Ethernet switches for audio - Part B: why a regular switch will suffice.

[sir sanders zingmore]

29 minutes ago, dbastin said:

 

 

This could take us really well off topic, and I'd rather not spiral this thread out of control.  I am by far not an expert in that subject, so I'm just going from life experiences.  Try googling "science can't explain" ... plenty of potential examples but again the problem of determining if its truth or BS.  I came across this ... https://www.livescience.com/11345-top-ten-unexplained-phenomena.html

 

I don't wish to invoke an emotional response .. emotion is probably an example of something we all experience that science can't fully explain at this point in time.  Just a thought ... perhaps we seek evidence to quell fear of the unknown?

 

Sorry this is philosophical, these are thought provoking quotes:

 

https://www.goodreads.com/quotes/tag/unknown

https://www.awakenthegreatnesswithin.com/20-inspirational-quotes-on-the-unknown/

Oh…..I thought we were talking about audio in general and ethernet switches in particular

 

 

[Dropbear67]

Try googling "science can't explain"

1 hour ago, dbastin said:

.  Try googling "science can't explain" ...

 

No thanks, because you're no longer in the realm of rational repeatable or measurable .... I'll stick to my "noisy" $50 switch which delivers bits from A to B quite well without  affecting "midrange definition"

[Stereophilus]

2 hours ago, davewantsmoore said:

ANY measurements.   Anything.... absolutely any evidence of any difference between the signals will suffice.

 

All that needs to be done is measure the difference between the two signals.   If the signals are different the we CAN measure and show that.     If we're measuring at the wrong place in the system ..... then measure in a different place..... until a difference is shown.

 

At this point.... many people will shrug and say "well, maybe there's things we can't measure".

 

No... the sigal leaving my DAC, or amp, or whatever is just an electrical signal with current and voltage varying with time.    It is quite simple (relative to other things science concerns itself with).

 

Now, once we have a meaasured difference..... imagine that difference is teeny tiny.   At that point, "science would say".... that's going to be inaudible....   and at that point you could pull out "perhaps science cannot explain why this IS audible" ..... and the answer to that might be "yes, maybe".

Good summary Dave. Thank you.

 

My perspective is that DACs are highly sensitive to RFI and EMI.  These issues occur beyond the threshold of human hearing (in terms of both frequency and amplitude), but act at the point of conversion from D to A to alter the analog waveform produced, thus becoming audible. 

 

I am talking about the effects of jitter.  But to measure these issues fully we need our network, streamer, DAC and analog gear all connected up.  The jitter inducing spuria travel along any connection between equipment, especially through power leads and, seemingly, network connections.

 

The trap we fall into with Audio is to disregard the sensitivity and vulnerability of the DAC.  We assume it can handle the spuria thrown at it.  We assume that digital signals that measure the same and are "bit perfect" are therefore not to blame.  Which indeed they are not, but the connections along which they travel are conduits for spuria into the DAC.

 

A final area of concern is how our test equipment interacts with our spuria sometimes obscuring our results and findings.

 

So if you are measuring the effects of network switches on audio reproduction, measure at the DAC output, with a full system connected, using care to avoid a bypass loop through your measuring equipment.

[sir sanders zingmore]

As I recall, ASR measured the impact on a DAC of an uptone etherRegen thingie. 
Take a stab as to what impact it had?

[davewantsmoore]

4 minutes ago, Stereophilus said:

My perspective is that DACs are highly sensitive to RFI and EMI.

Yes.  Can be.

 

... but you can see this in their output.

 

You can see it in their output at frequencies you can hear.... you can see it in their output at frequencies light-years outside of what you can hear.

 

 

4 minutes ago, Stereophilus said:

But to measure these issues fully we need our network, streamer, DAC and analog gear all connected up.

Yes, if you bench test a DAC in a lab, it is possible that there are other things affecting it once you put it in a system.

 

4 minutes ago, Stereophilus said:

The jitter inducing spuria travel along any connection between equipment, especially through power leads and, seemingly, network connections.

It's going to depend a whole lot on the DAC as to what it makes of such things.

 

It seems super important to ruthlessly isolate sensitive circuits from this sort of thing... and keep the system a simple as possible, with as few cables as possible, etc.    I've had as many boxes as anyone in my system at times over the years (think like 15+ power cables) .... and now I am as far away from that as I can get.

 

4 minutes ago, Stereophilus said:

The trap we fall into with Audio is to disregard the sensitivity and vulnerability of the DAC.  We assume it can handle the spuria thrown at it. 

I don't think anyone should disregard or assume anything.

 

4 minutes ago, Stereophilus said:

So if you are measuring the effects of network switches on audio reproduction, measure at the DAC output

1000%

3 minutes ago, sir sanders zingmore said:

As I recall, ASR measured the impact on a DAC of an uptone etherRegen thingie. 
Take a stab as to what impact it had?

I don't think the tests ASR do are necessarily enough

 

.... but, OTOH.... it's quite possible to do the needful  (guess why it isn't done)

[sir sanders zingmore]

27 minutes ago, davewantsmoore said:

I don't think the tests ASR do are necessarily enough

The “jitter” catch-all was mentioned. They measured that. 

