The Ethernet cables, Switches and Network related sound thread. Share your listening experience only.

[griff500]

I thought this was a thread for sharing your personal experience…

 

[audiobomber]

I thought this was a thread for sharing your personal experience…

You're right. I'm still auditioning the Supra+ and AS Signature cables, not ready to say for a few more days.

 

[chesebert]

The key factors in Ethernet performance: transformer coupling on both ends and correctly built shielded cat6 or better cable with metal jackets.

 

[Mark200]

correct but most people cant or dont want to understand how it works. there is a whole complex chain of events with networking to make sure that the bytes of data sent from one end get to the other end intact but you can still get corruption of data. your not sending 1.s and 0.s your sending algorithms and code in a voltage/light carrier in chunks of data which is compiled/decompiled. depending on the compiler. yes at the very core of every compiler you are dealing at machine code level and 1-0. you need the same one at the other end to decompile. then there is handshaking, memory stacks so on and so on. its all a very very complex matter and works at the blink of an eye. if you are truly interested then take an online course with CISCO.

instead of voltage or light wireless networking uses radio waves as the carrier signal.

IIR audio data also has modulation of the 1-0 so a 1 can be high voltage where the next 1 can be low voltage but 0 is always off (sort of).

but hey i could be wrong......

Cisco network devices are typically much more complex that one would use in a home network, or even in a music studio. That is because they are typically "managed" devices that run code to continuously monitor the performance, provide security, report on errors, etc.

 

[Mark200]

Digital audio is always so interesting to read about. I still do not understand switches and never had one.

Switches for home networks just give a way to get more than 4 hard-wired Ethernet connections to your LAN than is typically provided by your ISP supplied WiFi Router. So a switch is kind of like a power strip extender (but for Ethernet, not power) that you connect to one of the 4 Ethernet connections from your ISP supplied router to get more Ethernet connections to your LAN.

A managed switch from companies like Cisco are used on commercial network applications and can have remote monitoring, security, and other features built into them, not suitable for home use.

Audiophile switches sit between the router and end component, and supposedly remove impurities in the digital signal (which is not possible).

 

[F208Frank]

Switches for home networks just give a way to get more than 4 hard-wired Ethernet connections to your LAN than is typically provided by your ISP supplied WiFi Router. So a switch is kind of like a power strip extender (but for Ethernet, not power) that you connect to one of the 4 Ethernet connections from your ISP supplied router to get more Ethernet connections to your LAN.

A managed switch from companies like Cisco are used on commercial network applications and can have remote monitoring, security, and other features built into them, not suitable for home use.

Audiophile switches sit between the router and end component, and supposedly remove impurities in the digital signal (which is not possible).

When you wrote (which is not possible) are you implying that switches do not much?

Or am I absorbing your information wrong?

 

[iFi audio]

Its electrical signals. High or low voltage. Yes in binary code form.

That is factually correct

 

[Mark200]

When you wrote (which is not possible) are you implying that switches do not much?

Or am I absorbing your information wrong?

You mean do not "do" much? They are like splitters of Ethernet networks to expand the number of connections on the LAN, sort of like Ethernet hubs. Although hubs also expand the number of possible hardwire connections in a Ethernet network, the total bandwidth of all the combined connections is limited to the bandwidth of a single main line going into the hub, but a switch is much more efficient when there are multiple connections to the switch. Switches are more used in large enterprises for that reason.

Audiophile switches claim to clean up the signal for better sound. Some people claim to hear differences in the sound with these audiophile switches, which I don't understand because the the signal is in digital format and since it is Ethernet and TCP/IP, there can be no data corruption or errors (if there are errors, the data is automatically retransmitted), unless there is a total failure to connect. That doesn't mean one can't hear difference is cables that carry analog signals, which is certainly possible.

The only way a digital signal transmitted by Ethernet and TCP/IP through an Ethernet cable or audiophile switch can sound different than no audiophile switch is if the switch actually changes the digital signal (Ethernet cables don't have any electronics in them so they cannot change the digital information). It is possible that they reduce the amount of electrical noise along the cable path, but that noise is not in the digital signal, and any downstream device can just as easily have noise filters to keep out any noise that travels with the cable (but outside the digital data).

 

[AxelCloris]

Folks, this thread has shifted too far from listening impressions and personal experiences, so let's please bring the discussion back on topic. If anyone would like to debate non-subjective thoughts, we welcome you to join one of the threads already underway in Sound Science or to start a discussion thread there of your own. Let's please keep this thread on topic of the OPs original question regarding first-hand listening experiences between components. Thanks everyone.

 

[bluenight]

I thought this was a thread for sharing your personal experience…

I see you have the innous phoenixnet. Any good? heard a difference from what you previously used?

 

[cpurdy]

correct but most people cant or dont want to understand how it works. there is a whole complex chain of events with networking to make sure that the bytes of data sent from one end get to the other end intact but you can still get corruption of data. your not sending 1.s and 0.s your sending algorithms and code in a voltage/light carrier in chunks of data which is compiled/decompiled. depending on the compiler. yes at the very core of every compiler you are dealing at machine code level and 1-0. you need the same one at the other end to decompile. then there is handshaking, memory stacks so on and so on. its all a very very complex matter and works at the blink of an eye. if you are truly interested then take an online course with CISCO.

instead of voltage or light wireless networking uses radio waves as the carrier signal.

IIR audio data also has modulation of the 1-0 so a 1 can be high voltage where the next 1 can be low voltage but 0 is always off (sort of).

but hey i could be wrong......

You got most of the terms wrong, but I almost kind-of follow what you're trying to say.

