Ethernet Routers for Audio

[keinesorge]

I'm using a Netgear DG834v4.

An ADSL hand-me-down.

Edited September 15, 2020 by keinesorge
its late.

[Hydrology]

Trying to get my hands on a Waversa WSmartHub.

New Australian distributor so stock shouldn’t be too far away.

 

 

 

 

 

[Assisi]

On 13/09/2020 at 5:19 PM, recur said:

 

All I hear is anecdotal evidence and confirmation bias. Spend your money but I’d wager 100:1 that if the same money was spent on room treatment you’d get a much better outcome.

@recur,

To suggest that those who perceive a benefit from audio quality network components as just mere confirmation bias is to me of itself confirmation bias to justify one’s respective position.  Rather than allude to anecdotal evidence it would be interesting to actually experience the benefits of a network comprising audio quality components.  You may be surprised.

If what comes out of your speakers is low quality then room treatments will not improve the SQ.  It seems to me that all a room treatment will do is maybe help so that the listening experience is not further degraded by room imperfections.

John

[Assisi]

On 13/09/2020 at 5:46 PM, rmpfyf said:

I think everyone here is well-intended though the amount of money spent in some cases is crazy for a problem that can be solved easily.

 

Whilst (IEEE 1588) compliant switches are light on the ground they're growing in number and hopefully the audio community takes heed. AES67 is an excellent standard.

It seems to me that you and a few others are taking a simplistic perspective to what is possibly a little more complex for many of us.

 

Some members of Stereo Net have the knowledge and skills to design and construct well performing audio components.  Most members are probably not so blessed.  Most need to buy what they need or want.  Even less members probably have the knowledge and skills to put together and setup networks with the components as discussed in this and other threads.  I have to buy the components and connect them up and hope that the network actually works.  I am still surprised that mine does and especially the listening outcome. 

 

In simple terms what is the problem that can be easily solved?  How would a person with no network skills go about doing it?

 

John

Edited September 16, 2020 by Assisi

[BugPowderDust]

3 minutes ago, Assisi said:

@recur,

To suggest that those who perceive a benefit from audio quality network components as just mere confirmation bias is to me of itself confirmation bias to justify one’s respective position.  Rather than allude to anecdotal evidence it would be interesting to actually experience the benefits of a network comprising audio quality components.  You may be surprised.

 

If what comes out of your speakers is low quality then room treatments will not improve the SQ.  It seems to me that all a room treatment will do is maybe help so that the listening experience is not further degraded by room imperfections.

 

John

John, I'm just a tad experienced in networking and have decades of experience delivering far more time sensitive and bandwidth heavy applications than audio.

 

I've been involved in design and implementation of super low latency, high frequency trading networks here and in the US and also ran the team that delivered the biggest IP based video distribution network in this country. I also personally know people who design the switches that run our banking and internet industries and the IP backbones for companies like Google.

 

Your assertions that you hear a difference because you spent money on something is understandable and I won't dispute it, but it doesn't make it anything other than anecdotal in nature.  

 

 

 

 

[pwstereo]

1 hour ago, Hydrology said:

Trying to get my hands on a Waversa WSmartHub.

The Ethernet portion of that device is described as a hub, so all traffic repeated on all ports, it's not even a switch. Doesn't the thread title request a router?

[BugPowderDust]

I think the name might be misleading. I've seen it advertised as both a hub and a switch interchangeably, although they are very different things in reality. It's hard to tell when the OEM website is predominantly all in Japanese. @Ittaku can you read the WsmartHub bit and translate for us?

 

But yeah, not a router.

[Ittaku]

1 minute ago, recur said:

I think the name might be misleading. I've seen it advertised as both a hub and a switch interchangeably, although they are very different things in reality. It's hard to tell when the OEM website is predominantly all in Japanese. @Ittaku can you read the WsmartHub bit and translate for us?

Link?

[BugPowderDust]

https://waversasystems.com/wsmarthub

[Ittaku]

2 minutes ago, recur said:

https://waversasystems.com/wsmarthub

It's a network switch. They call it a hub to describe that it handles both ethernet and USB.

