Hi Bill,
According to my brief reading of the following link, streaming of digital audio via RPi with, say, for example, pulseaudio can or does rely on on TCP, rather than UDP:
At least that's what the following command would seem to imply:
load-module module-native-protocol-tcp auth-ip-acl=127.0.0.1;192.168.1.0/24
https://raspberrypi.stackexchange.c...ream-all-audio-output-from-my-pc-to-my-stereo
Maybe I'm missing something though? Haven't looked into it in detail yet.
Also found some more discussion at:
In any case, I was responding to a post that struck me as describing the robustness of TCP, rather the possible benefits of UDP.
According to my brief reading of the following link, streaming of digital audio via RPi with, say, for example, pulseaudio can or does rely on on TCP, rather than UDP:
At least that's what the following command would seem to imply:
load-module module-native-protocol-tcp auth-ip-acl=127.0.0.1;192.168.1.0/24
https://raspberrypi.stackexchange.c...ream-all-audio-output-from-my-pc-to-my-stereo
Maybe I'm missing something though? Haven't looked into it in detail yet.
Also found some more discussion at:
In any case, I was responding to a post that struck me as describing the robustness of TCP, rather the possible benefits of UDP.
Last edited:
I think you misread me. I was pointing out streaming generally doesn't use TCP. But you really are embarrassing yourself with random snips from the internet and coming no closer to a plausible theory why the Nordost switch is not pure snake oil compared to a $15 unit from any box shifter.
Bill,
I don't claim to know if the Nordost switch can do anything useful or not. I have never tried one and do not use streaming. However, I do find reports from people who stream over wired Ethernet, to the effect that such devices have been found to help sound quality, as potentially plausible. If they do help, then there there must be one or more physical mechanisms in play.
Again, this line of discussion is similar to the old arguments that power cords can't possibly make any difference for sound reproduction. I hope we all understand now that power cords can act as AC power line EMI/RFI noise filters, and such noise can intermodulate with audio, which can then in some cases produce audible effects. Sure, it means an audio device affected in that way may be deemed as 'pathological.' But its not all that uncommon to find in the real world.
Likewise, so long as people try ethernet and or USB noise filter devices and find some real benefit, then there is most likely some kind of noise problem involved.
And of course, yes, people sometimes imagine stuff that isn't real. We all know something about that too.
My only aim in responding to a particular post in the first place was to try to head off some of the wrongheaded engineers-know-everything smug superiority that has helped polarize the so-called objectivist verses the so-called subjectivists differences in the past. IOW sometimes audiophiles are right and engineers aren't fully thinking things through. IMHO better to dampen some reactionary excitation a little bit early on rather than let it grow into grow into something not good for civility. Maybe you recall how bitter some of that once became? So, maybe if in your technical wisdom you would like to help contribute to a healthy peace around here and can do better, then please be do so. I will applaud your, perhaps sometimes imperfect (if like me), efforts. Otherwise, I will speak up when it seems appropriate to me.
I don't claim to know if the Nordost switch can do anything useful or not. I have never tried one and do not use streaming. However, I do find reports from people who stream over wired Ethernet, to the effect that such devices have been found to help sound quality, as potentially plausible. If they do help, then there there must be one or more physical mechanisms in play.
Again, this line of discussion is similar to the old arguments that power cords can't possibly make any difference for sound reproduction. I hope we all understand now that power cords can act as AC power line EMI/RFI noise filters, and such noise can intermodulate with audio, which can then in some cases produce audible effects. Sure, it means an audio device affected in that way may be deemed as 'pathological.' But its not all that uncommon to find in the real world.
Likewise, so long as people try ethernet and or USB noise filter devices and find some real benefit, then there is most likely some kind of noise problem involved.
And of course, yes, people sometimes imagine stuff that isn't real. We all know something about that too.
