advertisement


Ethernet switches connection to router etc

Deleted my post as I don't wish to offend anyone but I suspect 'audiophile' switch threads are the new cable threads.

If you enjoy spending £££s on posh ethernet cables you may enjoy spending £££s on posh ethernet switches.
 
miktec said:
And the ASR review (video linked to above) proves fairly conclusively otherwise.

"Conclusions
Computer, networking and streaming architecture instructs us that an external switch cannot have any effects on an Audio DAC. But it is always good to put some hard data behind this. And that is what the measurements show. That no matter how deep we dig down in the waveforms coming out of the DAC, no difference exists between a cheap, generic switch and UpTone EtherRegen. Even when going down to incredible -160 dB which is equivalent to a 27 bit audio word (3 bit more than any 24 bit content), there is still no difference.

Measurements conclusively demonstrate that EtherRegen did not change jitter, noise or distortion of the DAC. It further had no impact on its clock speed, or output voltage."

https://www.audiosciencereview.com/.../uptone-audio-etherregen-switch-review.10232/

As always - YMMV
Click to expand...

Miktec
ASR states that I own the 'best' Amplifier (Benchmark AHB2) - and whilst it is great to my ears - an Audio Note Ongatu sounds 'even better' (but is out of my price range)
I presume you do see differences ('improvements') in the photos?
I do accept that these are static photos and it may be possible to critise the methodology but it does appear to be an honest attempt to determine what matters (in this case what is seen). If you do see differences then I woudl suggset you try an audition of the Etherregen . If purchsed in the UK you can return within 30 days (in my case the unit stayed). Alternatively try a £50 Cisco on ebay and sell on if no improvemneyt is noted
 
andrews said:
it may be possible to critise the methodology
Click to expand...
To be criticised, there would have to be a methodology in the first place.

Regarding Ethernet in general, the coupling transformers that must be present are very good at blocking noise up to several MHz. There is nothing a switch can do to improve matters.

Netgear switches are good value and generally a safe choice.
 
Suggesting a network switch can improve sound is equivalent to suggesting freezing a CD will make it sound better...or coating it with some magic resin...
The only possible impact the switch could have (not on the signal, which is 1/0s, but on the 'system') is a) introducing mains noise into the system, or b) performing so badly that parts of the signal are lost - signal dropouts - which would mean that the network simply wouldn't function anyway.
Complete and utter foo....
IMHO.
 
andrews said:
Miktec
ASR states that I own the 'best' Amplifier (Benchmark AHB2) - and whilst it is great to my ears - an Audio Note Ongatu sounds 'even better' (but is out of my price range)
I presume you do see differences ('improvements') in the photos?
I do accept that these are static photos and it may be possible to critise the methodology but it does appear to be an honest attempt to determine what matters (in this case what is seen). If you do see differences then I woudl suggset you try an audition of the Etherregen . If purchsed in the UK you can return within 30 days (in my case the unit stayed). Alternatively try a £50 Cisco on ebay and sell on if no improvemneyt is noted
Click to expand...

I read the review and and studied the frame images and remain completely unconvinced - even more so as I have plenty of experience in both areas (photography and video)
Particularly agreed with comments further down criticising the obvious flaws in the methodology - just before the review sponsor locked the thread ... noting that the 'sponsor' was none other than the distributor of the device being reviewed.

Hmmm....
 
Not to put too fine a point on it ...
If you take a properly focussed picture of any TV screen, including ultra HD, the pixel grid will be sharp and noticeably well defined regardless of the images they are reproducing.

The 'pixel level' images used in the review very clearly aren't in focus ... even more so in some of the images which have been used to illustrate 'bad' tech that are clearly softer than the paired 'good' examples. No amount of 'noise' in the signal would do this to the TV's inherent pixel structure.
https://audiophilestyle.com/uploads...-RAL.jpg.4d11228ae60689fe845a874f18e0ea9a.jpg

Sorry ... but that is just not cricket
 
Last edited:
From the EtherRegen white paper:

"We are going to explain how jitter of the clock—e.g., in a network switch—will show up in the data coming out of that switch"
https://cdn.shopify.com/s/files/1/0...enson_EtherREGEN_white_paper.pdf?v=1583429386

I'm sorry but that's simply not how TCP/IP works. A switch can't modify the payload of a TCP/IP packet - it doesn't know how. If the payload is corrupted in some way it will fail the checksum and be rejected by the destination.

TCP/IP was designed from the ground up to cope with poor connections (i.e. nuclear war!) and is incredibly robust. The data is chopped up, wrapped in a numbered packet, sent down the line and re-assembled in order. If packets are lost or corrupted they are resent - hopefully within the time it takes to buffer the audio.

The data is either bit-perfect or missing entirely. There's no halfway.

I'd have more sympathy if these switches were using some form of Quality Of Service to prioritise audio streams over general network traffic but as far as I can see they're not even doing that.

