Author Topic: Switch! in line  (Read 4774 times)

Offline Crankshaft

Re: Switch! in line
« Reply #30 on: July 16, 2021, 12:31:35 PM »
The only place where digital actually 'exists' is in the conceptual realm.

Everything else is physical / analogue.

Not sure If I'm using the right words, making a fool of myself or both.

Offline chrisc

  • Trade Count: (+180)
  • AVForums Grandmaster
  • *****
  • Posts: 13,747
  • Total likes: 1101
  • Condition your mains to condition your ears
Re: Switch! in line
« Reply #31 on: July 16, 2021, 01:04:20 PM »
Sorry, not good with computers and diagrams:

All wired connection

Incoming Fibre
        l
Router
        l
ROON ROCK------External HDD
         l
  CAT 5 from remote position
         l
 ( Local to Sound System)

CROAK SWITCH-------Apple Airport Express (As Wi Fi Extender for Roon Remote)
          l
WATTS SWITCH
          l
Brinkmann Streamer/DAC



What is the purpose of a second (Watts) switch ahead of the first (Croak) switch?   Cannot one switch do the whole job?   The sole purpose of a switch is to expand the number of available LAN ports if the router has insuffient

My Mikrotik router has 5 ports, 1 WAN and 4 LAN

1 - Fibre ONT (WAN port)
2 - To the unmanaged switch
3 - To Apple Airport Extreme somewhere else in the house
4 - To a DVR
5 - To my neighbour's house as she shares my internet
Music is the shorthand of emotion

Offline Nikkel

Re: Switch! in line
« Reply #32 on: July 16, 2021, 01:05:43 PM »
What is the purpose of a second (Watts) switch ahead of the first (Croak) switch?   Cannot one switch do the whole job?   The sole purpose of a switch is to expand the number of available LAN ports if the router has insuffient

My Mikrotik router has 5 ports, 1 WAN and 4 LAN

1 - Fibre ONT (WAN port)
2 - To the unmanaged switch
3 - To Apple Airport Extreme somewhere else in the house
4 - To a DVR
5 - To my neighbour's house as she shares my internet
:popcorn:

Offline chrisc

  • Trade Count: (+180)
  • AVForums Grandmaster
  • *****
  • Posts: 13,747
  • Total likes: 1101
  • Condition your mains to condition your ears
Re: Switch! in line
« Reply #33 on: July 16, 2021, 01:07:28 PM »
She does pay me R250 a month, which is about 1/4 of the fee, but they could not afford a stand-alone service

I ran a Cat 6 cable through the roof, in a pipe down the wall, 20m across to her house and back in their roof
Music is the shorthand of emotion

Offline naboo

Re: Switch! in line
« Reply #34 on: July 16, 2021, 01:17:09 PM »
I don't know why I struggle to "get" this. I can accept that inside the DAC, from when the information or data hits the DAC *device* , there can be things that affect the signal from that point onwards (even before the DAC chipset).

But the role of the switch is to get the information - the data - the digital file that IS the music, to the DAC device. You could copy the thing there, or read it from a USB stick, for all I care, the digital representation of the artefact must be same, until the DAC starts processing it. If it 30 016k on your pc, then 30 016k must land at the DAC (whether copied in one go, or streamed over a network). I think I've made this point, so I'll stop now - clearly it is not as simple as I'm trying to make it. :nfi:

Offline chrisc

  • Trade Count: (+180)
  • AVForums Grandmaster
  • *****
  • Posts: 13,747
  • Total likes: 1101
  • Condition your mains to condition your ears
Re: Switch! in line
« Reply #35 on: July 16, 2021, 01:21:59 PM »
The "bits" become purified and shaped

Obviously nonsense, but for an august and experienced listener like Johan there must be something else going on that we have yet to determine.  My cheapie switch (R110 from SCOOP) had a rudimentary PSU, so I made a better one, fully regulated and stablilised.  This made 0% difference
Music is the shorthand of emotion

Offline johanhp

Re: Switch! in line
« Reply #36 on: July 16, 2021, 02:06:10 PM »
I don't know why I struggle to "get" this. I can accept that inside the DAC, from when the information or data hits the DAC *device* , there can be things that affect the signal from that point onwards (even before the DAC chipset).