My only aim in responding to a particular post in the first place was to try to head off some of the wrongheaded engineers-know-everything smug superiority that has helped polarize the so-called objectivist verses the so-called subjectivists differences in the past. IOW sometimes audiophiles are right and engineers aren't fully thinking things through. IMHO better to dampen some reactionary excitation a little bit early on rather than let it grow into grow into something not good for civility. Maybe you recall how bitter some of that once became? So, maybe if in your technical wisdom you would like to help contribute to a healthy peace around here and can do better, then please be do so. I will applaud your, perhaps sometimes imperfect (if like me), efforts. Otherwise, I will speak up when it seems appropriate to me.
Last edited:
Mark, you are aguing on a melting iceberg. Now I personally doubt a switch will help unless the network is chronically overloaded with teenagers streaming movies/games/pron in the house and the DAC is pathalogically bad. In a studio running AES-67 everywhere then there is a clear place for switches. I fail in other circumstances to see anything plausible.
If anyone is really worried about noise they should run fibre. Oddly no one is selling ethernet DACs with SFP NICs so you can select wire or fibre. Maybe industry will pick up on that. Or start putting dual NICs in audiophile servers so the DAC gets a dedicated 100-baseT connection.
But if a switch does help (and a testable theory is required on that) then a $15 switch will do the job. A $3000+ switch is most clearly snake oil and a company making a solution for a problem that does not exist unless of course their target is customers who relish buying the latest solution to problems that don't exist. In which case I rest my case that it should be on this thread.
I don't think Howie reads this thread, but he'll have real world experience of AES-67...
If anyone is really worried about noise they should run fibre. Oddly no one is selling ethernet DACs with SFP NICs so you can select wire or fibre. Maybe industry will pick up on that. Or start putting dual NICs in audiophile servers so the DAC gets a dedicated 100-baseT connection.
But if a switch does help (and a testable theory is required on that) then a $15 switch will do the job. A $3000+ switch is most clearly snake oil and a company making a solution for a problem that does not exist unless of course their target is customers who relish buying the latest solution to problems that don't exist. In which case I rest my case that it should be on this thread.
I don't think Howie reads this thread, but he'll have real world experience of AES-67...
Well, how does this "Resonance Synchronizer" qualify as snake oil? Price is suspiciously low ~800$
https://www.nordost.com/qrt/qpoint-resonance-synchronizer.php
The QPOINT Resonance Synchronizer emits a subtle field which manipulates all electromechanical resonances within its immediate proximity so that they resonate in unison with each other. By syncing these resonances, the QPOINT eliminates internal electrical noise, enhancing the coherency and timing that is typically lacking in even the most high-end audio systems.
In order to accommodate its effects to all of the various electronics in the industry, including but not limited to solid state or tube amplifiers, digital or analog source components, traditional line stages or modern digital processors, and even to tailor results to the personal preferences of the user, the QPOINT has been designed with two modes. Mode I and II each generate unique resonances, which emphasize different musical characteristics as they lower the noise floor. Users can switch from Mode I to Mode II instantaneously to evaluate which sonic properties fit the needs of each unique electronic device in their hi-fi system.
The QPOINT is modular in nature. Its effects are cumulative and are best used alongside other Nordost QRT products. Nordost’s QPOINT Resonance Synchronizer ensures your electronics are all working in concert, in order to enhance the coherency, articulation, and musical structure of your system.
Only positive reviews!
https://www.theabsolutesound.com/articles/nordost-qrt-qpoint-and-qrt-qsource-resonance-synchronizer/
https://www.monoandstereo.com/2019/06/nordost-qpoint-resonance-synchronizer.html
https://hifiplus.com/articles/nordost-qpoint-harmonisers-and-qsource-power-supply/
https://www.soundstageaustralia.com...ance-synchroniser-qsource-linear-power-supply
http://bostonconcertreviews.com/qua...es-ac-supply-power-distribution-audio-system/
ps. I had a day off from work...
https://www.nordost.com/qrt/qpoint-resonance-synchronizer.php
The QPOINT Resonance Synchronizer emits a subtle field which manipulates all electromechanical resonances within its immediate proximity so that they resonate in unison with each other. By syncing these resonances, the QPOINT eliminates internal electrical noise, enhancing the coherency and timing that is typically lacking in even the most high-end audio systems.
