you don't seem to understand this (or want to deny or ignore it)
BTW, please leave the insults off your posts

 

[BiggerHead]

I'm firmly on the fence on this.
 
Ok digital audio was invented to prevent degredation of sound information right up to the point that it's converted to analog, but not beyond.  Digital audio was not invented to prevent noise from leaking into the output stage of an analog device from whatever directions noise might be coming in and in the end a DAC has an analog output.   There is nothing at all irrational at least in general about speculating that noise from ground connectiosn (nothing to do with digital cables other than incidentally) might impact that.  In practice how likely this is probably depends on many many factors, including how the sound output of the dac is referenced.  It doesn't take much googling though to find real credible looking cases of people reporting noise on their usb dacs coming from their computers.  I don't see why this is so irrational.
 
What does seem a bit far fetched is this whole noise only in inaudible range, even after 10,000 amplification, but enough even without that to cause noticable IMD.   That seems like a stretch.

 

[mmerrill99]

I'm firmly on the fence on this.

Ok digital audio was invented to prevent degredation of sound information right up to the point that it's converted to analog, but not beyond.  Digital audio was not invented to prevent noise from leaking into the output stage of an analog device from whatever directions noise might be coming in and in the end a DAC has an analog output.   There is nothing at all irrational at least in general about speculating that noise from ground connectiosn (nothing to do with digital cables other than incidentally) might impact that.  In practice how likely this is probably depends on many many factors, including how the sound output of the dac is referenced.  It doesn't take much googling though to find real credible looking cases of people reporting noise on their usb dacs coming from their computers.  I don't see why this is so irrational.

Thank you - exactly so.

What does seem a bit far fetched is this whole noise only in inaudible range, even after 10,000 amplification, but enough even without that to cause noticable IMD.   That seems like a stretch.

I don't see the difficulty here - out of band noise (i.e noise above the audible frequency) is a well known issue in DACs & one that DAC manufacturers attempt to minimise. Alias images are out of the audible range but it's known that they can have audible effects by folding back into the audio band. Out of band signals (which includes RF signals) can have significant effects on downstream analogue devices - the opamps on DAC outputs, for instance. None of this is any mystery or anything "magical" to those who know about these issues - I guess it might be seen as "magical" to those who haven't encountered this before but on a section called sound science I would have thought this shouldn't be the case.

 

[mmerrill99]

One understanding that also seems missing here is that we are not necessarily talking about noise riding on the signal, although this is possible/likely, but also noise on the ground connection. This is typically what a ground loop noise is - disturbances on the ground which should ideally be fixed at 0V & unwavering i.e not fluctuating.

Ground is a reference signal that all other signals are referenced to - if this reference signal is fluctuating, then it has an effect on how signals are processed

I think there's also a problem here in people's understanding of the linearity of systems - IMD is a reflection of the non-linearity of systems i.e in a linear system a 19KHz & 20KHz tone will not produce any signals at other frequencies - look at any DAC & you will see small signals produced at other frequencies - this is as a result of the non-linearity of the system.

With non-linear systems we get signals produced at other frequencies not in the original signal - for instance if we have two signals at 22KHz & 24KHz, we can have signals arising at 2Hz due to non linearity

 

[BiggerHead]

One point is yes the ground itself fluctuating means all signals are referenced to the "wrong" thing, however, if you just plug your headphones directly into that, your headphones are also referenced to the wrong thing and there isn't necessarily a problem.  Reality can be complicated though and ground loops find their way around.
 
I'm no expert on the IMD stuff but it just seems like a) the dac should have some filtration against very high frequency junk b) the stuff would need to get through at pretty high levels.
 
c) Ok, if this is really something going on,  fine but
   c.1) Assuming these fancy cables can fix it, wouldn't the makers of these kinds of cables to be showing some measurements about this?  Maybe they have.  Have they?
   c.2) How the heck are you going to fix a ground loop with a fancy cable that still connects the grounds?  Ferrite cores maybe?, ok.. and that costs what? Unless it actually isolates the grounds I don't see where the expense is justified (and even then, some of these cost serious bucks).

 

[gregorio]

Originally Posted by BiggerHead /img/forum/go_quote.gif
 
Digital audio was not invented to prevent noise from leaking into the output stage of an analog device from whatever directions noise might be coming in and in the end a DAC has an analog output.

 
Digital audio was invented to prevent any noise entering the analogue output from the digital data. Once the digital signal is converted to an analogue signal it is, by definition, not a digital signal any more and there's nothing digital audio can do about it. However, a USB cable is the input to a DAC, at the point of conversion the data has by definition already exited the USB cable. In other words, the USB cable has completed it's role of transferring digital audio data at the input of the DAC, before the conversion to analogue and therefore it is bound by the fundamental principles of why digital audio was invented! In addition to the digital audio data, the USB cable may also be supplying power to the analogue stage of the DAC but that's the completely separate issue of power supply management within an analogue signal chain and has nothing to do with digital audio. But I would say that as power supply is part of the USB specification, if the DAC cannot adequately manage the USB specification power it is being supplied then it's incompetent, maybe even to the point of not being worthy of actually calling itself a USB specification device.
 
