Audiophile Network Switches for Streaming ... really ?

[chiily]

Link please. Or search criteria which deliver same as top hit. I want to ensure we’re talking about the same vid, and I want to see what they’re measuring.

There are so many ways that “test” could be limited to certain configs that I’d like to see it. Given that a switch makes a clearly audible difference in my system, it would be disingenuous/simplistic/<insert adjective of choice> to extrapolate from one video to “all systems everywhere”.

sq is just talking about a buffer in the steamer. Simply remove the ethernet cable to the streamer whilst it is playing, and thereby removing any influence the switch might have on the sound....the buffer in the streamer will continue to clock its buffered data through its DAC until the buffer is empty.

Do you notice a difference in the sound when the switch is disconnected and playing its buffered data?

All streamers will have a buffer due to the nature of the network's availability.

 

[chiily]

Sweeping generalisation warning!
I presume you have tried this in your system and didn’t hear a difference. Or perhaps you did!
There could be all sorts of explanations either way but we’d need to know more about your system and your test.

I don't think you do need to know the rest of the system as it is just an A B comparison, where A is with switch connected and B without the switch connected. Once the streamer is disconnected from the network it is just clocking it's cache data, just like it does to all incoming data, it'll be a FIFO queue.

 

[sq225917]

Seems to make sense that if a switch improves things by reducing noise on the connections that flow through it, data and mains, then unplugging it must break these routes entirely, at least making things sound as good, if not perfect.

After all what could be better than a streamer unencumbered by the noise of other electrical connections playing back bit perfect data from its own local buffer...

 

[Fourlegs]

All things being equal, I don't think so.

Unless noise is leaking across PSUs of different devices and the devices are failing to filter them out (course one pays good money for your kit so you'd expect it to have a good PSU, quiet PSU, good A and D grounding - though you wouldn't be the first to be sold a box containing some poor or inadequate engineering for a high price, remembering that marketing slogan, quality costs!), there will be no difference. Unless of course you measured the different device's outputs, subtracted one from the other and forgot to post the result to confirm your findings.

If opinion and hearsay is at the forefront of progress then we are certainly doomed.

I suppose your playback chain is ‘bit perfect’ with no digital processing. In which case have you ever tried different server/player software (but still bit perfect)? The strange thing is that these bit perfect apps sound different. In some cases very different. Will you be trying to tell me that I and the rest of the audio world is also mistaken about that as well? Are you trying to tell me that Roon, Innuos Sense, LMS etc all sound the same?

Indeed do all streamers sound the same to you? It is just the same unaltered bits but I have yet to find anyone faced with hearing two different streamers who would claim they sound the same. Indeed have you tried two different Innuos streamers where the only difference is the HDD vs SSD and an uprated internal power supply, ie the Zen and the Zenith, where the actual processing boards are the same? Those two otherwise identical streamers sound different. Or are you trying to tell me they sound the same? (Edit, I see that a fishy on the USB bridge thread is saying, “Windows SQ is a shit show, always has been.” Surely he is mistaken as well according to you and assuming unaltered bits digital streaming?)

 

[kernow]

Life is far too short to be worrying about this, I use a pi and haven't really thought twice about my wireless network - being a network engineer and all

Oh, and Spotify becauuuse I lovvve that lossy

 

[TheFlash]

sq is just talking about a buffer in the steamer. Simply remove the ethernet cable to the streamer whilst it is playing, and thereby removing any influence the switch might have on the sound....the buffer in the streamer will continue to clock its buffered data through its DAC until the buffer is empty.

Do you notice a difference in the sound when the switch is disconnected and playing its buffered data?

All streamers will have a buffer due to the nature of the network's availability.

Yes I get the principle, I’d just like to see the configuration of the system used. Hence link request.

I can do this but not for very long; certainly not long enough to constitute “evidence” of the robustness some here would demand.

With/without switch in my system is clearly audible so I presume the switch is replicating, or getting close to, the ideal of being disconnected entirely.

Of course it does so in a way which actually allows me to listen to close-to-disconnected music quality for hours on end!

 

[TheFlash]

Seems to make sense that if a switch improves things by reducing noise on the connections that flow through it, data and mains, then unplugging it must break these routes entirely, at least making things sound as good, if not perfect.