In order to accommodate its effects to all of the various electronics in the industry, including but not limited to solid state or tube amplifiers, digital or analog source components, traditional line stages or modern digital processors, and even to tailor results to the personal preferences of the user, the QPOINT has been designed with two modes. Mode I and II each generate unique resonances, which emphasize different musical characteristics as they lower the noise floor. Users can switch from Mode I to Mode II instantaneously to evaluate which sonic properties fit the needs of each unique electronic device in their hi-fi system.
The QPOINT is modular in nature. Its effects are cumulative and are best used alongside other Nordost QRT products. Nordost’s QPOINT Resonance Synchronizer ensures your electronics are all working in concert, in order to enhance the coherency, articulation, and musical structure of your system.
Only positive reviews!
https://www.theabsolutesound.com/articles/nordost-qrt-qpoint-and-qrt-qsource-resonance-synchronizer/
https://www.monoandstereo.com/2019/06/nordost-qpoint-resonance-synchronizer.html
https://hifiplus.com/articles/nordost-qpoint-harmonisers-and-qsource-power-supply/
https://www.soundstageaustralia.com...ance-synchroniser-qsource-linear-power-supply
http://bostonconcertreviews.com/qua...es-ac-supply-power-distribution-audio-system/
ps. I had a day off from work...
The comments section on this review are getting quite fruity https://www.stereophile.com/content/nordost-qnet-network-switch-qsource-linear-power-supply#comments . A couple of the objectivists should have thought before posting as well there 
Interestingly one hard line subjectivist is claiming the uptone etherregen is PROOF that ethernet cleaning works. I had a look a the 'white paper' on that https://cdn.shopify.com/s/files/1/0...enson_EtherREGEN_white_paper.pdf?v=1583429386 All I can say is that the quality of white papers is going downhill. Talking about ground plane noise for ethernet where the signals are galvanically isolated and the cable has no ground is a very odd way of justifying their product. But at $640 it's at least merelyoverpriced rather than obscenely overpriced.
Interestingly one hard line subjectivist is claiming the uptone etherregen is PROOF that ethernet cleaning works. I had a look a the 'white paper' on that https://cdn.shopify.com/s/files/1/0...enson_EtherREGEN_white_paper.pdf?v=1583429386 All I can say is that the quality of white papers is going downhill. Talking about ground plane noise for ethernet where the signals are galvanically isolated and the cable has no ground is a very odd way of justifying their product. But at $640 it's at least merelyoverpriced rather than obscenely overpriced.
Hi Bill,
Thank you for your thoughtful technical comments. Thinking about it some more seems to me I read somewhere else here in the forum that some people have had problems with wired ethernet affecting dac sound. IIRC one guy said he tried using fiber for isolation, but still had a problem. IIUC the topology that may tend to work is to put the ethernet isolator just before the streamer. The failure of a fiber link to help solve the problem suggested to me that noise might be produced in the network equipment at the dac end of the fiber connection. However there was another guy who found that inserting some public domain diy filter in the wired ethernet connection between two pieces of network gear helped his problem. Therefore, anecdotal reports would seem to suggest that the cause of problems is not always at the same point every ethernet system. May vary with particular equipment brands and or models. Don't recall what brand ethernet isolator was claimed to help in that discussion.
Regarding pricing of high end audio equipment, as we have discussed on other occasions sales volumes for high end audio products tends to be very low. Some products may only ever sell 100 or maybe 200-300 units over the lifetime of the product. Therefore retail markups tend to be a bit higher than the usual 3-5 times seen in consumer markets. For high end the markup often has to be at least 6 times. So the total incremental cost of manufacturing, retail packaging, manufacturer marketing, apportioned engineering development costs, etc., for a $3,000 box might be around $500 per unit. Given the low sales volumes typical in that market, doubtful they are getting rich. An EE could probably make more money working for big company, plus get benefits a small company often can't offer.