But the role of the switch is to get the information - the data - the digital file that IS the music, to the DAC device. You could copy the thing there, or read it from a USB stick, for all I care, the digital representation of the artefact must be same, until the DAC starts processing it. If it 30 016k on your pc, then 30 016k must land at the DAC (whether copied in one go, or streamed over a network). I think I've made this point, so I'll stop now - clearly it is not as simple as I'm trying to make it. :nfi:

This reminds me of myself a few years ago. I since realized that open mindedness is doing a lot more for my music enjoyment than understanding. My process used to Research -> Understanding -> Audition -> Implementation. This has deprived me from some significant improvements to my system. I now do it more like this Research / Listen to people I trust -> Audition (Sometimes optional) -> Implement -> Understand. This has been much more rewarding and more fun to be honest. Understanding in the end is optional because the improvements that something can bring to your system does not depend on your understanding of it. I have no idea how an audiophile switch will bring improvements and I do understand Network protocols and networking quite well. But if the people who's opinion I trust in this thread say it does I am more than willing to explore the possibility.

Offline kinosfronimos

  • Trade Count: (+9)
  • AVForums Veteran
  • ***
  • Posts: 705
  • Total likes: 84
  • I've found music to be the therapy of choice
Re: Switch! in line
« Reply #37 on: July 16, 2021, 02:07:49 PM »
Just to throw my confusion into this thread:
Does this 'analog environment noise' change 0's to 1's and 1's to 0's arbitrarily?
I can understand EM signals affecting the analogue path, but not the digital one to this extent?

Sent from my MI CC 9 using Tapatalk

Monitor Audio Speakers: BR6's, BX-FX's, RX-C; ASW100 Cambridge Audio: Asure 540R V3 & 540D V2; Samsung LED: 46EH5300; Xonar U3 DAC; Bose SoundTrue Around-Ear I & Sony WH-1000M3 Headphones; Denon DM41 mini-Hi-Fi

Offline pwatts

Re: Switch! in line
« Reply #38 on: July 16, 2021, 02:28:17 PM »
Concerning PTP, quite right it has no use for stereo audiophiles. It is meant for synchronizing networks to a very high accuracy, compensating for packet residency time across multiple switches. It is used for video, but also audio where you need extremely low and precise latency: a good example is live audio where several microphones and instruments have to go to a mixing console. Using Ethernet with conventional RTP (or worse, TCP) and best-effort traffic is never going to fly with its non-deterministic nature and high latency.

On its own, PTP is not really enough, so it is added to another collection of protocols to make it usable for other applications. The first will be a transport protocol. The most precise will work on layer2, not layer3, and some (IEEE1722) can be used to transport several media streams (audio or video), with their own embedded media clock. These media clocks are independent/asynchronous to the PTP 'grandmaster' clock, but PTP is used to synchronize all the multiple endpoints together when the media clocks are recovered. If the PTP clock is bad, the media clock is bad. The grandmaster clock source can be elected, in some designs automatically so that the whole network does not crash down if the elected grandmaster fails. However, most often the grandmaster clock is elected to be the switch's clock since the switch is a) typically central to the most endpoints and b) the switch tends to be more reliable than endpoints. All of this is why I just mentioned why there is merit in using a good quality clock for some Ethernet switches from a demonstrable academic perspective, but this is wholly independent of audiophool switch tweakery.

Lastly, one needs to be sure that streams are reserved and traffic is shaped. Madonna on stage will not be pleased if her music cuts out because the bored DJ streamed five 4k youtube videos whilst connected to the same network. As such, media streams get bandwidth reservation and traffic is 'shaped' in a way to eliminate bursts that can cause buffer overruns. What makes it so tough is that not only must the precision be very high across all devices, the latency must also be very low. Low latency means short buffers, means very tight control requirements to avoid empty/overflowing media buffers.

This is a very deep rabbit hole, and my software colleagues will shoot me if they read this; some serve on the IEEE802.1 committee this very week discussing these very topics.

Coming back to dreary audiophiles though, little of this matters or is needed fortunately. The closest some may have experienced is if a system with very large audio buffers (and thus added delay) is used together with a TV and causes lip-sync problems, but this is often solved by just adding a rough static delay to the video stream.

Another note should be that there are FAR simpler, and better, methods than all of this malarkey. However, most of them will involve either using something else than Ethernet. Firewire is good for example but since it's all but abandoned, useless. Or better yet, just use two cables. One for regular Ethernet and another for audio. But no, modern society wants a single cable for everything (or even no cable at all), and that's where things get messy.

Nonetheless, some of this is used in the consumer world though: two examples are Ascendo with their cinema speakers and the B&O Beolab 90 where it's used to synchronize the master speaker containing the analog & digital inputs to its slave. I was involved with both. There likely are several others but details are often not published.