What do you think they should have measured that might give a clue as to why an Ethernet device would improve the sound (or change the sound at all)?

Edited June 25, 2021 by sir sanders zingmore
Added the mark of questioning

[Assisi]

It seems to me that the posters to this thread who have doubts etc as to whether Ethernet audio switches and associated items can have a beneficial listening outcome expect some sort of substantive proof of the outcome.  I don’t have the proof nor am I able to ascertain the proof what ever it might be.  To me it is all about the hearing, the listening and importantly the overall quality of the ultimate emotional experience. 

 

 

Have any of the posters who doubt my experience and that of many others ever actually listened to a system with audio switches etc.?  As somebody said to me today the maybe the only way to resolve some of the matters is to have a GTG.  Any body interested?

       

I am not sure how to do such a GTG.  The complication of a GTG would be to do a comparison of with and without the audio switches and other network bits etc.  It is not a simple as changing a DAC for another DAC.  The real difficulty is to put all the components and cables back after they have been removed.  A couple of times I have removed just one switch to demonstrate it elsewhere in another system.  In the past getting everything back working again in my system has been a stressful nightmare.  It is not something I look forward to doing.  Even when everything is connected correctly then there is settling in time which can be at least day.

John

Edited June 25, 2021 by Assisi

[Stereophilus]

5 hours ago, sir sanders zingmore said:

The “jitter” catch-all was mentioned. They measured that. 

What do you think they should have measured that might give a clue as to why an Ethernet device would improve the sound (or change the sound at all)?

I'm disappointed Amir didn't do a full sweep of the audible spectrum for jitter.  Low frequencies are most affected by jitter to my ears.  He only did test tones at 12kHz and 20kHz.  I really think a spectral analysis would be more useful.  Maybe throw in some impulse responses as well. 

 

It was almost as if Amir went into the testing with a predetermined idea he would find nothing, so he skimmed over it.  As such, I really don't think you can say he excluded jitter as a cause for percieved differences.

[sir sanders zingmore]

1 minute ago, Stereophilus said:

Low frequencies are most affected by jitter to my ears.

 

how do you know this?

[Stereophilus]

1 minute ago, sir sanders zingmore said:

 

how do you know this?

It's in the quote... "to my ears"

[Stereophilus]

2 minutes ago, sir sanders zingmore said:

 

how do you know this?

And regardless of my impressions about where to measure it wasn't measured, so we don't know.

[sir sanders zingmore]

1 minute ago, Stereophilus said:

It's in the quote... "to my ears"

 

I read that, I guess what I was asking is how do you know that whatever it is you are hearing at low frequencies is caused by jitter (as opposed to something else)?

 

 

[sir sanders zingmore]

2 minutes ago, Stereophilus said:

And regardless of my impressions about where to measure it wasn't measured, so we don't know.

 

Sure, but he measured jitter at the same frequency as he does for all the DACs he measures. The comment below from you implies that he was somehow doing something different or misleading for the etherRegen test

9 minutes ago, Stereophilus said:

It was almost as if Amir went into the testing with a predetermined idea he would find nothing, so he skimmed over it.  As such, I really don't think you can say he excluded jitter as a cause for percieved differences.

 

Edited June 25, 2021 by sir sanders zingmore

[Stereophilus]

Also, how was the test tone network signal fed to the ER? From the APx 500 directly? If it was then this is hardly a good way to prove a network switch rejects network borne (jitter inducing) spuria.  I don't know how this was done.

Edited June 25, 2021 by Stereophilus

[Stereophilus]

2 minutes ago, sir sanders zingmore said:

 

Sure, but he measured jitter at the same frequency as he does for all the DACs he measures. The comment below from you implies that he was somehow doing something different or misleading for the etherRegen test

I don't read much at all on his website tbh.  If that is his normal routine then I'm doubly annoyed.  As I said, there was much more he could have looked at, and should have looked at before dismissing the ER.

[sir sanders zingmore]

2 minutes ago, Stereophilus said:

Also, how was the test tone network signal fed to the ER? From the AR-555 directly? If it was then this is hardly a good way to prove a network switch rejects network borne (jitter inducing) spuria.  I don't know how this was done.

 

"EtherRegen Audio Measurements
I powered on the Matrix i and selected Network input. I connected its balanced XLR output to my analyzer for the tests you are about to see (at the end, I also tested RCA but made no difference). First our usual dashboard of 1 kHz with just my cheap, generic TP-LINK gigabit switch feeding the Matrix i via Roon streaming:"

[sir sanders zingmore]

7 minutes ago, Stereophilus said:

I don't read much at all on his website tbh.  If that is his normal routine then I'm doubly annoyed.  As I said, there was much more he could have looked at, and should have looked at before dismissing the ER.

My understanding of the jitter test is limited but I don't think it's correct to see it as a frequency test at 12khz.