Last point: I didn't object to the Nordost device inclusion in this thread, that wasn't my point. I only commented when I felt there was some over-confidence expressed to the effect that it was impossible for the Nordost unit to help with sound problems. From reading what you said a couple of posts ago, you have doubts but don't claim its impossible. Good enough for me. Have a nice day
Thank you for your thoughtful technical comments. Thinking about it some more seems to me I read somewhere else here in the forum that some people have had problems with wired ethernet affecting dac sound. IIRC one guy said he tried using fiber for isolation, but still had a problem. IIUC the topology that may tend to work is to put the ethernet isolator just before the streamer. The failure of a fiber link to help solve the problem suggested to me that noise might be produced in the network equipment at the dac end of the fiber connection. However there was another guy who found that inserting some public domain diy filter in the wired ethernet connection between two pieces of network gear helped his problem. Therefore, anecdotal reports would seem to suggest that the cause of problems is not always at the same point every ethernet system. May vary with particular equipment brands and or models. Don't recall what brand ethernet isolator was claimed to help in that discussion.
Regarding pricing of high end audio equipment, as we have discussed on other occasions sales volumes for high end audio products tends to be very low. Some products may only ever sell 100 or maybe 200-300 units over the lifetime of the product. Therefore retail markups tend to be a bit higher than the usual 3-5 times seen in consumer markets. For high end the markup often has to be at least 6 times. So the total incremental cost of manufacturing, retail packaging, manufacturer marketing, apportioned engineering development costs, etc., for a $3,000 box might be around $500 per unit. Given the low sales volumes typical in that market, doubtful they are getting rich. An EE could probably make more money working for big company, plus get benefits a small company often can't offer.
Last point: I didn't object to the Nordost device inclusion in this thread, that wasn't my point. I only commented when I felt there was some over-confidence expressed to the effect that it was impossible for the Nordost unit to help with sound problems. From reading what you said a couple of posts ago, you have doubts but don't claim its impossible. Good enough for me. Have a nice day
Last edited:
Just to completely clarify my position, I believe it is impossible for the Nordost to improve on any L2 switch in the scenario where such a switch is needed. Thinking about this, and given that most broadband routers have an L2 switch in the problems would start to arise where there are not enough ports and a L1 hub is put in. But short of packet loss or a really poor implementation at the receiver I'm at a loss to find a testable mechanism.
Well, suppose we set out to see if we could measure a difference between Nordost and an L2 switch? Noise affecting a dac would often tend to manifest in one or both of two (or possibly more) measurable ways:
(1) a change in FFT noise floor (hopefully measured with and without audio signals), and or
(2) a change is spectral line skirt widths of high resolution FFTs.
If there is a measurable difference which correlates well with reports of changed sound, then perhaps there might be something to the claims?
What if someone you trust posted measurements of such a difference in measurements, say, perhaps KSTR, might that influence your thinking?
Note: One may recall that KSTR once showed that two identical digital audio data files stored on two different storage devices can come out of a USB dac as measurably different in the analog domain depending on the particular storage device. Of course, one might expected that to be impossible before proof and an explanation was presented. Once understood though, it becomes quite plausible...
(1) a change in FFT noise floor (hopefully measured with and without audio signals), and or
(2) a change is spectral line skirt widths of high resolution FFTs.
If there is a measurable difference which correlates well with reports of changed sound, then perhaps there might be something to the claims?
What if someone you trust posted measurements of such a difference in measurements, say, perhaps KSTR, might that influence your thinking?
Note: One may recall that KSTR once showed that two identical digital audio data files stored on two different storage devices can come out of a USB dac as measurably different in the analog domain depending on the particular storage device. Of course, one might expected that to be impossible before proof and an explanation was presented. Once understood though, it becomes quite plausible...
No I don't recall those measurements and a single test as you well no is not proof.