[pwstereo]

From the makers website (also repeated on resellers websites):

"The LAN and USB hub circuits that affect sound quality are ultra-high-end specifications that operate with the high-precision power supply of the built-in battery of WSmartHub."

[BugPowderDust]

I managed to squeeze it through Google Translate.

 

They say it is a 100 Mbps "switching hub", which I'll say is probably an unmanaged switch. Networking might be complicated to the uninitiated, but there's no excuses for getting the terminology wrong if you're a manufacturer.

 

I continue to love the furphy that abounds that a 100 Mbps port is better than a gigabit port because "noise". 

 

The big noise removal in the "switch" is apparently by getting rid of those LEDs that most switches have to let you know the thing at the other end of your cable is connected and turned on.

 

Points loss for claiming LAN cables are directional, not knowing the difference between a hub and a switch and for thinking a better clock on the switch will make any difference to the IP traffic running above it. 

[pwstereo]

There's a .kr makers website in English.

It says

 

"Key Features

Remove/isolate digitial noise from USB/Ethernet source equipment

Built-in USB2.0 Hub and Gigabit Ethernet Hub"

[PCOWandre]

7 minutes ago, recur said:

Points loss for claiming LAN cables are directional, not knowing the difference between a hub and a switch and for thinking a better clock on the switch will make any difference to the IP traffic running above it. 

Can we take away another couple of points for a billet CNC job that looks like a "My First Fusion360 Project"?

 

10 minutes ago, recur said:

getting rid of those LEDs

Getting rid of some passives and adding an LCD with controller and backlight! Yippie!

[Ittaku]

39 minutes ago, recur said:

I managed to squeeze it through Google Translate.

 

They say it is a 100 Mbps "switching hub", which I'll say is probably an unmanaged switch. Networking might be complicated to the uninitiated, but there's no excuses for getting the terminology wrong if you're a manufacturer.

I'm a Japanese translator... I explained it above.

[Assisi]

1 hour ago, recur said:

John, I'm just a tad experienced in networking and have decades of experience delivering far more time sensitive and bandwidth heavy applications than audio

I do not doubt for one moment the extent of your networking skills, knowledge and experience.  As I said of myself and others, we are not so blessed. 

I seem to remember some time ago that @rmpfyf  suggested to me to try and change a switch to 100 Mbs instead of 1gb as there maybe a benefit.  With the switch that I had it wasn't possible to change it.

This morning I disconnected my audio switches and replaced them with a just a Netgear 108 with an ordinary smps.  I kept the cables in as it was too complicated to remove those.  Played something that I use for comparison purposes.  The result was good as I had expected it would be.  Put the audio switches back with no other changes at all.  Result was better with more life and decay. 

John

[pwstereo]

5 minutes ago, Ittaku said:

I'm a Japanese translator... I explained it above.

You explained the name of the product, but their own site (and reseller sites) specifically call the USB part a hub, and the Ethernet part a hub. So the whole things is a SmartHub, but it contains separate USB and Ethernet hubs.

[pwstereo]

3 minutes ago, Assisi said:

...more life and decay. 

What does decay mean in this context?

[PCOWandre]

There's always going to be a little friction in these discussions because people who actually do networks (and other IT) for a living deal in hard reality and not wishy thinking.

 

Networks are observed in terms of bandwidth and latency and packet loss and CRC errors and other measurable, real world, concrete details. Performance problems are measured in terms of "the database is I/O bound because the interfaces to the NFS head are running hot" or "packet loss is killing the connection between the firewall and the load balancers, I better go fix that". Or a hundred and one other problems.

 

Network switches, for the most part, either work or they're faulty. There's no shining improvement by putting fancy clocks on them, because if they were clocking out frames at the wrong rate the next device would just discard the frames and hopefully count the CRC errors or even isolate the jabbering device. 

 

We know that despite what Audioquest might claim in between describing hard drives as a "record player" with "surface noise", ethernet frames are not like I2S with a bitstream that can suffer from jitter and latency. The switches have memory and store 'n' forward. Routers have queues. NICs have buffers which are copied to memory which are unpacked and copied to memory and so on and so on. Geez, now everything is increasingly https, you've got so many more memory copies for decryption.

