HOW AND WHY we test network switches...
HTML-код
- Опубликовано: 16 ноя 2022
- We love streaming audio at Alpha Audio. But it's not because it's plug and play... because it isn't. There is a lot involved when you want to get the best possible performance. We have been into network switches for a long time now. In this video we discuss the WHY and the HOW we test network switches. And we explain the 'ins and outs' of the next big test of network switches.
You want to support Alpha Audio? You can! / alphaaudio
More information:
www.alpha-audio.nl
www.alpha-audio.net 
Наука
Thank you for doing this research, I do appreciate effort your putting into it. Looking forward to your findings.
Thanks!
Love it, thanks for dedicating time to this!
Looking forward to this. I recently got the LHY SW-8 switch on the back of great results with the LHY OCK-1 with my Gustard R2R dac. I was hopeful of some improvements but I was a little sceptical, my expectations were modest, I mean packets are packets, error correction etc, yes there may be a little electrical noise, but how much would impact would it make. I am sceptical no longer.
Love it... i'd call myself quite carrier-isp-grade ex-networking osi-guru-guy and was just loughing my ass off some years ago ;-) THIS CHANGED COMPLETLY WHEN I started more seriously my audio journey and also to put in a Paul Pang Dual OXCO (i know, you dont believe it in ;-) and now I play quite a lot with diy psu and cabling... thanks for those efforts!!!
I also laughed .y ass off a couple of years back. Until i started experimenting with fiber. The difference was pretty big. After that I wanted to know why... Well... Still working on it! :-)
@@TheAlphaAudio when will you be publishing the white paper ! Should be a good read.
@@r423sdex this is our latest article...
alpha-audio.net/background/how-a-network-switch-affects-audio-playback-an-extreme-deep-dive/
awsome thank you very much, cant wait
Good idea to have different people do the measuring and the listening parts of the tests. And it's interesting to see here how the measuring is done
Thanks!
I don't think I will ever spend any serious money on these, but still interested in the theoretical background. Thanks for explaining this as simple as possible, and showing what is happening. See you in the next one!
Most switches shown are a between 30 and 100 Euro and an iFi iPower2 is 70 Euro. The test to be conducted will show you what you get for your money in sound differences. Watch this space for the livestream announcement!
@@jmtennapel OK thanks, those prices can be within budget. Will look for your livestream!
Yes I am interested. I heard big differences between power supplies on my Meraki ms220 switch. An oversized Jung Super Regulator sounded the best.
I can imagine.
Looking forward to hear about your findings Jaap and team. I presume that you add the white noise to simulate controllable " traffic" ? FYI have you ever heard of passive intermodulation ? Just trying to help with no intention to be the " smart " guy " 🙂I did in the past many RF tests and became aware that when you inject to much energy on a coil it can go into saturation and that coil will create passive intermodulation. I was thinking about that because likely the special audio tuned switches have isolation coals on the ethernet output? Great work ! and again I'm looking forward to hear about your findings.
Fantastic video Jaap. Thanks. Now can't wait for the juice of the story....Understand you have many measurements to carry out but hopefully you will be educating us in a series of videos as you are processing the test results? Adding another dimension, if you wish to consider this, in your reviews....upgrade-ability in relation to cost.... I do not have the budget, unless starve for a couple of months 🤔...to purchase the quad Paul Pang....But I do have the budget to buy the baseline SOTM, then add the internal clock with input...then add external clock.... Cost in both cases is same. Result, I hope you can tell us which is better. But one is an option for me whereas the other one is not.
Hi Magmamin,
Thanks for the comment. I'm not a fan of the basic SOTM. The basic adapter is pretty bad. The PP Quad is very good.
@@TheAlphaAudio Yeah, fully took into account your review of the 7 switches. But as you know there is an SOTM model with the internal clock and input for external clock. The SOTM plus extra internal board is at Euro 2k...Together with an Afterdark clock or equivalent and a LPS the total would be 3K. So quite close to the PP Quad....The question then is how do the two systems at about same price compare rather than the baseline model which is a third of the price. Would much appreciate if you could arrange to include such a comparative. (I know that you consider that clocks don't make much of a difference, which makes it even more interesting!)