Fearing my sanity I did look at the ASR review of the uptone etherregen. https://www.audiosciencereview.com/.../uptone-audio-etherregen-switch-review.10232/ diddly squat difference on the tests done and nothing that would point to a mechanism. And a lot of confusion over the protocols being used again...
Fearing my sanity I did look at the ASR review of the uptone etherregen. https://www.audiosciencereview.com/.../uptone-audio-etherregen-switch-review.10232/ diddly squat difference on the tests done and nothing that would point to a mechanism. And a lot of confusion over the protocols being used again...
I'm failing to remember I ever did something like that with a positive result but mabye I'm getting old, care to dig that one up?
Found this post and some others leading up to it: https://www.diyaudio.com/community/...der-as-sufficient.325234/page-35#post-5518826
Also this comment on your post by Scott Wurcer:
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5518969
Some earlier posts:
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5520450
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5520566
Also this comment on your post by Scott Wurcer:
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5518969
Some earlier posts:
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5520450
https://www.diyaudio.com/community/...ou-consider-as-sufficient.325234/post-5520566
The point is that the problem was not found to be digital with the ones and zeros. There was an analog effect that some people never bothered to consider, or they assumed such a thing was impossible. Doesn't sound like such a weak correlation to me.
Again, its like the oversimplification of power cords: can't be any effect because its only very small amount of the total resistance going back to the power generator. Or, its impossible because ethernet is isolated at both ends. Blah, blah, blah...
Why do we always tend to over-simplify the physics when people claim they hear something? How about we try to find a more complete model of the system and see if the claims make sense in that context? If we are too lazy, fine. Doesn't mean we have done to the work to conclude its impossible.
Regarding your remark about people being banned, so what? Scott Wurcer commented. Are you suggesting there was something wrong with that?
Again, its like the oversimplification of power cords: can't be any effect because its only very small amount of the total resistance going back to the power generator. Or, its impossible because ethernet is isolated at both ends. Blah, blah, blah...
Why do we always tend to over-simplify the physics when people claim they hear something? How about we try to find a more complete model of the system and see if the claims make sense in that context? If we are too lazy, fine. Doesn't mean we have done to the work to conclude its impossible.
Regarding your remark about people being banned, so what? Scott Wurcer commented. Are you suggesting there was something wrong with that?
actually the point was there that if you change the file (leading zeros) something odd happens on a very poor DAC. I still fail to see how that relates to a snake oil switch vs a cheap L2 switch. The packets in that case come from a server and the switch is broken if it re-assembles them differently from Rx to Tx
Why are you so desperate to find an effect where logic suggests there should not be one? Are top studios shouting the praise of magic switches? And why are you throwing random arguments against other snake oil against this one device. What do power cords have to do with L2 ethernet? Argue the point or admit you don't have a credible theory to work on.
You are having a serious sense of humour failure today. I commented on the banned contributors as I thought it was amusing how they all gathered like flies to dung on threads like that. I didn't mention Scott or any of the more serious contributions there.
Why are you so desperate to find an effect where logic suggests there should not be one? Are top studios shouting the praise of magic switches? And why are you throwing random arguments against other snake oil against this one device. What do power cords have to do with L2 ethernet? Argue the point or admit you don't have a credible theory to work on.
You are having a serious sense of humour failure today. I commented on the banned contributors as I thought it was amusing how they all gathered like flies to dung on threads like that. I didn't mention Scott or any of the more serious contributions there.
Bill, IMHO there are enough independent reports of problems with wired ethernet sound quality to take at least some of the reports seriously. That's all. I can't and don't attribute them all to hallucination. IMHO KSTR is on the right path in general to suspect analog noise mechanisms that are not obvious in superficial analysis.