Getting back to 'regular' stereo and why seemingly nonsensical things matter for streaming UDP with e.g. Tidal: audiophile switches do not make the ones and zeroes better. It also doesn't bring any improvement that can be seen on Wireshark. It does not improve layer2 or above, and layer1 is both yes and no. The first hurdle, that I found hard to cross, was that measurements do not always show everything. I am fully aware of the ASR tests of the EtherRegen. I also know of Paul McGowan's opinions, even though I see him as a lovechild of Peter Walker and a televangelist. In an Audio Precision, doing a loopback test between a Mogami and some $$$ interconnect shows zero difference. Same goes if you swap out the Audio Precision's own fuse with a gold Hi-Fi Tuning Fuse. Same goes if you power it with a decent shielded Tachii power cable or a $$$ one. I've done all of those tests.

However, there is a very large part of the audio community, and even non-audio spouses, that can easily tell these apart in listening tests, as Francois often attests to.

As engineers, this is a difficult one. If I design a switch for my industrial or automotive customers, there is an Acceptance Test Procedure. If it passes a series of environmental, electrical and performance tests, it passes and that is the end of the discussion. If it fails afterwards it means the test was wrong or incomplete, and then the test procedure is amended. The technicalities are absurdly complex but the concept is simple.

For audio there is no such binary test. I came to ask myself, if things like cables and fuses are claimed to make a difference by such a large group of folks who have no money to gain (only credibility to lose) from it, and if the same group claims that switches make a difference, there may be merit. The only difference to me is that I have zero interest in cable manufacturing, but very much in switches. In absence of useful measurements of the actual DAC output (since ASR already proved that there is none to be seen), I go on theories followed by sonic evaluation. The theory here is that it has nothing to do with Ethernet or what protocol is used with which amount of buffering or CRC checking. It actually has a lot in common with power cables, where everyone will point to the aluminium and brass cabling from the substation; same for what comes into our routers. It's the Last Mile that matters, and the closer to the final load it comes the more it helps.

Ethernet is galvanically isolated, this is true. But no transformer is perfectly isolated, there's a fair amount of inter-winding capacitance that happily passes through RF noise. This noise goes downstream and upsets whatever it's connected to. Even if the endpoint is an Ethernet to S/PDIF bridge, it has to be kept in mind that the S/PDIF encoder is very much analog and sensitive to noise. On a streaming DAC like the Nyquist the challenge is even bigger. The immediate question is of course, use fiber. IMO it's a great idea, except for it to optimally work, the fiber part must be on the endpoint/DAC and not between router and switch or anywhere downstream. Strangely enough I have not seen this on endpoints yet. Also though, fiber transceivers use a laser that creates extreme EMI with the sharp current draws. So, working on that theory one wants to silence whatever can be silenced. I carefully elected a switch IC with the most voltage rails exposed so they can be individually regulated. Even its PLL has a separate regulator. Next you deal with shielding of conducted and radiated noise.
Lastly, clocking. This is for me one of the more difficult ones to get to deal with. Some audiophile switch manufacturers merrily print eye patterns on their marketing, except they conveniently omit to mention that for Ethernet it does not matter as long as it passes the standard tests. My theory here is that a master clock frequency that is jumping around (even if within the typical 25ppm spec of most Ethernet equipment), the jitter causes a harder workload for the PHY clock recovery circuits than a more stable one that has to keep tracking a wandering clock. Harder workload means more asymmetrical power consumption etc. that once again trickles down the chain. Most of the things I optimize I can measure. What I cannot measure is the actual result on audio and I wish there was, but for now measurements are done with a biological instrument, Agaton in this case.

There are several other things that are puzzling, e.g. why daisy-chaining switches help if only two ports on each are used. For several of these I have theories, and several I look forward to exploring further. Perhaps All I need is time, R&D capital and my bio-instruments.

Offline Agaton Sax

Re: Switch! in line
« Reply #39 on: July 16, 2021, 02:34:43 PM »
What is the purpose of a second (Watts) switch ahead of the first (Croak) switch?   Cannot one switch do the whole job?   The sole purpose of a switch is to expand the number of available LAN ports if the router has insuffient

My Mikrotik router has 5 ports, 1 WAN and 4 LAN

1 - Fibre ONT (WAN port)
2 - To the unmanaged switch
3 - To Apple Airport Extreme somewhere else in the house
4 - To a DVR
5 - To my neighbour's house as she shares my internet

I have absolutely no idea what or why. To me, it sounds better -dramatically so. I have no idea how this stuff works but I suspect it all has to do with lower noise and better timing. I think the Switch act as a sort of filter to keep crud from the router out of the DAC. As for the timing thing, I have sort of gotten similar results by using a re-clocker between my Apple Airport Express (before Roon)  and between my CD transport and the DAC. My CD transport is a Studer. Studer was early pioneers in the pro-digital world so I assume they knew something about clocks, yet the re-clocker improves sonics-much like the super Switches. Why the Watts Switch so dramatically improves the scale of the musical reproduction is even more of a mystery to me; higher, wider, deeper, more dynamic, but I do not know why.

