Jump to content

Fibre-optic ethernet connection


was_a
 Share

Recommended Posts

7 hours ago, was_a said:

To any of you who've tried this, what was your configuration? 

@was_a,

 

A thread was started on this exact topic almost a year ago by @Stereophilus.  Regrettably the thread was closed.  Maybe you should look at the thread.  I did not think that the thread should have been closed.  Others think the same.

https://www.stereo.net.au/forums/topic/252541-ethernet-network-streaming-tweak/?tab=comments#comment-3966338

 

John

Link to comment
Share on other sites
  • 3 weeks later...

Hi folks,

 

I'm also disappointed the other thread was closed.  This topic is very 'now'.

 

I have an Antipodes Edge as server/Roon Core and Devialet Pro as renderer.

 

I had these connected via a cheap $30 switch with $5 ethernet cables.  I was quite surprised at the improvement gained from separating Roon Core and Renderer functions .. like the Antipodes EX-CX model.  This improvement was immediately obvious.

 

On the weekend I replaced the cheap switch with a spare ADSL iinet router/switch (with WiFi); it just functions as a swithch and access point, not router (the router is elsewhere).  It sits about 30cm from the Devialet and has a 50cm ethernet cable to the Edge.  I configured the Devialet to receive signal via wifi rather than ethernet, and disconnected the ethernet from the Devialet.  So there is not opportuity for electrical noise to get into the Devialet via the ethernet cable.  USB is also disconnected.

 

I did not do direct AB comparison, but replayed several tracks I had played earlier in the day and considered there to be an improvement, it took a few tracks of quazi analysis to notice, some things jumped out as better in some tracks, and some things were more subtle.  I suspect the improvements will become more apparent with more listening.

 

I'm curious if the media converter (MC) > fibre > MC arrangement will be better than this wifi link ... what's your experience? 

 

I'm now considering how to improve upon the link from switch to Edge.  Isolation for the switch's noise could be acheived with fiber (or wifi).  As Edge doesn't have wifi built in, an ethernet cable is required (I'll upgrade that).  That can connect to a MC (or possibly wifi access point (AP)) which could inject its noise into the Edge, but that's unavoidable (there's filters that coud be used if that is a problem, eg. SoTM ISOCat, SoTM CAT7 cable, Acoustic Revive).    What would you suggest based on your experience?

 

With the MC>fibre>MC arrangement, does having a better power supply to the 1st MC improve the outcome (given its isolated from the Edge via fibre)?

 

And the final query, is the Aqvox switch better SQ than all this? 

 

Sorry if this starts an unpleasant debate that goes OT.  I wont start a debate about if/when switches that reclock are better than those that dont. Fortunately the Devialet reclocks incoming signals and synchornises with the DAC clock.

 

For interest, I've noted a few audiophile switches, SoTM (base model plus various upgrades), the Linear Solution (a few models), Waversa, Paul Pang and of course the 2 uber costly M12 solutions via JCat.

 

Cheers

Edited by dbastin
Link to comment
Share on other sites

FIbre will not give you any gains, its only normally used for very high speed links or very long runs ( as in over 100 meters or kilometers long, or both high speed and long runs)

 

8 or 10gig and up is where we use fibre into SFP's in a switch or using DAC cables depending on the application (LAN/WAN/MAN or fabric networking for storage.etc...) Its also expensive.

 

better off with a decent no nonsense switch like an HP Aruba and some decent quality cat 6 cable.

  • Like 2
Link to comment
Share on other sites
14 minutes ago, sjay said:

FIbre will not give you any gains, its only normally used for very high speed links or very long runs ( as in over 100 meters or kilometers long, or both high speed and long runs)

 

8 or 10gig and up is where we use fibre into SFP's in a switch or using DAC cables depending on the application (LAN/WAN/MAN or fabric networking for storage.etc...) Its also expensive.

 

better off with a decent no nonsense switch like an HP Aruba and some decent quality cat 6 cable.

So, you tried it and it was not better?