Also, we know we can measure thresholds of audibility for various things. By definition that means that 50% of the population can hear below the threshold and 50% can't. Suppose there is a threshold of audibility for the kind of effects noisy ethernet can have on some dacs. Doesn't mean you will hear it. Doesn't mean someone who hears it in their system will hear it on some unfamiliar system. Yet an effect may still be quite real. For people that actually do hear real problems that you might not, it is extremely insulting and disrespectful for you to denounce them as fools (or the implied equivalent) in threads like this. Its also a disservice to denounce companies that are seriously trying to help.
Yes, I know its difficult or impossible to know who is sincerely trying to help, who is actually helping, and or those who are trying to take every sucker for as much money as possible. Doesn't mean its right to always assume the latter just because off the top of your head you can't see why something might be a problem. What if you turn out to be wrong? That is possible you know.
Also, we know we can measure thresholds of audibility for various things. By definition that means that 50% of the population can hear below the threshold and 50% can't. Suppose there is a threshold of audibility for the kind of effects noisy ethernet can have on some dacs. Doesn't mean you will hear it. Doesn't mean someone who hears it in their system will hear it on some unfamiliar system. Yet an effect may still be quite real. For people that actually do hear real problems that you might not, it is extremely insulting and disrespectful for you to denounce them as fools (or the implied equivalent) in threads like this. Its also a disservice to denounce companies that are seriously trying to help.
Yes, I know its difficult or impossible to know who is sincerely trying to help, who is actually helping, and or those who are trying to take every sucker for as much money as possible. Doesn't mean its right to always assume the latter just because off the top of your head you can't see why something might be a problem. What if you turn out to be wrong? That is possible you know.
We must also remember that correlation is not causality. Whenever someone reports a problem spends $$$$$ and the problem goes away there are so many confounders you cannot draw any valid conclusion.
As a side comment; calling a L2 device "switch" is an old sales critter language that wanted to
distance themself from the crowd, Calpana (*) was the name of the company. Most of all others
could not resist but renamed their devices "switch" as they did not want to sound like "old fashion"
The correct name is "bridge" ( any device working on L2 in the OSI stack is a bridge) A decice
that works on level 1 is a "hub" a device thet works on L3 is a router.
The common household gateways is really a combination of a router bridge and a level 7
component "network address translator ( NAT)"
Switch is stuff that is used to make or break power cables.
(*) calpana was a pioneer with "cut through " where a L2 packet was forwarded as soon as the ethernet
header destination was received, the rest of the packet was clocked out with only about 10 bytes delay,
checksum errors and collisions was transmitted which caused problems. The technology was later
abandoned as 100 and 1G ethernet became common.
distance themself from the crowd, Calpana (*) was the name of the company. Most of all others
could not resist but renamed their devices "switch" as they did not want to sound like "old fashion"
The correct name is "bridge" ( any device working on L2 in the OSI stack is a bridge) A decice
that works on level 1 is a "hub" a device thet works on L3 is a router.
The common household gateways is really a combination of a router bridge and a level 7
component "network address translator ( NAT)"
Switch is stuff that is used to make or break power cables.
(*) calpana was a pioneer with "cut through " where a L2 packet was forwarded as soon as the ethernet
header destination was received, the rest of the packet was clocked out with only about 10 bytes delay,
checksum errors and collisions was transmitted which caused problems. The technology was later
abandoned as 100 and 1G ethernet became common.
Yes the Broadband router is indeed L3 but here we are talking about a specific L2 device that 'improves quality'. I'd personally argue that 'bridge' is more confusing than 'switch' as you need L3 to bridge 2 networks together, but that's being pedantic
As a managed cisco 48 port L2 device is a couple of hundred dollars and a medical grade 4 port is similar order and there is no plausible theory as to what could go wrong I am still happy to part the switch in question here under 'fleecing'. And as long as the punter realises he is being fleeced to score him points with his mates so be it.
As a managed cisco 48 port L2 device is a couple of hundred dollars and a medical grade 4 port is similar order and there is no plausible theory as to what could go wrong I am still happy to part the switch in question here under 'fleecing'. And as long as the punter realises he is being fleeced to score him points with his mates so be it.
- Status
- Not open for further replies.