 

 

from stereophile (I admit I don't fully understand it)

 

the J-Test, where J stands for jitter, was developed in the mid-1990s by the late Julian Dunn of PrismSound to investigate the jitter rejection of digital datalinks in which the clock is embedded in the audio data: the balanced AES/EBU or AES3 link, for example, or the unbalanced S/PDIF link. The test signal comprises a high-level tone at exactly one-quarter the sample frequency, or Fs/4, to which is added a squarewave at the level of the least significant bit (LSB), at Fs/192. With the twos-complement PCM encoding used by CD data, this low-level squarewave exercises all the bit transitions simultaneously, which is very much the worst case for stimulating jitter. The high-level, high-frequency tone is thus modulated by the jitter and sidebands spaced at the frequency of the Fs/192 squarewave, and its harmonics will appear in the reconstructed analog signal's noise floor.

[Stereophilus]

Is it a J-test? He doesn’t tell us in the report…. If it is a J-test, fair enough, but even then, I’ve seen quite a few questions raised about J-test validity.  Most of what I read suggests jitter phenomena are complicated and not easily simplified into 1 test routine.

[sir sanders zingmore]

15 minutes ago, Stereophilus said:

Is it a J-test? He doesn’t tell us in the report…. If it is a J-test, fair enough, but even then, I’ve seen quite a few questions raised about J-test validity.  Most of what I read suggests jitter phenomena are complicated and not easily simplified into 1 test routine.

I think we can assume it's a J-test 

 

I don't know if it's sufficient, but most of the stuff that suggests it isn't is (coincidentally?) written by people selling jitter-related stuff

[Stereophilus]

5 minutes ago, sir sanders zingmore said:

I don't know if it's sufficient, but most of the stuff that suggests it isn't is (coincidentally?) written by people selling jitter-related stuff

That’s a bit cynical.  Jitter is widely acknowledged as a complex phenomena.  Distilling complex phenomena into what could be argued as an overly simplistic measurement remains a point of reasonable contention.

[BugPowderDust]

4 hours ago, Assisi said:

I am not sure how to do such a GTG.  The complication of a GTG would be to do a comparison of with and without the audio switches and other network bits etc.

John, I believe you don’t live too far from me.
 

I have a 25+ year career in networking and would be happy to help facilitate something like this, as well as ensure your system gets reinstated to its original state afterwards.

 

I’m a total sceptic but would still entertain the experience if it meant we could share some knowledge in both directions. I’d be very happy to explain how the networking piece works, down to the frame level if anyone wants a crash course in how tcp/ip works too.

 

If you’re really keen, PM me and we can chat. 

[MLXXX]

22 hours ago, Stereophilus said:

The trap we fall into with Audio is to disregard the sensitivity and vulnerability of the DAC.  We assume it can handle the spuria thrown at it.  We assume that digital signals that measure the same and are "bit perfect" are therefore not to blame.  Which indeed they are not, but the connections along which they travel are conduits for spuria into the DAC.

 

If the above is a possible trap, the converse would be a possible trap, i.e. a fear that well designed equipment will almost certainly fall foul of "spuria".  As davewantsmoore has mentioned:

 

22 hours ago, davewantsmoore said:

22 hours ago, Stereophilus said:

My perspective is that DACs are highly sensitive to RFI and EMI.

Yes.  Can be.

 

... but you can see this in their output.

 

You can see it in their output at frequencies you can hear.... you can see it in their output at frequencies light-years outside of what you can hear.

 

 

As I have already suggested, setting up two [identical model] stream receivers and recording their audio output with two identical ADCs  would enable capture and analysis to be done. One stream receiver would be fed through a standard ethernet switch, and the other through a reputedly "enhanced" ethernet switch.

 

Then if a passage of received streamed audio was identified as different as between the two recordings, it could be played back to a panel of individuals to evaluate. As an initial threshold question it would need to be established whether the difference in the recordings was audible. If that were established, people could be asked to go on and evaluate the seriousness of the difference (e.g. "hardly noticeable", "noticeable but not annoying", "mildly annoying", "irritating", etc).

 

 

Let me ask this question:  for those people who have heard differences, is it a continuous effect noticeable almost all the time, or is it sporadic?

Edited June 26, 2021 by MLXXX

[dbastin]

22 hours ago, sir sanders zingmore said:

Oh…..I thought we were talking about audio in general and ethernet switches in particular

Sorry, I seem to have misunderstood your question which I thought was in the context of my statement "This can apply to many things, not just audio gear performance."

 

An example in answer to your more specific question:

Why does daisy chaining ethernet switches often provide a perceived improvement in sound quality?  I gather science and engineering haven't figured this out yet.

 

My general point is maybe we should accept that some things cant yet be explained or proven.

 

phenomenon

a fact or situation that is observed to exist or happen, especially one whose cause or explanation is in question.

[sir sanders zingmore]

1 minute ago, dbastin said:

Sorry, I seem to have misunderstood your question which I thought was in the context of my statement "This can apply to many things, not just audio gear performance."

 

Looks like I'm the one who needs to apologise @dbastin. I completely missed the "not just audio gear" bit….. sorry

[dbastin]

This post is about a product that would likely perform very well as a switch that is non audio grade.

Edited January 22, 2022 by dbastin