Offline johan.pretorius

Re: Switch! in line
« Reply #40 on: July 16, 2021, 02:37:05 PM »
There are several other things that are puzzling, e.g. why daisy-chaining switches help if only two ports on each are used. For several of these I have theories, and several I look forward to exploring further. Perhaps All I need is time, R&D capital and my bio-instruments.

It's clearly very complicated. Thank you for taking the time to share your knowledge.
There is no better idle time than Thorens time.

Offline chrisc

  • Trade Count: (+180)
  • AVForums Grandmaster
  • *****
  • Posts: 13,747
  • Total likes: 1101
  • Condition your mains to condition your ears
Re: Switch! in line
« Reply #41 on: July 16, 2021, 03:55:21 PM »
Maybe a summary might be:

How can I achieve the same result for myself, without spending all the housekeeping money?

Some of the similarly purposed commercial products sold overseas are mouth-wateringly expensive

Here are a few...

Afterdark - Project ClayX Constellation Optical LAN isolated network switch - US$1580.00 (Made in China)

   


Innuos PhoenixNET - 2749,00 - very elegant I must say

Music is the shorthand of emotion

Offline Agaton Sax

Re: Switch! in line
« Reply #42 on: July 17, 2021, 07:58:36 AM »
  My cheapie switch (R110 from SCOOP) had a rudimentary PSU, so I made a better one, fully regulated and stablilised.  This made 0% difference

A while ago Croak posted 3 videos of a  low to midline Clearaudio turntable with a not too fancy cartridge playing music. There were 3 videos. One with a basic SMPS, one with a better SMPS and one with a linear supply. The differences were quite obvious. I could easily pick them out (it was a blinded test-he only revealed what he had done later) on an extremely basic system using a Scarlet 212 interface and Yamaha HS 5 monitors. I  absolutely do not have golden ears, just an unprejudiced approach. I cannot find the post but here are the videos:


https://youtu.be/LIu1ygcxQ7o

https://youtu.be/LSXte0uM4s0

https://youtu.be/i-ME13u4xJE

Why does a turntable PSU affect the sound it reproduces? My thoughts are that it is not so much what the PSU does to the turntable motor but more about what noise that PSU injects into the mains line. I would guess this to be the same with the Switch. I now have enough time with the Switches to believe that, with the Audio grade ones, it is both the quality of the supplied DC and what is injected into the mains that affect the sound. My LPS is still with my Tech. So I used an iFi power supply( super duper noise-cancelling SMPS) last night with the Croak Switch only. The Watts Switch needs 12 V so was not in the circuit. The sound was certainly different to what I know. Not necessarily worse, just different, more like I would expect good quality digital to be rather than the more analogue (dare I say "natural"?) sound of the linear supply.

I grant you that your LPS did not improve the sound quality of your Switch but its effect on the line? If I am correct you use a fairly simple set-up with a computer, said Switch and Devialet? By simple I mean not lots of other sources etc. The Devialet is a very high tech digital amp/DAC/streamer/tea maker. I would assume they have taken heroic measures to make its audio circuits immune to digitally induced noise.

I would therefore conclude that in your set -up the PSU on a standard Switch does not make much difference to sound quality. However for those with conventional amplifiers and other analogue sources it does.




« Last Edit: July 17, 2021, 08:07:37 AM by Agaton Sax »

Offline naboo

Re: Switch! in line
« Reply #43 on: July 17, 2021, 02:31:29 PM »
Was it this thread, @Agaton Sax ?

https://r.tapatalk.com/shareLink/topic?url=https%3A%2F%2Fwww%2Eavforums%2Eco%2Eza%2Findex%2Ephp%3Ftopic%3D93489%2E0&share_tid=93489&share_fid=19931&share_type=t&link_source=app

True, even I could distinguish differences there...

Can't you record your DAC with the different switches, somehow?

Offline chrisc

  • Trade Count: (+180)
  • AVForums Grandmaster
  • *****
  • Posts: 13,747
  • Total likes: 1101
  • Condition your mains to condition your ears
Re: Switch! in line
« Reply #44 on: July 17, 2021, 03:49:44 PM »
The Limetree bridge also came with what they describe as a "medical-grade" PSU.   For fun I attached the Chinese cheapie that came with an old router I am now using as a switch somewhere else.   There was a noticeable loss of definition.   It didn't need many minutes to realise this

So I agree 100%, perhaps more, with your conclusions
Music is the shorthand of emotion