G

 

[BiggerHead]

oh and aren't usb grounds required to return

   
 However, a USB cable is the input to a DAC, at the point of conversion the data has by definition already exited the USB cable. In other words, the USB cable has completed it's role of transferring digital audio data at the input of the DAC, before the conversion to analogue and therefore it is bound by the fundamental principles of why digital audio was invented!
 
G

 
You're treating this as a schematic definition of what it should do.  In reality though, it doesn't just do what's it's told to do and provide 1's and 0's.  It also connects the ground of  an analog device to the ground of a computer, and creates a loop in the process.  I'm pretty sure the digital side of the DAC and the analog side are on the same ground.  You keep acting like this is a just an information path.  Forget the data.  Pretend the usb cable is optical instead.  Now just connect a wire from the dac case to the computer.  Who cares what the purpose of the wire is.  The fact is, you just connected the ground of an analog audio device to a noisy computer and created a ground loop attached to that analog audio device.

 

[BiggerHead]

OK, so I should back up.  I did start playing along presuming that in fact the cable does connect the two case grounds, just to give the benefit of the doubt that there could be an issue.  This isn't necessarily true, and could depend on the DAC design.  Maybe any reasonable DAC simply doesn't connect the USB case to ground on the DAC side.

 

[mmerrill99]

OK, so I should back up.  I did start playing along presuming that in fact the cable does connect the two case grounds, just to give the benefit of the doubt that there could be an issue.  This isn't necessarily true, and could depend on the DAC design.  Maybe any reasonable DAC simply doesn't connect the USB case to ground on the DAC side.

Now we are getting somewhere
Most current USB DACs operate at USB 2.0 high-speed i.e 480Mbps (480MHz). There are only a couple of devices that galvanically isolate (isolate signals & grounds) at these speeds - the Intona device & the Corning optical device.

There is no current galvanic isolation chip that can be used internally in a DAC that will cater for this USB 480MHz two-way communication needed.
There are some but very few DACs that galvanically isolate after the USB receiver inside the DAC - usually when the signal has been converted to I2S & it's the slower I2S signals that are isolated. Some level of immunity to upstream noise is conferred in this arrangement but not complete immunity

So if immunity to incoming noise is the criteria for whether a USB DAC is broken or not then we can conclude that all USB DACs are currently broken - that's why I regularly asked for an example of an "unbroken DAC"

So, if we accept that USB DAcs are not immune to incoming noise, we are left with the question - what are the possible sources of this noise & can a cable have any effect on the spectrum or level of this noise?

One needs to know about the various types of noise & study the system itself in order to be able to answer this. USB signalling is a differential signalling system that also reverts to single ended for certain signals. Differential signalling is used to provide some immunity to noise but it relies on many factors in order to achieve this noise reduction. How this works in theory is that any external noise which impinges on the cable will, because of the twisted pair construction of the cable cause an equal level of disturbance on bothe USB signal lines, D- & D+ The USB receiver is a summing receiver & only voltage differences between the D+ & D- are sensed as signal. Any signal of the same voltage that appears on both D+ & D- will be cancelled out & not sensed. As you can see this cancellation relies on the amount of perfection in the balance. Lots of cable factors involve the balance between the two differential signals, both in timing & amplitude. So there are very good & well known reasons why there may be a difference between USB cables. Are they audible? That depends on the implementation & is not just a black & white binary answer.

 

[ralphp@optonline]

Okay so last night I get tried of this discussion we're having here and watch a movie and then go off to bed. I wake up this morning and check in to see if any progress has been make only to find that yet another detour has been successfully thrown up by mmerrill99 with everyone falling his trick.
 
An audible ground loop is a problem. A program means that something is broken and needs to be fixed. I've had ground loop problems several times over the years and they very annoying and not very easy to correct. However they can and should be corrected and not simply with a shielded cable. Excessive RF noise so that it becomes audible is also a problem that needs to fixed.
 
What is not being said is that various level of noise exist in all electrical systems and unless those noise levels are excessive, as stated in the paragraph above, the noise and any affect that noise might have on the audio signal are all well below the levels of human hearing.
 
All of this little detour is high end audio marketing 101 - identify a "problem" and then offer a "solution". Then rinse and repeat. Works like a charm and works almost every time.
 