@@administratormagmamin8397 I will try to get samples.
Thank you for the work. It's time people get to know more about digital (specialy) audio streaming done right. Already have a network with cheaper switches without extra power supply. One switch direkt behind the router that one does the work to transport the signal trough the house to my system than it goes to the second switch which communicates with the server from there it goes in to the third switch which only does the one job to take the signal from server to the network streamer/ bridge... Also the critical pathway was chosen to go trough the almost shortest (and no interference between pathway) "port to chipset to port" signal phatway etc. Everything cat7 which I also think it's the best?... That was my theory of an network. Right now I'm in to the bigger projects with much better switches and also I'm considering extra power supplys. I'm curious and thankful you guys are going in deep one the matter.
Welcome!
Sorry for newbie question, but what about wifi router units? Good technology? Bad technology? I do all streaming vie wifi from a wifi unit attached to a broadcom cable modem. Am I just ruining the data/music I stream?
Don't be sorry... no problem. Wifi can definately work. But it completely depends on the implementation of the wifi-chipset in the streamer. Some do a great job, some don't. By the way: the data will be the same: cable or wifi. It's all about the noise you introduce with an extra chipset / antenna that is doing its job.
Hi,
I love your comparisons. Always good value. When are the results due to be broadcast?
This is of particular interest to me as I am anticipating the release of the Brooklyn Bridge II which has a roon core. I’ll be needing a switch to connect it to the Ubiqity Router. I’m running cat 8 patch cables and have a media converter to provide isolation of the internet signal just before the BBII. The media converters powered by Allo shanti dual 5v linear power supply. I’ve heard that a clean switching supply is better for switches so not sure how that applies for the converters.
Cheers and keep up the good work.
The 18th of december we have the live test.
Fiber or AC wireless are the way to go to dump that extra noise.
I'm confused about how this would work - not meaning it can't/doesn't, but how to implement. I have fiber cable service coming into the house. That is converted to ethernet on a provided device that only has a single port. That feeds a 24 port switch which handles all the hardwired stuff in the house, including my streamer. I guess I could terminate from the fiber box into a smaller, better switch and use it to feed my streamer and the other switch. Does that make sense?
No. That is not optimal. The small switch that connects to your streamer should be as close as possible to the streamer.
What Ethernet patch leads are you using in this test, please?
Stock CAT6 standed patch cables in the test.
Why are you loading a switch with a signal (white noise) that:
1: It's never going to see in the wild
2: Isn't designed to deal with
Why not grab a computer, install IPerf and generate real traffic in a controlled manner?
Easy. Besides power supplies, sonic differences are from harmonic spurri, intermodulation distortion, clocking artifacts.
Yes
Most of network switches not have "hard" signal ground base because they using external AC/DC power supply. This floating signal ground can be often rather big problem source and cause increased noise levels and random disturbance signals (switches own clock signals can then easily leak to the another device thru LAN cable) . It can very easily detect by spectrum analyzer. Best method to avoid such problem is to use double shielded CAT8 LAN cable and do proper grounding on network switch and/or the device where to connect to LAN cable. This was just my own experiment.
You are right. Grounding the switch - some have a screw on the case to do that - makes a huge difference.
Check out our new article about the influences of switches on audio clocks:
alpha-audio.net/background/how-a-network-switch-affects-audio-playback-an-extreme-deep-dive/
See Audiosciense measurements, they measured the output of the DAC. We don’t listen to switches but music. The DAC determines what we hear….
We have done measurement on the clock jitter in combination with different switches. There is a difference in jitter. That is audible.
@@TheAlphaAudio When using an USB connection to a good DAC this should not be the case, the clock op the DAC wil take the lead.
You say the data will be fine.
Audio transferred through a switch is also only data. If you can hear or measure a difference, then the data is not OK.
What kind of datatransfer (audio-transfer) are we talking about?
