Ethernet Switch- what's the point?


I run an Ethernet cable between my router (standard issue from Verizon) and my streaming transport. I note that some use an ethernet switch between between the router and streamer. Assuming I got that right, what is the point- what does a good switch do? I've been into audio since the 70's but when its comes to streaming, I'm definitely a newbie- 

Thanks all!

128x128zavato

I appreciate your detailed explanation. Some of it is beyond my understanding, I’ll admit that. One thing about digital: it’s been almost 20 years ago now but I began to notice wow and flutter while playing records. CDs had sensitized me. I replaced the belt which helped but I also had to refurbish the suspension. My turntable today has imperceptible wow and flutter. Mostly in the highs is where I used to hear it. But interestingly I find some vinyl has better pace than digital- a few rare cases. Since putting in the new, expensive network switch I have noticed that not so much. It might be that I have not been playing as much vinyl. Streaming sounds so good now and I’m exploring a universe of music. Love discovering 1960s rock and it sounds so good. Nothing like how I heard it back in the day on AM radio.

@tonywinga

 

They’re measuring the analog output. I’m not sure it’s all FFT, may be wavelet or whatever. I think it’s very safe to say that no spurious white noise or random tones or phase or timing errors are sneaking by the measuring equipment, unless they’re 140 dB or more down in level, in which case they fall below the noise floor and will not have any effect on soundstage or tonality for anyone no matter how good their ears or their playback gear.

For imaging, there may potentially be timing issues between the left and right channel. We are more sensitive to that kind of thing, but I think there’s a limit to that. I’m not sure what it is. We tend to notice directionality associated with timing differences between left and right ear starting at about 0.1 millisecond. Jitter involves timing errors vastly smaller than this. I think we can safely say we’re not hearing jitter in any way at all unless our gear is severely malfunctioning.

Might be something else though.

I’ve heard some suggest ringing of digital filters as digital’s big problem. This idea was popular for a while but some audio reviewer got ahold of a DAC that would let him apply and listen to filters he could program himself. He intentionally made some terrible filters in terms of having obnoxious amounts of pre and post ringing. To his surprise and amazement, these didn’t have much of an effect on the sound at all. It’s really not that amazing because the ringing is so far down in level, otherwise the test equipment would easily pick it up as distortion. So I think it’s safe to say it’s not digital filter ringing that we are hearing 99 percent of the time.

Could be something else though.

What could it be? We know that the signal in the range of human hearing is being very faithfully produced at the output of just about any DAC that isn’t broken. A cheap switch normally will provide the digital signal intact, allowing the DAC to function within it’s normal operating parameters. If adding an expensive switch to the signal path is doing something good it has to mean that something was going wrong with the previous switch. What could it be?

LP being able to present us with great sound is proof that we’re really not that bothered by noise and distortion, we may even perceive it as an enhancement. It may help us suspend our disbelief. I did an experiment once with digital photography, taking over 100 pictures of the same scene and then blending them together to eliminate digital noise. It worked. And the result was --- color banding! Without noise it’s impossible for there not to be banding across pixels over gradients. Think about it - how would you show a black to white gradient if you only had a picture 30 pixels wide? You’d see 30 bands of grey going from black to white. What if we created the same scene using a very small piece of photographic film? You’d see film grain and a lot of noise and a sort of blurriness, but no banding. If you add noise to the digital picture you get the same effect. The banding is no longer apparent. You notice the noise but you also perceive a smooth gradient behind it. Your mind can separate the noise from the signal, and the signal is remarkably perceived as being more detailed and natural.

DACs do add noise on the output, and they even shape that noise to make it less audible, so no equivalent of imaging banding is happening on a DACs output. 44.1 16 bit audio is the equivalent of an extremely high resolution photo viewed from an appropriate distance. But, our brains react to easily detectable noise, giving us an impression of the content behind the noise that’s different than when the noise is not noticeable. I find that adding easily visible noise to digital photos can sometimes have remarkable positive emotional effects in how I perceive the photo. I suspect that the perceptual problem with digital audio for some people is it’s lack of noise and distortion, not an excess of something that our ears can pick up on but measuring equipment can’t. Some noise gives our brain wiggle room to imagine, or I should say deduce, what’s not actually being heard. 

Post removed 

just like teachers, professors, friends, colleagues, politicians and so on, that we have all been subject to in our travels...

there is a person’s message or point, then, there is how they choose to deliver it

so many examples of horrendous, unpalatable styles of delivery chosen that totally turn off potential listeners, such that whatever merit the substance of the message may be becomes entirely moot

no one wants to be subjected to a heavy handed, argumentative, defensive blowhard

Back in the day when everything audio was analog, we all understood what was noise.  We had hiss from the electronics, hiss from the tape decks, hiss, hum and crackles from records and phono cartridges and even an AM radio station could sometimes bleed through to the speakers.  Digital noise manifests itself differently.  I see these "tech" guys on the internet showing FFT plots of Signal to Noise of a DAC and judging this cheap DAC better than that expensive DAC because the noise plot looks better.  But what are they really showing?  Is it the FFT plot of the digital signal or the analog signal that they are showing?  If it is a post DAC analog signal then the question is how good is good?  CD players have been claiming 110dB-120dB S/N since the early 1980's.  In the end, these "tech" guys are merely reciting technobabble.  No relationship to the sound.  Pretty graphs impress people.  I watched execs fall for the pretty graphs in the business world for years.