After all what could be better than a streamer unencumbered by the noise of other electrical connections playing back bit perfect data from its own local buffer...

Indeed this is all a typical unmanaged (in the jargon) switch just before the streamer can do: break the noise chain to get as close to disconnected from-the-buffer sound quality as possible while staying connected.

 

[John Phillips]

What did you need this for? Addtional devices? As a shorthand way of describing your setup, how long were your cables:

• router to switch
• switch to streamer
• streamer to DAC (if separate)

For the avoidance of doubt, I am not suggesting that cable lengths have anything to do with the digital signal. I merely seek to confirm the way you had your switch installed. Thanks

The D-Link 5-port was only in my Mk2 streaming system and yes, to get an extra port for the streamer because of the other Ethernet devices.

Actually, if it's network topology that matters, I think a more relevant change was between my Mk3 and Mk4 network when I dropped the 15m Ethernet cable from router to streamer in favour of 5 GHz WiFi. I wanted to understand if the WiFi aversion I commonly see had any substance. There was then no Ethernet cable at all to pick up and/or conduct anything that isn't wanted (I think that's what the switch in question is meant to deal with). I did not perceive a change. With either WiFi or Ethernet cable I could still appreciate in detail what was on the recording and relax into enjoying the music. The audio path to my streamer has remained wireless since then.

Perhaps you could characterize the differences you perceive to give me a clue about its significance. An example scale that means something to me is the difference between listening to a Steinway from about stalls row H in the Wigmore Hall (quite close to the stage, 500 or so seats) and stalls row 5 in the Royal Albert Hall (not close but almost as close to the stage as you get if not in the arena, 6,000 or so seats). The only time I have heard a significant difference on that scale was from changing loudspeakers. Updating my network and streamer has never ever even registered on that scale so please forgive me for thinking that re-arranging my network into a new topology and putting back my network switch really isn't for me.

And BTW having listened to a dCS Bartók at my local dealer (which sounded very fine into my loudspeakers' bigger brothers), I still wasn't tempted away from the rest of my humble digital kit (see the information page on my profile).

 

[tuga]

Perhaps you could characterize the differences you perceive to give me a clue about its significance.

I would say that it's top-end related: sometimes "glassy", "hard", "grainy", also "hazy" or loss of sense of decay/space (from Stereophile's Audio Glossary). This may not be readily apparent or even an issue for most people. It's down to one's requirements and expectations, also how much the issues are audibly obvious in a particular system.

From Stereophile's Audio Glossary:

glare An unpleasant quality of hardness or brightness, due to excessive low- or mid-treble energy.
glassy Very bright.
hard Tending toward steeliness, but not quite shrill. Often the result of a moderate frequency-response hump centered around 6kHz, sometimes also caused by small amounts of distortion.
grainy A moderate texturing of reproduced sound. The sonic equivalent of grain in a photograph. Coarser than dry but finer than gritty.
haze, haziness A moderate smearing of detail and focus. The audible equivalent of viewing something through a gauzy veil or a dirty window.

 

[chiily]

Of course it does so in a way which actually allows me to listen to close-to-disconnected music quality for hours on end!

I do too, my streamer is connected via WiFi.

I hear the creaking hinges of another box of opinions opening

 

[Blackmetalboon]

My CD player has a USB input so whether I use CD or a locally stored digital file they all use the internal DAC of the CD player. My set up at home has allowed me to compare the original CD to a ripped version stored on my music server and my media server.

I have compared the following;

CD -> CD Player
Ripped CD -> Music Server -> USB Cable -> CD player
Ripped CD -> Media Server -> Ethernet Cable -> ISP Router -> Ethernet Cable -> Music Server -> USB Cable -> CD Player

I could not detect one iota of difference. My results didn’t change if I disconnected any cables (USB or Ethernet*) when not in use for the tests.

*The cables used were all bog standard, run of the mill cables.

 

[John Phillips]

I would say that it's top-end related: sometimes "glassy", "hard", "grainy", also "hazy" or loss of sense of decay/space (from Stereophile's Audio Glossary). This may not be readily apparent or even an issue for most people. It's down to one's requirements and expectations, also how much the issues are audibly obvious in a particular system. ...

Your comment is interesting. Looking back at my personal notebook from 2004 when I auditioned some CD players, they all sounded fine except one about whose sound I wrote "glassy". However, I haven't ever heard "glassy" from any of my systems - regardless of the networking arrangement.