All DACs are electrical devices and are subject to RF noise and ground loop noise and all the other "noise" present in electrical signals. The thing is that if the DAC is properly designed and built all that noise does not make into the analog output. So let me repeat this yet again - if there is audible noise and distortion coming from the analog output of the DAC then something in the system is BROKEN and needs to FIXED. This is not a problem that replacing a USB cable will fix UNLESS it was the USB cable that was BROKEN.
 
Now mmerrill99 may or may not have a degree in electrical engineer but I'm pretty sure that he has a degree in marketing.

 

[ralphp@optonline]

Now we are getting somewhere
Most current USB DACs operate at USB 2.0 high-speed i.e 480Mbps (480MHz). There are only a couple of devices that galvanically isolate (isolate signals & grounds) at these speeds - the Intona device & the Corning optical device.

There is no current galvanic isolation chip that can be used internally in a DAC that will cater for this USB 480MHz two-way communication needed.
There are some but very few DACs that galvanically isolate after the USB receiver inside the DAC - usually when the signal has been converted to I2S & it's the slower I2S signals that are isolated. Some level of immunity to upstream noise is conferred in this arrangement but not complete immunity

So if immunity to incoming noise is the criteria for whether a USB DAC is broken or not then we can conclude that all USB DACs are currently broken - that's why I regularly asked for an example of an "unbroken DAC"

So, if we accept that USB DAcs are not immune to incoming noise, we are left with the question - what are the possible sources of this noise & can a cable have any effect on the spectrum or level of this noise?

One needs to know about the various types of noise & study the system itself in order to be able to answer this. USB signalling is a differential signalling system that also reverts to single ended for certain signals. Differential signalling is used to provide some immunity to noise but it relies on many factors in order to achieve this noise reduction. How this works in theory is that any external noise which impinges on the cable will, because of the twisted pair construction of the cable cause an equal level of disturbance on bothe USB signal lines, D- & D+ The USB receiver is a summing receiver & only voltage differences between the D+ & D- are sensed as signal. Any signal of the same voltage that appears on both D+ & D- will be cancelled out & not sensed. As you can see this cancellation relies on the amount of perfection in the balance. Lots of cable factors involve the balance between the two differential signals, both in timing & amplitude. So there are very good & well known reasons why there may be a difference between USB cables. Are they audible? That depends on the implementation & is not just a black & white binary answer.

Pure marketing garbage. Lots of fancy technical stuff that is just not relevant to the types of systems under discussion (home audio system with a computer as a source component connected to a USB DAC. The one question not asked: if there are all these problems and all this horrible noise inherent in using the USB output for audio then why are so many people happily listening to music using a computer and USB DAC. Are all those people just immune to all that noise and distortion or are they just plain deaf?
 
Just another example of the marketing mantra of "everything matters and anything is possible". Fall for it only your wallet is fat because this will surely help to keep your wallet nice and slim.

 

[mmerrill99]

It does appear that rational discussion is beyond the capability of some who are more concerned with marketing than an objective discussion.
If I knew that the discussion was above my level of knowledge I would either bow out or more likely ask questions to inform myself - what I wouldn't do is continually attack someone just because I don't understand what they are saying. But then keyboard warriors, armchair engineers & egos are the norm on audio forums - pity

 

[BiggerHead]

mmerrill99 you seemed to have slipped back into talking about noise on the actual signal toward the end of that last post.  I think someone covered that usb does have some error correction and resending capability.   I don't think you want to start discussing the 1's and 0's.   Let's hope not.

 

[ralphp@optonline]

It does appear that rational discussion is beyond the capability of some who are more concerned with marketing than an objective discussion.
If I knew that the discussion was above my level of knowledge I would either bow out or more likely ask questions to inform myself - what I wouldn't do is continually attack someone just because I don't understand what they are saying. But then keyboard warriors, armchair engineers & egos are the norm on audio forums - pity

The sad thing is that you do actually have a pretty understanding of the issues but you also have a strong need to prove that the "issue" that you've have managed to dig up using some clever google searches is a real issue and the sky is falling. Well chicken little I would like to ease your mind and tell you that the sky is staying right where it is.
 
What you very deliberately refuse to acknowledge is that the issues that you are basing your entire argument on (noise in the signal) are well known to the audio engineers who design the equipment and that competent designers address this issue when designing the equipment. Another thing that you refuse to acknowledge is that if all that nasty noise ends up having an audible effect on the analog output then something in the system is broken or operating well outside of specifications.
 
Again high end audio marketing 101.
 

  mmerrill99 you seemed to have slipped back into talking about noise on the actual signal toward the end of that last post.  I think someone covered that usb does have some error correction and resending capability.   I don't think you want to start discussing the 1's and 0's.   Let's hope not.

 
Why not, since he just loves loops, all kinds of loops - ground loops, RF noise loops, so why not logical loops.
 
The only thing missing is the pretty girl holding up the round number between rounds.
 
Wait I hear the bell. Time to start round #7 (or is it 8, 9, 10 or 25?)