If anyone here on these forums worked along side Al Gore to invent the internet, wrote the IEEE specs for internet gear and/or designs and builds internet gear raise your hand.  You might have some wisdom regarding these topics but even design engineers don't always fully understand their products once they go out into the world and are applied in ways they never considered.  (Consider the Fast and Furious movies.  I know full well that GM never tested a Pontiac Fiero in outer space.)

The bulk of us on these forums are Application Specialists (mostly self trained) of Audio Systems.  We have varying levels of understanding of acoustics and electronics.  Some of us might even be able to do the math up to some level.  So I don't understand why so many here have to try to explain or convince others how digital works or doesn't work or why cables work/don't work, etc.

Back to digital noise.  I don't hear noise with digital in the form of background hiss or hum.  What I hear as I added fiber optics and then a high quality network switch, LPS's and better ethernet cables was 1) Smoother highs, 2) better bass, 3) improved clarity, 4) Better, more focused imaging and 5) bigger soundstage.  I can listen to streaming all day.  It sounds that good.  I have a great vinyl rig.  I play records and love the sound.  I play CDs and love the sound.  I can stream and love the sound.  It is a great time to have a stereo, I think.  Streaming or local 96/24 files seem to sound the best.  It rivals the soundstage of vinyl.  So here is a puzzle- if vinyl has such a low S/N compared to digital, why is the soundstage and detail of records so good?  (That's a rhetorical question.  Please- no technobabble about why vinyl sounds better/worse than digital.)

Post removed 

I am thankful for all the tech people who contribute.

Wholeheartedly agree. Could it be that 222’s heavy-handed a smidgen and that is perhaps why ....

@12many well that is nice, not sure why you called me out as a good many agree he is a one trick pony. I am sure you learned a lot from this thread. 

@recklesskelly "I would hardly call it persistence, more like trolling as this is the only subject matter he contributes to. And reading through his archives he is consistently chirping off on it." 

I have found that people who only speak on one topic usually know that one topic very well.  I have personally meet people who profess to know everything about every topic, and well, they usually don't.  I think that is called jack of all trades, master of none.  I am thankful for all the tech people who contribute.     

So a good network switch can reduce noise and distortion, some of which comes from jitter. Anything else? How much of a difference should one expect? Like going from -90 dB to -95 dB noise to signal? Is it something very subtle but still worthwhile for some, or something quite loud and obvious to some, more so than typical room reflections and diffraction effects? I get the impression it's something far to subtle for me, and yet I can easily hear the difference between live music and any high end 2 speaker system. 

@jmbumgarner01  @zavato 

I was also using my streamer wirelessly too, but wasn't happy with the sound quality vs. CD play, or playing directly from my HD 'server' connected to the streamer directly.

As I had a router/extender being used elsewhere, but not critical, I moved that extender down to my Audio rig location and utilized it to hardwire to my streamer direct. This alone helped, but adding the converters between the router/extender and the streamer, made a another nicer improvement in SQ.

@bkeske Thank you for a voice of reason.

OP this is a great and simple summary of what bkeske uses. I have decided to go this route as I have been streaming wirelessly from my router with no noticeable issues but want to see if I can hear any differences. In a system well north of $50k this seems like a small investment worth trying.

https://www.psaudio.com/blogs/copper/improving-digital-streaming-with-a-gigabit-ethernet-media-converter-system

 

 

@jon2020 sorry for the confusion, you do need a switch. You do not need a $1200.00 or even $800.00 switch as you are converting from the dirty Ethernet side to clean noiseless fiber. Look at the converters on Amazon and you can do finder (short runs) for unsolder $200.00. 
 

@recklesskelly 

"Fiber and be done."

Why does fibre not need a switch? I have fibre and am thinking of getting one until your post.

Thanks in advance. 

Jon

Nate - a switch, apart from being a convenient device to increase the number of ports you have available for attaching things to your router, will often improve the sound. To do this it needs to be a good quality switch. It does this by reclocking the data and removing ‘noise’ (rfi/emi noise) which your router may inject into the digital path through its poor shielding and power supply. 
My experience tells me you need a good quality switch but not necessarily an expensive audiophile one. For example a second hand Cisco 2960 to my ears sounds as good as audiophile ones I have tried.   Cisco make high quality business switches in vast quantities hence prices are low (especially second hand prices which are next-to-nothing) for arguably a better quality piece of hardware than a small production run audiophile switch.