I did drop that CD player from my short list just in case that defect which I had to listen for (so quite small relative to my "Steinway piano sound" scale) might annoy me after having recognized it; even when listening to music rather than listening to equipment.

 

[miktec]

I suppose your playback chain is ‘bit perfect’ with no digital processing. In which case have you ever tried different server/player software (but still bit perfect)? The strange thing is that these bit perfect apps sound different. In some cases very different. Will you be trying to tell me that I and the rest of the audio world is also mistaken about that as well? Are you trying to tell me that Roon, Innuos Sense, LMS etc all sound the same?

Indeed do all streamers sound the same to you? It is just the same unaltered bits but I have yet to find anyone faced with hearing two different streamers who would claim they sound the same. Indeed have you tried two different Innuos streamers where the only difference is the HDD vs SSD and an uprated internal power supply, ie the Zen and the Zenith, where the actual processing boards are the same? Those two otherwise identical streamers sound different. Or are you trying to tell me they sound the same? (Edit, I see that a fishy on the USB bridge thread is saying, “Windows SQ is a shit show, always has been.” Surely he is mistaken as well according to you and assuming unaltered bits digital streaming?)

....... and if my horse was a donkey would it still have no horns?

 

[miktec]

I could not detect one iota of difference. My results didn’t change if I disconnected any cables (USB or Ethernet*) when not in use for the tests.

*The cables used were all bog standard, run of the mill cables.

I reported precisely the same (lack of change) result way back on page 11 (and others also have)

Despite repeated attempts, there was absolutely no discernible difference to the sound with or without an ethernet switch.

.... obviously I hadn't incanted the right spell or scattered enough flower petals.

To give them their dues though - the faithful are dogged in the pursuit of our 'enlightenment' ... thread now up to 45 pages (and it's not even Xmas yet...)

 

[chiily]

I suppose your playback chain is ‘bit perfect’ with no digital processing. In which case have you ever tried different server/player software (but still bit perfect)? The strange thing is that these bit perfect apps sound different. In some cases very different. Will you be trying to tell me that I and the rest of the audio world is also mistaken about that as well? Are you trying to tell me that Roon, Innuos Sense, LMS etc all sound the same?

Indeed do all streamers sound the same to you? It is just the same unaltered bits but I have yet to find anyone faced with hearing two different streamers who would claim they sound the same. Indeed have you tried two different Innuos streamers where the only difference is the HDD vs SSD and an uprated internal power supply, ie the Zen and the Zenith, where the actual processing boards are the same? Those two otherwise identical streamers sound different. Or are you trying to tell me they sound the same? (Edit, I see that a fishy on the USB bridge thread is saying, “Windows SQ is a shit show, always has been.” Surely he is mistaken as well according to you and assuming unaltered bits digital streaming?)

Nope, I'm not talking about DSP or whatever from different servers, players, whatever. No need to confuse the question. Hence the phrase, all things being equal, that means the same, constants, which one needs to be able to isolate the variable/effect. Simple experimental procedure.

With everything the same apart from the network switch changing, play a piece of audio, capture the output. Then, with not changing anything else, change the switch, play the same piece of audio, capture the output. Subtract the first capture from the second and post the graph or table of variance here.

And do that at least ten times, a small sample, but hey, better that doing it just once!

Honestly that simple...go for it...

Oh, and record the system in test, the configs the devices, etc so the test is repeatable.

 

[Zombie]

Don't forget to daisy chain some isolators, too

 

[chiily]

Yes I get the principle, I’d just like to see the configuration of the system used. Hence link request.

I can do this but not for very long; certainly not long enough to constitute “evidence” of the robustness some here would demand.

You could simply record the output and diff between with cable and without cable. The difference should be clear if it is at a level that is audible.

 

[Roog]

It is like all of these 'fix-it' type products, do any of these people know if there was a problem to be fixed in the first place?

 

[TheFlash]

It is like all of these 'fix-it' type products, do any of these people know if there was a problem to be fixed in the first place?

This presumably rhetorical question has been asked several times in this thread.

 

[TheFlash]

You could simply record the output and diff between with cable and without cable. The difference should be clear if it is at a level that is audible.

You mean with a microphone? Is this what happened in the test which I can apparently find online?