  • Like 1
Link to comment
Share on other sites
2 hours ago, sjay said:

FIbre will not give you any gains, its only normally used for very high speed links or very long runs ( as in over 100 meters or kilometers long, or both high speed and long runs)

 

8 or 10gig and up is where we use fibre into SFP's in a switch or using DAC cables depending on the application (LAN/WAN/MAN or fabric networking for storage.etc...) Its also expensive.

 

better off with a decent no nonsense switch like an HP Aruba and some decent quality cat 6 cable.

The idea for using it here, is that instead of the opportunity for electrical interference to be carried over the ethernet cable from the switch (to the computer) ...... instead you are dealing with the fibre to ethernet converter at the computer end..... which could (potentially) be made less electrically noisy (than the switch).

  • Like 3
Link to comment
Share on other sites

I have gone from wifi to a wired connection.
I found that an ethernet filter made a positive difference.  I'm using a DIY unit, similar to the SOTM item.
Hmm, this is puzzling.

I assumed signal being transmitted thru air (or light in the case of fibre) has minimal potential to be affected by electrical noise and coloration from cable components. But I'm willing to try anything that has shown to be beneficial despite theory.

In my case, it seems my ethernet cable was poor compared to the quite good wifi arrangement tat replaced it.

What is the sgnal path, before and after?

Maybe the wifi transmission was compromised.
How far did the wifi signal need to travel? Was the signal strong? Are there obstacles?

I've read that if the receiver has to work hard to 'correct' transmissions, that could be detrimental.
Link to comment
Share on other sites
The idea for using it here, is that instead of the opportunity for electrical interference to be carried over the ethernet cable from the switch (to the computer) ...... instead you are dealing with the fibre to ethernet converter at the computer end..... which could (potentially) be made less electrically noisy (than the switch).
Well, sjay has indicated fibre has ample technical capability for this purpose. Yes, some is very costly but there are lower cost options, which in the realms of expenditure on hifi gear could be cost effective.
Link to comment
Share on other sites
29 minutes ago, dbastin said:
12 hours ago, davewantsmoore said:
The idea for using it here, is that instead of the opportunity for electrical interference to be carried over the ethernet cable from the switch (to the computer) ...... instead you are dealing with the fibre to ethernet converter at the computer end..... which could (potentially) be made less electrically noisy (than the switch).

Well, sjay has indicated fibre has ample technical capability for this purpose.

No he didn't.

 

He correctly described it's typical use as long and/or high speed interconnection.   However.... It's use in audio is not for that reason at all.

  • Like 1
Link to comment
Share on other sites

7 hours ago, dbastin said:

Hmm, this is puzzling.

I assumed signal being transmitted thru air (or light in the case of fibre) has minimal potential to be affected by electrical noise and coloration from cable components. But I'm willing to try anything that has shown to be beneficial despite theory.

In my case, it seems my ethernet cable was poor compared to the quite good wifi arrangement tat replaced it.

What is the sgnal path, before and after?

Maybe the wifi transmission was compromised.
How far did the wifi signal need to travel? Was the signal strong? Are there obstacles?

I've read that if the receiver has to work hard to 'correct' transmissions, that could be detrimental.

Hi dbastin,

 

In using a Raspberry Pi with a DAC attached to it, the difference with and without wifi is audible.   This is likely not going to the be the case if using a digital output.  There is no attenuation of the wifi signal that would be a concern.

 

At a guess, it could be the additional overhead in the OS for the wifi, power consumed by the transmitter, additional noise on the DAC circuits etc (its a radio after all).

 

I'm using shielded cat6 cables with an ethernet filter.

[Netgear ADSL wifi router] --- [Ethernet filter] --- [Raspberry Pi & DAC]

Link to comment
Share on other sites
21 hours ago, davewantsmoore said:

The idea for using it here, is that instead of the opportunity for electrical interference to be carried over the ethernet cable from the switch (to the computer) ...... instead you are dealing with the fibre to ethernet converter at the computer end..... which could (potentially) be made less electrically noisy (than the switch).