 

So any time the idea of a thousand dollar ethernet cable or an audiophile switch with clever clocks and magical properties comes up, the response is pretty much like an oncologist hearing that a patient is going to treat their cancer with homeopathy -- there's an instinct to offer advice, to suggest that's not the right course of action; to scream quietly. And when we read that people "hear more air" or "more PRaT" or "warmer presentation" from switching ethernet gear around, we're going to say confirmation bias because unless we're going to accept paranormal outcomes, it just doesn't work that way. And there's a bit of a drive to push back against the companies selling this stuff for the same reason that oncologists push back against quack cures for cancer -- we don't like seeing people taken advantage of. 

[Ittaku]

58 minutes ago, pwstereo said:

You explained the name of the product, but their own site (and reseller sites) specifically call the USB part a hub, and the Ethernet part a hub. So the whole things is a SmartHub, but it contains separate USB and Ethernet hubs.

Nations that use English as their second language infamously misuse English terminology either completely or - far more commonly - miss the nuance. If the title and description ends up using these loan words, then unless they have a completely renamed product for the English speaking world, it will propagate. Getting hung up about it is pointless. The question was whether it was a network switch or not and it is, and to them it's a "networking hub".

[Guest rmpfyf]

2 hours ago, recur said:

I continue to love the furphy that abounds that a 100 Mbps port is better than a gigabit port because "noise". 

 

No, it's a timing source used to generate a periodic output offers inherently better timing accuracy at lower generated frequencies - simple maths. Not a noise thing. This sounds crazy to a network person as the effect in packet networks is indirect, the usual 25MHz crystal exists to run the board and the effect on packet periodicity is limited. Packet network exist so that it can be a 2c crystal not a precision OCXO and transmission still works. If you want to force timing, use PTP.

 

2 hours ago, Assisi said:

It seems to me that you and a few others are taking a simplistic perspective to what is possibly a little more complex for many of us.

 

Some members of Stereo Net have the knowledge and skills to design and construct well performing audio components.  Most members are probably not so blessed.  Most need to buy what they need or want.  Even less members probably have the knowledge and skills to put together and setup networks with the components as discussed in this and other threads.  I have to buy the components and connect them up and hope that the network actually works.  I am still surprised that mine does and especially the listening outcome. 

 

In simple terms what is the problem that can be easily solved?  How would a person with no network skills go about doing it?

 

It's simple, and I've mentioned it before. 

• You can reduce conducted noise >> simplistic perspective is a transmission methods developed to be immune (fibre, used in high-noise environments)
• You can reduce the amount of traffic >> simplistic perspective is to sub net it, VLAN, etc
• You can make the periodicity of communications more regular >> simplistic approach is to force timing through a standard that allows a master clock source, and is engineered to compute time lag throughout network components (AES67 using PTP and IEEE 1588, which are common in industrial automation - note this is very different to bandwidth arguments)

 

The audiophile industry has directly or indirectly observed these challenges and done differently:

• You can reduce conducted noise >> go long trying to make a copper cable act like fibre, all sorts of (fibre!) isolators, absorption in the box, cascaded switches (more iron isolation) etc
• You can reduce the amount of traffic >> cascaded switches (depending how they're managed)
• You can make the periodicity of communications more regular >> switches with better regulated clock sources though having no direct effect on packet periodicity (there is an indirect effect), expensive cabling with very high conductivity enough to better preserve waveforms, rising edges, etc (which very often is longer than is ideal for the application)

 

Now, whilst the solutions are conceptually simple actual diffusion in the market is hard. I had a good post here on page 1.

[Stereophilus]

1 hour ago, PCOWandre said:

There's always going to be a little friction in these discussions because people who actually do networks (and other IT) for a living deal in hard reality and not wishy thinking.

 

Networks are observed in terms of bandwidth and latency and packet loss and CRC errors and other measurable, real world, concrete details. Performance problems are measured in terms of "the database is I/O bound because the interfaces to the NFS head are running hot" or "packet loss is killing the connection between the firewall and the load balancers, I better go fix that". Or a hundred and one other problems.

