I was told many years ago that routers are good at routery bits, bit not so good at the switchy bits, hence why adding a separate switch can help. Maybe that’s old hat now of course.
My thinking on this, and I have to admit my real knowledge of computers is effectively nil, is that as my nas and streamer are connected via a switch, I can turn the router off and the music continues happily and it sounds no different whether the router is on or off. So unless the router is deliberately throwing out evil signals down the wire to the switch, it cannot make any difference. If, on the other hand, everything is connected to the router, the router could well make a difference, just as a switch can.
“How can 0 and 1 be affected by electrical noises?”
The 0 and 1 themselves exist as analog waves of electricity, as Dozey notes above.
So electromagnetic noise is the bath of other electrico-magnetic activity in which those digital signals exist and are transmitted.
This is also (literally) streaming through the wires and circuits from your broadband entry point into the house and from your connected electrical circuits into numerous parts of your hifi right up to the speakers (minus any optical sections).
It’s that background unintended noise that alters the operation of your DAC and other parts of your system.
[Noises are something else - like your neighbour’s dog barking].
A 0 is a 0, and an 1 is an 1, there is nothing in between, there is no noise whatsoever that can change the data, and if the inbound data to the streamer is not what the streamer is expecting, it will be rejected. That is all I am saying.
If you believe that you can hear the differences when you switch some switches, then it is fine, it is your money as I said earlier.
Not sure about your background, but I suggest you try to enroll some courses about TCP/IP network or socket programming, or at least some IT networking courses?
A very noisy system will sound worse than a very low noise system, not just different, but worse.
Switches, no matter how good, can only add noise, never remove it.
The above two beliefs would indicate then that removal of switches should lower noise entering your streamer, and hence might benefit SQ.
There are usually other devices, unrelated to music streaming, attached to your home network (Sky box, computers, back up NAS, etc), that could introduce noise.
Disconnecting these unrelated (to music reproduction) devices from your home network (even temporarily) would remove the noise they introduce.
This would leave just a server/music store and a streamer connected to your router.
If, and it is a big if, you believe in all of the above, then removal of switches and unnecessary (for music) devices, could improve SQ, right?
If this ‘device pruning’ actually reduces noise, then is it possible to ‘split’ the network into two within the router so you end up with one network dedicated to music and another network for everything else? If so would the other network still introduce noise to the music network?
Thoughts please before I rip out my two switches, Sky box, NAS and Mac Book Pro. (I suspect the Sky Box and the Mac Book are noisy bu**ers).
I know, sorry.
I thought it to be such an innocent question, I did not for a minute think it would be like poking a wasps nest.
Certainley not feeling the love.
I will stick to knitting fur balls for my pot cats
Hi, in a static file of binary data you are completely correct.
We are discussing serial data streams, where 1 and 0 are analogue voltages and are separated by time controlled by clock which will have noise… now upto a certain amount of noise energy the stream can be recovered, but that noise will couple into connected systems.
Also it’s worth remembering that with modern Ethernet ‘1’ and ‘0’ are represented by many interim analogue states such as word coding, phase and voltage… and even with SPDIF a 1 can look like either 0 or 1 depending where it is in the bit stream sequence… so digital 1 and 0 are a high level encoding supported by analogue physical variables.
Of course all of this gets in the way of the simplified view of ‘digital’ for the layman… but the real world is far from simple, and can have many side effects, which I am sure you agree.
Thank you for that - it is the first time (I can remember) anyone explaining it so clearly. I had also thought (erroneously it seems) that it was a binary thing
Gary… I think people get confused and think of the network digital protocols and digital encodings introducing or creating noise … of course that doesn’t happen … unless there is information/data corruption which the protocol state machines themselves will recover the lost data or close the connection. This leads to layman argument of 1 and 0s that you see in simplified texts ad nauseam . But of course this oversimplification completely misses the point… and ultimately I believe causes more confusion.
The digital data has to be transported and encoded over the network. The physical encoding uses good old traditional analogue physical values. Depending on the medium used there are side effects in encoding these analogue values. It is these dried effects of the physical analogue encodings that can interfere with sensitive audio electronics by modulating the ground for example.
There are many engineering design white papers in existence that suggest approaches to mitigate this side effects and reduce coupled interference. Of course this is not unique to voice, but any sensitive equipment like medical and sensitive radio devices for example.
With audio the noise distortion caused is non harmonic, and is more noticeable to our auditory systems.
So as I say it has little or nothing to do with local area network itself, but it has to do with the physical environment and physical designs used to support the network. It’s this physical electrical analogue layer than can cause noise through coupling… completely independent to the digital values carried by the higher level network protocols.
Given that you can see I am far from sitting on a fence, but being assertive in as accurate way as I can be appropriate for this forum.
Noise can be measured and displayed. Spectrum Analyzers are useful. Bandscopes are useful for looking at coupled RF harmonics.
The consequence of noise can be measured as well such as seeing the frequency variability or modulation introduced into a clock through powerline, EM or groundplane modulation… again an Oscilloscope or Spectrum Analyzer can be effective.
Because of the non harmonic nature of digital noise the human ear is good at detection, but not necessarily measurement.