@thyname first off I am not Earl, Jeff or any other name in your head. You also need to get a sense of humor the joke was Freddy is the bear. I will never address you ever again. The EJDS runs deep in you. 

@thyname 

Just ignore the likes of Freddy….it’s best not to engage them. I am glad to hear you’re loving the Muon Pro. 

Jeff / Earl: my apologies. Not sure who controls the Audiogon posting today, but I thought I’d apologize. You and I are in the same “camp” most of the time (other than The Orange related stuff). But you keep attacking me for some unexplained unexpected reason, whatever that is.

Post removed 

@lalitk : except someone here (like Freddy) will accuse you of pushing expensive stuff to innocent people 😂😂.

 

I totally agree. I have tried many switches, although not yours, and by far, the Muon PRO beats them all hands down. And it’s very convenient to use, as it is passive (not powered, no need to plug anything on the wall outlet)

@milpai

I suggest you insert a passive filter like Muon or Muon Pro between the wall ethernet port and your amazing DIY server. And you’re done!

@milpai

I read recently in these instances a long run of optical cable using an ethernet media converter at each end ensures no unwanted noise inserted along the length.

@zavato

thank you for the directed nature of your comments. Yours is an interesting thought. Question though- do we know that converting to optical doesn’t also include converting “junk” as well?

The optical cable is ‘clean’. My understanding is the media converters are what actually cleans the signal from the router. The optical cable simply keeps that signal clean before being ‘fed’ into the streamer (important to use a very short ethernet cable from the 2nd media converter to the streamer).

I’m a vinyl guy, so no expert here, but did this at the recommendation of someone more knowledgeable about Digital than I.

And a relatively inexpensive solution to obtain better SQ when streaming from the internet. I’m very pleased on the return of investment ($80~100).

 

@bkeske - thank you for the directed nature of your comments. Yours is an interesting thought. Question though- do we know that converting to optical doesn’t also include converting “junk” as well? 

Interesting thread and have a question:

I have a Motorola Modem and Asus ZenWifi mesh routers. I have a switch near the router and I have hard wired the smart TV, Roku, NAS, etc, and one of the ports from this switch has a cat6 cable connected to the wall ether net port. The other end of this port has another ether net cable (~25 feet) all the way to my basement where my audio room is located. I am quite happy with the fact that I was able to get a CAT6 cable routed into the audio room before the drop-in ceiling was replaced with a regular ceiling.

What suggestions do you have to make sure that the noise from the network switch on the floor above is not transmitted into the audio server that makes use of this hard wire? It is not that I hear any noise, but based on previous experiences in cables, I believe that there could be room for improvements.

@zavato

To answer your question, as you are going from your router directly to your single streamer, a high end ‘switch’ may not be necessary. I had the same question, and was advised to get 2 Media Converters, tethered with an optical connection : Router > 1st Media Converter > 2nd Converter tethered with optical > Streamer.

This helped to ‘strip’ unwanted noise from the router, and indeed sounds better. All in? About $100 via Amazon. I’m pleased with the SQ increase.

This is the Q&A I sent to Bluesound, answered by a ’Technical Analyst - Level 3" Scroll down to the end but it seems streaming audio is sent over TCP which I THINK means bit checked packets which if true, is CONTRARY to what I "learned" on this forum.

Question:

I have the Node 130 and mostly stream using Qobuz.

I’m reviewing my modem, router equipment and have a question about streaming protocols.

Does audio data transport to the Node via HTTPS/TCP (so bit checked packets) or a continuous stream using some other protocol?

Answer:

The ports used with the Bluesound Players are as follows;
Required TCP/IP:
· 80 (HTTP)
· 135 (SMB)
· 139 (SMB)
· 1025 (SMB)
· 445 (CIFS)
· 3400 (UPnP incoming events)
· 443 (Rhapsody, Napster, and SIRIUS Internet Radio)
· 4070 (Spotify incoming events)
· 3689 (DAAP for iTunes Sharing and AirPlay)
· 5353 (mDNS connection to mDNS responder)
Required UDP:
· 136-139 (NetBIOS)
· 2000 (BluOS® Proprietary)
· 1900 (UPnP advertisements / device discovery)
· 1901 (UPnP m-search responses)
· 2869 (UPnP connection to Windows Media Player 11 Network Sharing Service)
· 6969 (Getting Started process)
· 10243 (UPnP connection to Windows Media Player 11 Network Sharing Service)
· 10280-10284 (UPnP connection to Windows Media Player 11 Network Sharing Service)
· 11000 (BluOS® Proprietary)
· 11430 (BluOS® Proprietary)
· 5354 (mDNS connection to Bonjour)

Most times, in fact almost all the time, you won’t have to open or forward any ports for our products.