I am not so much into tech details about different protocols, but are we talking about some "casting" protocols that are one-way without error check/correction?
It's about the - mostly - common mode noise getting through the switch in to the streamer. That noise will influence the clock and eventually, possibly d/a-conversion. That is what we are trying to find out.
@@TheAlphaAudio But to find that out you have to measure the output of the DAC....which you not did.
Have you thought about doing a waveform comparison from the output of the DAC? That may tell a better story.
Not really, actually. It doesnt say anything about timing for example. Also: I would have to do some A/D conversion. That is problematic, because no 2 recordings are exactly the same. Even with exactly the same equipment.
@@TheAlphaAudio Only assuming that the signal noise on the input of an streamer or the usage of different switches does have an impact on the output of the DAC is not enough.....
Either proof that by measurement results or a real blind test....
I think people shoud stop saying HIFI grade "Lab Grade"= limit device to device noise
?
I guess what I am saying is a properly designed device would not allow the galvanic isolation to fail.
When that works all that needed is shield and additional EMI/RFI through a RF Choke
Ignore the trolls who don't want to listen and just want to mock, don't feed the trolls!
Thanks....takes some self restraint... ;-)
Ignorance is bliss ! Your ears can measure, but electrical instruments cannot ! Very strange don't you think. Remember it's what comes out of the speakers you need to measure, I could care less about anything else. Set up a microphone, record the music with and without the switch and take one from the other and play that file back. Very simple !
@@TheAlphaAudio what does it matter what happens before the music comes out of your speakers. I could care less about noise, just record some actual music with and without the switch. And then show us the files, if you can hear a difference it will definitely show !
@@r423sdex ok
why not just record the difference so we can hear it then? I'd like to hear it but somehow it's too hard to plug a DAC connected to various switches into a MIC and measure🤷♂
The switch makes a BIG difference in my system for sure. Thanks for the video
My experience with streaming changed the day I got the English electric switch.
Yes... Nice one.
@@TheAlphaAudio Have to tried different power supplies with english electric switch? Can you recommend any?
@@geir8849 there are many. What is your budget? IFi is a nice budget option. LHY is a nice midrange option and Farad is ultimate.
@@TheAlphaAudio 500-800 euro. Farad, is better than Sbooster you think?
@@geir8849 ow... Definately!
not related to audio quality at all but super interesting test anyway.
It is related... Different switches sound different. Power supplies as well.
@@TheAlphaAudio objectively analyzing the audio samples??? Or subjectively?
@@net_news what do you mean? We don't analyze samples. In these tests we have objective measurements and a blind test.
This again.... DACs can't process noise. They can only process signal. So noise below any level that leaves the signal discernible is irrelevant.
Rubbish. Noise affects all the circuits in a DAC.... and indeed ANY audio device.
@@steven2809 Feel free to link to any evidence that noise in the digital signal (at a level that leaves the signal discernible) has an effect on the analogue output signal of a DAC.
@@PitchnYaw Feel free to learn some electronics....
@@steven2809Ha! That's as clear a "no I can't produce any evidence" as I could get. Thanks for playing.
@@PitchnYaw Feel free to provide evidence that noise and EMI do NOT affect electrical circuits....🤔 I've heard this argument before that 'good DACs are unaffected'...frankly it's nonsense. You believe what you want to believe...if it makes you feel happy......🙂
... when the results form these tests?
The 18th is the livestream. After that the written review with all results.
Are you trying to make the switches sing? Are you trying to listen music direcly from switches? This is very misleading topics!
How is this misleading.
@@TheAlphaAudio Ok let me tell you:
- You are injecting white noise (full bandwidth) into switch. This is not actual normal working of network switch. So why bother? Why don’t you just test for normal operation for all the switch and compare?
- With differenct filter capability of the “white noise” of different switch, it does not mean they can make the sound output of the system different! Why don’t you measure the actual analog output of the DAC in the chain with different switch?! If they sound different, you can of course can show it on the analyzer! Why didn’t you do that?
What you showing in this video is only for noobs!