You still would convert it into an electrical signal when the fibre terminates into the SFP into the switch. Sorry but I am not buying it.

 

I have a Cisco switch here at home that takes SFP's , I also have the fibre and spare SFP at work should i be so inclined but i do not want to waste my time as it makes no sense to even bother testing these claims out. Its a nice idea and all but I see zero benefit from it. Even if I run fibre  to upstairs, how am I going to terminate it into a my mac, I would need another switch ( more noise?) and more cable....

 

OK sure you can install a very expensive HBA into your desktop or server computer to take direct fibre, assuming is does not need a just as experience SFP interted into it as well. Even doing it this way, its still a lot of moving parts and its these extra links that degrade signals generally speaking. 

 

Less is almost always more.

 

meh

Edited by sjay
  • Like 2
Link to comment
Share on other sites

 

1 hour ago, sjay said:

Its a nice idea and all but I see zero benefit from it. Even if I run fibre  to upstairs, how am I going to terminate it into a my mac, I would need another switch ( more noise?) and more cable....

The idea is that you would use a fibre to ethernet converter on both ends.    This means you audio equipment is exposed to the electrical noise of only the fibre to ethernet converter .... and not to the noise of the ethernet switch.

 

 

Whether it would make any difference at all... depends on how noise the switch is (that you are avoiding) .... how noisy the fibre to ethernet converter is .... and how susceptible your audio equipment is.

 

1 hour ago, sjay said:

OK sure you can install a very expensive HBA into your desktop or server computer

You could ..... but that's not what they're suggesting.    Often the devices they're suggesting doing this for are not (desktop/server) computers.

 

1 hour ago, sjay said:

its these extra links that degrade signals generally speaking. 

They'll be very unlikely to degrade the network data by any relevant amount.

 

1 hour ago, sjay said:

Less is almost always more.

This is a very very poor generalisation.    It very much depends on the situation/details.

  • Like 3
Link to comment
Share on other sites

I am not buying it.
PS.  I should probably say, before anyone gets the wrong idea.... that my approach is to buy audio equipment which is not susceptible to noise (carried on the ethernet cable).
 
ie. I'm not a proponent of 'tweaks' like this.
I agree, I'd rather have less than more. However I suspect nothing is immune, just degrees of suceptability to noise. I'd guess even a wifi receiver creates noise.

What peoducts have you found to be least susceptible? And how did you determine that?

Devialet has advised me that on the Core Infinity board of Pro units ...

"The ethernet signals are isolated and filtered. The audio extracted from Ethernet streams is relocked."

and ..

"We use a fully asynchronous internal clock structure for internal our buffers, the computer clock used by Roon is generated by our central clock engine, which also generates the clocks for all the system. Of course, de-jittering occurs."
Link to comment
Share on other sites

@sjay, my experience has been noise may not really noticed, until some is removed.

 

You might find a media convertor that is less noisy than your Cisco switch and improves your SQ.

 

But if you're content with you set up, leave it.

 

I respect your view. Maybe you can share some knowledge with our community if they wish to explore the fibre option.

 

If one were to go down the optical ethernet rabbit hole, do you know if GPON is suitable? Or too complicated?

 

Also, would these lowish cost (I've seen much higher cost) SFPs be suitable?

 

https://ubwh.com.au/Ubiquiti/UFiber-and-Fiber/UF-MM-1G

 

 

Link to comment
Share on other sites
44 minutes ago, dbastin said:

What products have you found to be least susceptible?

Hard to generalise... but some products are designed for an realistic level of performance, which doesn't fluctuate under "adverse" (ie. real world) conditions.

 

It's an extremely controversial position to take in hifi, but I will say that price is the great "bamboozler".

 

44 minutes ago, dbastin said:

And how did you determine that?

Measurement of performance.

 

As your hearing of very small distortion and very quiet sounds is very unreliable and subjective.... then measurement with sensitive electronic equipment is the only way to determine what is really there.