All that said, if people hear a difference and enjoy what they hear then what do I know :) It's a hobby not a science after all.
 
I have such, works fine just make sure it is a gigabit switch rather than the old 100/10 and use suitable cables. An old 100/10 connection only gives me about 9 megabytes/s max whereas the gigabyte connection gives me between 30 and 80 ie considerably faster. In my case I have a notebook as a server ( good for its low power useage for 24/7 , I connect to it using remote desktop) that I use to down load various multimeda then pump the results across to a number of media computers on the end of a about 10m of cable connected by such a switch. Mine is a 3COM switch been in use for donkey years
 
paulfromcamden said:
Had anyone like Stereophile measured any of these switches?

It's not hard to produce clear data on how a switch performs.
Click to expand...

not sure on that so much depends on the files being tranfered. Small files have a large over head large files in 100s' Megabytes transfer much faster. There is also what else the computers are doing
 
Doesn’t matter as the MTU of the ethernet frame is the same each time, so large file or small file the payload is still (default) 1500... you can reduce processing overhead by using jumbo frames, frames greater than 1500, but this has to be enabled end to end.
 
mercalia said:
not sure on that so much depends on the files being tranfered. Small files have a large over head large files in 100s' Megabytes transfer much faster. There is also what else the computers are doing
Click to expand...

There are plenty of solutions to stress test ethernet switches and produce quantitive performance measurements.

Here's an example of how a large/complex Cisco backbone switch was tested using Spirent hardware:
https://www.networkworld.com/article/3212805/cisco-nexus-9516-review-how-we-did-it.html
 
NRG said:
Doesn’t matter as the MTU of the ethernet frame is the same each time, so large file or small file the payload is still (default) 1500... you can reduce processing overhead by using jumbo frames, frames greater than 1500, but this has to be enabled end to end.
Click to expand...

do these effect lots of small files rather than large ones? Its the small files that are the problem in my case.
 
paulfromcamden said:
I'd have more sympathy if these switches were using some form of Quality Of Service to prioritise audio streams over general network traffic but as far as I can see they're not even doing that.

All that said, if people hear a difference and enjoy what they hear then what do I know :) It's a hobby not a science after all.
Click to expand...
I am not defending said switch makers here but apparently Melco use Buffalo switch mainboards which do support 802.1p QoS. If audio data streams were traffic shaped/prioritised I’m curious to know what you think the reason/reasons would be that could provide an uplift in sound quality? I’ve used QoS a lot for VOIP and Bandwidth management for various types of data but not specifically at home for Qobuz type services for example.

I also agree people are entitled to buy and enjoy whatever they like, it’s a hobby, I feel that these threads can turn somewhat ugly and a bit of give and take along with less aggressive language would be welcome. I struggle to understand how some of the IT related gear aimed at Hi Fi these days could work but am open minded so try things myself and make my own mind up.
 
Amber Audio said:
I’m curious to know what you think the reason/reasons would be that could provide an uplift in sound quality?
Click to expand...

Hope I didn't suggest it would improve sound quality. Reduce packet loss and jitter perhaps but as you rightly imply on a home network hard to see that being an actual problem.

As I stated, the contents of the payload can't be changed. It's bit-perfect or toast.

Edit: Anyways, I'll leave this thread alone. Not my intention to offend anyone and I suspect everyone has already made up their mind.
 
I think a lot of the problem with this whole area is a lack of understanding about how digital encoding/decoding actually works and what the impact of errors in the encoded signal might actually have. Instead there is a tendency to simply transpose understanding of what affects the familiar area of analogue signal paths into the 'unknown' realm of digital reproduction.

This is much more than simply a difference twixt chalk and cheese .... much much more. So fundamentally different in practice in fact that knowledge of the analogue might actually actively inhibit understanding of the real problems that might be encountered in the digital.

For instance, the notion that any random interruption to the stream of data might consistently affect only certain aspects of the reconstructed audio image (such as tonality) and not others I find more than just fanciful.
 
paulfromcamden said:
Hope I didn't suggest it would improve sound quality. Reduce packet loss and jitter perhaps but as you rightly imply on a home network hard to see that being an actual problem.

As I stated, the contents of the payload can't be changed. It's bit-perfect or toast.

Edit: Anyways, I'll leave this thread alone. Not my intention to offend anyone and I suspect everyone has already made up their mind.
Click to expand...
I didn’t take it that you implied it would improve sound quality but when you mentioned QoS which does manipulate data I was just wondering if you had a use case where it might influence it, pure curiosity. I struggle to understand how a switch could improve a streamer when the data goes into a memory buffer, not having tried a Melco or Innuos I am willing to borrow them but it’s not high on my priority list of demos after COVID, speakers are :)
 
You must log in or register to reply here.


advertisement


Back
Top