The only exception is enterprise installation, but Bluesound is strictly designed for commercial installs and most routers can handle our product’s required protocols.

The only issue that we see often are routers not supporting Multicast/Unicast which is how our app "locates" players, as without this... you can’t see or use the device.

Streaming on our players operates the same as when streaming from an app, as the stream is sent over TCP for our players to receive the stream.

Actually it is not the case as the Ethernet run is less than 8 inches. Also signal in fiber into my house (frontier) and the 50 feet from the termination in my On Q to my listening room is all fiber. I have 8 inches of ethernet my listening space. 

Also find it interesting that the folks a WBF find this is very beneficial also. Different audience and more receptive. 

https://www.whatsbestforum.com/threads/sotm-snh-10g-switch-mods-tweaks.31092/

https://www.whatsbestforum.com/threads/sotm-snh-10g-network-switches-x2-the-new-king-of-usb-network-gadget-setup.27758/

https://www.whatsbestforum.com/threads/ethernet-tweak.36727/#post-885460

I think you nailed it.  Converting to digital and back may add more noise than the fiber prevents.  

Post removed 
Post removed 

Some questions are better asked on other forums where the trolls are kept a bay or are afraid to weigh in for being made fools of. 

It’s not just about the bits, it’s about the timing of the bits. Everything has a clock in it that supports perfect timing. Even if you have an expensive DAC, someone is going to offer a reclocker to more refine the signal speed. I believe all switches have clocks in them and some people can hear the difference and some people can’t.

My brother is perfectly happy listening to tunes from his phone.  Who am I to tell him he’s wrong and should get a better rig?

@zavato You asked a good question and got good answers.  Unfortunately it is up to you to sort the good answers from the misinformation....which is the key to success on the internet.

 

What the archives are for. @nonoise warned you this thread would grow legs. 

FWIW after all these replies, the most knowledgeable answer I got privately on this question was from an IT guy who is also an audiophile. He wrote to me:

“Being an IT guy, I fall into the camp where I don't see how the switch can make much, if any, of a difference. TCP/IP packets are not guaranteed to arrive in order, the receiving device has to sort them out and then feed them to whatever interface is doing the work. Meaning the idea of jitter is not plausible, since there is no possible timing on out of order packets. All Ethernet switches have galvanic isolation, it's part of the spec, they use transformers on the interfaces. Of course using WiFi exacerbates this, since multiple devices are contending for the same interface, but any streamer that doesn't drop out when playing is buffering the data, which is needed so it can re-assemble the packets, again, tossing out jitter and phase noise since there is no digital audio data stream, just data the computer has to put back together.

Of course I could be wrong, but it's really hard for me to buy into. As an aside, I added a 2.5Gbit switch to my network a couple of weeks ago and my NAS, which is also the Roon server, is connected to that, I noticed no difference, other than I can push data to and from the NAS faster, which was the goal.

As long as all the data arrives in sufficient time for it to be reassembled and fed to buffer, all is well. I have 40mbps download speed on my internet, on a good day that is, it's usually closer to 30, and I stream lossless stuff on qubuz daily, at this very moment in fact and it's rare that it has to stop."

  I started this thread. It's degraded to dibble. Thanks everyone- let's just forget about it. 

Post removed 
Post removed 

Puzzling, maybe for you.

But I would without hesitation join nonoise for a cup of tea.

You just pass on by. 

You're not the easiest to follow or discern. Something's missing. Must be me.

Post removed 

@zavato 

Apologies for the bickering.

Freddy22, panzerwaggin, cakhole, and others (12 many joins) continually try to ruin open discussions by spouting their 'networking' knowledge. Side with them at your peril.

The winners in this? Those who have researched, discussed, and tried network switches; made their mind up from there.

I have two. Well worthwhile in my system.

@erik_squires I think you nailed it.  Converting to digital and back may add more noise than the fiber prevents.  

We just had to splice a fiber line accidentally cut at our plant, and what a pain.  We had $100k worth of special equipment mounted on  it's own truck and 8 technicians.  It was important and time sensitive but wow, what a production.

Jerry

Solely based on @fredrik222’s courage and ability to keep his "emotions" at bay with many members here, I’d score his performance mid 90 percentile.

Now that is rich. He's kept nothing at bay by being haughty, condescending and insulting, twisting the results of past conversations (at least with me), displaying his usual arrogance. He may be your cup of tea but not others.

All the best,
Nonoise

Post removed 

"@yoyoyaya and we are back to bias.

the facts of how a switch works doesn’t change. It is not an intelligent device that can make decisions based on what is transported via the switch."

@fredrik222  That is amusing. Where was it suggested that a switch is an intelligent device?

Though to be fair, my switch doesn't post on this forum so, possibly, that does suggest it's an intelligent device.