 

44 minutes ago, dbastin said:

"The ethernet signals are isolated and filtered.

For example.... Injecting noise intentionally into the ethernet cable, will show if and how the amplifiers performance changes.    It will even show if the amplifiers performance changes in way which are much smaller than could ever be heard.

 

Link to comment
Share on other sites

Noise and there's quite frankly the Eth interrupt on your CPU also.

 

Low noise / very shielded cables aren't hard to find. You can buy CAT8-spec if you look hard enough... Not expensive either.

 

You'll get a decent upgrade just on limiting network traffic required only to your media player so it's not checking all network traffic all of the time. Easy way to rest if this works - try local playback with and without the network physically connected. Putting it on a separate subnet isn't expensive if you know what you're doing, and only moderately more so if you don't and call a professional networking tech. Don't buy into audiophile crap here.

 

Go hard wired. If you're worried about noise a wireless radio trying WiFi is almost always worse.

 

Swenson wrote a ton on mains induced ethernet noise... In short, no LPS required. 

 

If OS jitter is an audible issue you're welcome to improve clock quality on the last switch to your audio device.

 

Plenty of short run industrial and medical applications use optical ethernet or isolated hardwired connections. You're welcome to try (and Ethernet can pickup a good bit given the quality of most home cabling jobs) but buyer beware - a **** reaction to a lot of traffic over a highly isolated connected can still sound poor.

 

There is nothing in this space not solved in other domains. I'd be very wary of 'audiophile' grade gear... It's likely overpriced and underdone.

Link to comment
Share on other sites
14 hours ago, dbastin said:

Can you advise some specific products, rather than generalise?

There are reasons I don't want to do that.

 

People will use specific examples to try and invalidate the general advice ....  and the thing is that "both sides" could be right.

The product that might be used as an example of why my advice is "invalid" might be a good roduct ....  but it doesn't invalidate the generalisation.

 

The are other similar versions of this.   If I say product X .... then someone will say they had a bad subjective experience with "product x".... and therefore I don't know what I'm talking about, and all my general advice must be flawed.

 

 

At the risk of giving you more generalisations that you don't want .... the sort of products I am talking about, are designed to maintain their performance in a wide range of conditions.   For example, noisy power, signals, or clocks.... and these products are often found in "professional" product categories.

 

There's some audiophile mythology which seems to suggest that when you're in a situation where "every little change is audible".... that this is a good thing.    That you're done something like "stripped away all the veils, and now you can hear the little changes".   I think this is a big trap.    Sometimes that's right.... but other sometimes it's very wrong, and can be an example of equipment with a less than great design.   If I can inject noise down an ethernet cable and hurt the performance... then why didn't the design prevent that ?!?!??

 

 

In short, if I have very robust reclocking right before my DA converter ... and/or a very clear break between the source and converter side for the signal .... and I have well designed (ie. isolated) D and A power supplies in the converter .....  then what happens on the source side matters very little.

 

The trick is that this doesn't require expensive parts, or particularly fancy things.... just good engineering.   It could be found in a device which costs hundreds of dollars, not thousands.   Although if you have good performance, then it's typical to charge as much as you can for it.

Link to comment
Share on other sites
1 hour ago, davewantsmoore said:

There are reasons I don't want to do that.

 

People will use specific examples to try and invalidate the general advice ....  and the thing is that "both sides" could be right.

The product that might be used as an example of why my advice is "invalid" might be a good roduct ....  but it doesn't invalidate the generalisation.

Understand and respect that. But I still dont know what to look for and will use things already mentioned and perpetuate the hi priced bs product. 

 

What about discussing more broadly the characteristics of a well setup streaming rig with a focus on the network side ie modem- characteristic a b c, cable abc  router, abc, server abc which gives me/us a better chance of finding something that is more likely to work well.  The kinds of characteristics that might be found in specifications would be best (or something else?). Is that doable? 

Link to comment
Share on other sites
 Share

  • Recently Browsing   0 members

    No registered users viewing this page.





×
×
  • Create New...
To Top