 

Network switches, for the most part, either work or they're faulty. There's no shining improvement by putting fancy clocks on them, because if they were clocking out frames at the wrong rate the next device would just discard the frames and hopefully count the CRC errors or even isolate the jabbering device. 

 

We know that despite what Audioquest might claim in between describing hard drives as a "record player" with "surface noise", ethernet frames are not like I2S with a bitstream that can suffer from jitter and latency. The switches have memory and store 'n' forward. Routers have queues. NICs have buffers which are copied to memory which are unpacked and copied to memory and so on and so on. Geez, now everything is increasingly https, you've got so many more memory copies for decryption.

 

So any time the idea of a thousand dollar ethernet cable or an audiophile switch with clever clocks and magical properties comes up, the response is pretty much like an oncologist hearing that a patient is going to treat their cancer with homeopathy -- there's an instinct to offer advice, to suggest that's not the right course of action; to scream quietly. And when we read that people "hear more air" or "more PRaT" or "warmer presentation" from switching ethernet gear around, we're going to say confirmation bias because unless we're going to accept paranormal outcomes, it just doesn't work that way. And there's a bit of a drive to push back against the companies selling this stuff for the same reason that oncologists push back against quack cures for cancer -- we don't like seeing people taken advantage of. 

I completely understand where you are coming from on this.
 

If there was an explanation of why networks can affect the SQ of streamed music, would that change your perspective?  The best explanation of why network switches / routers and isolators affect audio SQ is here: 

 

It is, to my mind a very clear and persuasive argument for what I hear.  However, I am not knowledgeable enough in networking to know if it is factually incorrect.

[Assisi]

1 hour ago, pwstereo said:

What does decay mean in this context

The note hangs or last longer in the air and you move into a silent space for a moment.  It may be simply because there is a reduction in noise floor that may mask the delicate end of the note and the silent space.  All the audio switch may do, is reduce or remove interference or noise in the signal that you would otherwise hear.

 

 

With the virtually nothing that I know about networks and Ethernet, I am sure that the those who say it is all about the integrity of 1s and 0s being sent and received correctly is absolutely correct.  What I sure about though, is that there is something else that happens and is important with audio switches and cables.  The numerous gear that the network people refer to may be mission perfect for what it is designed for. 

 

 

Is it designed to remove interference/resonance that can pollute and spoil the audio signal?  If there is interference/resonance inherent in the referred to gear, is that critical to the performance of their networks?

 

 

Some people do not appreciate how noise floor can spoil what you are listening to.  When I first experienced a reduction in noise floor, initially I did not know at the time what had happened.  The minimisation of interference/resonance can be an expensive pursuit.  I have been pursuing the reduction in noise floor in every way ever since my first realisation to enhance my listening pleasure.

 

John

 

Edited September 16, 2020 by Assisi

[PCOWandre]

2 minutes ago, Stereophilus said:

It is, to my mind a very clear and persuasive argument for what I hear.  However, I am not knowledgeable enough in networking to know if it is factually incorrect.

It's pretty much voodoo wishy thinking, on the face of it. I'll give it proper re-read later, but the facts just aren't stacking up. If ethernet frames were a primary cause of "ground plane noise", the problems wouldn't be restricted to audio, you'd have bright flashes on screens and generally the whole thing would lock up because the timing for accessing memory (RAM) is thousands of times more strict than S/PDIF or USB audio.

 

Also, there's actual test equipment that can measure the quality of ethernet timing, and nobody selling "clever solutions" seems to have actually used it, nor have they ever had an independent third party assess it. If the timing of ethernet frames could have any audible impact, the first call to action would be to eliminate as many frames as possible. You'd be surprised how much broadcast traffic is on the average home network. 

 

That's a well-written bit for justifying the story, in the same way that Shakti have some pleasant to read bits around their "stones". However, despite their promise that they can increase engine performance and fuel economy in these times of increasing emissions standards, no car manufacturer has filled their engine bay with Shakti Stones because while the story sounds plausable prima facie, there's just no substance to it.

[pwstereo]

11 minutes ago, Assisi said:

The note hangs or last longer in the air and you move into a silent space for a moment.

Thank you for explaining.