And at a low level, there are no ones and zeros; there's just voltage. And for the last 20 years, it's not even about off vs. on (e.g. 0v vs .5v). Rather, there are multiple values encoded into a wave. So where we used to send 1 value in every wave, we might send 4 now, or more. Conceptually, those 4 values are 0 or 1, but in reality, it's more like a number from 0..15 (4 binary digits, i.e. 2^4 potential combinations). This is how the PCIe bus has kept doubling in throughput, for example, with the same number of wires: it is now sending data on the leading edge and the trailing edge of each wave (that is a simple doubling of data rate), etc.

I have found a noticeable difference in the sound produced when using different ethernet cables. Many others have reported the same (this being what the thread is supposed to be about). Some participants persist in banging on about data integrity, which is not what's being discussed.

I don't mean to be pedantic, but you have not found a noticeable difference in the sound produced when using different ethernet cables.

You may happen to like the same sound more when you plug in a more expensive ethernet cable, but the sound is exactly the same.

And your increased enjoyment that results from a more expensive cable makes your purchase worth it; there's nothing fake about the placebo effect. I have no problem with your approach; I do the same exact stupid thing with beautiful expensive cables (carrying digital data) that cannot make any difference in the resulting sound. Of course I enjoy the sound more when the cables are prettier; who wouldn't?!?

Yes but as soon as the data is received, it is stored in a buffer, which can ONLY store a binary value, so noise is eliminated because there is no state that can represent it. Obviously, when that data is sent onwards to the next device in the chain, as a digital (electrical) signal, it can pick-up some noise/interference again, until it enters the data buffer of that next device and so on down the chain until the data hits the data buffer in your DAC. This removal of noise picked-up in transit is why digital was invented in the first place.

It's even less noisy than that. As it is passed around and processed, digital data is never "stored" or "held" in a way that degrades, no matter how many trillions of hops it takes, and no matter how long it is held. Whether it's in a register in a CPU, or in a DRAM cell, that 0 is "any charge kind of close to 0", and that 1 is "any charge that isn't around zero". And many times every second, an automatic process occurs (called a refresh, e.g. see "DRAM refresh rate") that "tests" the data, and then "sets" the data to a "zero" or "one", based on whether the voltage was around zero or not.

So basically, a digital device like an audio player is juggling 100,000,000s of "hot potatoes" (bits) many many times every second, and never, ever, ever drops one of them. And every time it juggles one, it "re-cleans up" the bit.

(OK: Technically, a cosmic ray might randomly flip a single bit in your computer once every couple of months. See https://ieeexplore.ieee.org/abstract/document/658626 and https://www.science.org/doi/10.1126/science.206.4420.776 ... this can be a design issue in building hardware for outer space, for example.)

Like with anything i think there can be gradations of how much noise can enter the audio device. Can anything exist without gradations? Digital audio/video through ethernet does have a very High minimum level quality that will satisfy most people. But can be improved upon in my experience.

There are two different issues here.

First, there is the data being transmitted and received, and it is always received bit perfect. (Because if it is not received bit perfect, it is not received at all. That's why TCP exists: any packet that is not 100% bit perfect is just tossed, which forces TCP to re-send the tossed packet until it is received bit perfect. 99.99% of packets come through bit perfect on the first try. Do the math: 99.999999% of packets are bit perfect on the first retry, and 99.99999999999999% on the second retry, and so on.)

The key factors in Ethernet performance: transformer coupling on both ends and correctly built shielded cat6 or better cable with metal jackets.

I'm not really sure what the key factors are on the hardware side, and I work with this stuff. Each generation has tighter and tighter tolerances; just read the specs and tolerances for CAT5 vs 5e vs 6 vs 6e vs 7 vs 8 vs 8e.

The main thing, for this particular conversation, is that the data that comes through is digital and it is transferred bit perfect. There are no errors. There is no noise. No matter how bad the power supplies are. No matter how crappy the cables are. The data either comes through bit perfect, or not at all.

So the very first step in the chain that can affect audio quality is the DAC, because up to that point the data is digitally communicated in a bit perfect manner, whether that is Ethernet or USB or coax or fiber or whatever.

 

[gregorio]

As it is passed around and processed, digital data is never "stored" or "held" in a way that degrades, no matter how many trillions of hops it takes, and no matter how long it is held. …

Yes, I realise that. I was referring to the digital signal picking up noise/interference in transit though the Ethernet cable/s, much of which is rejected due to the differential signalling but what’s left is then removed by the process you described.

Back on topic though, there can be a difference in sound between Ethernet cables. Nothing to do with the digital signal passing through them of course but they do make some sound when you plug them in or step on them. Based on listening, a marginally different click or crunching sound, probably only once or twice in the life of the cable, is not worth buying more expensive Ethernet cables for me personally.

G

 

[griff500]

You got most of the terms wrong, but I almost kind-of follow what you're trying to say.

And at a low level, there are no ones and zeros; there's just voltage. And for the last 20 years, it's not even about off vs. on (e.g. 0v vs .5v). Rather, there are multiple values encoded into a wave. So where we used to send 1 value in every wave, we might send 4 now, or more. Conceptually, those 4 values are 0 or 1, but in reality, it's more like a number from 0..15 (4 binary digits, i.e. 2^4 potential combinations). This is how the PCIe bus has kept doubling in throughput, for example, with the same number of wires: it is now sending data on the leading edge and the trailing edge of each wave (that is a simple doubling of data rate), etc.


I don't mean to be pedantic, but you have not found a noticeable difference in the sound produced when using different ethernet cables.

You may happen to like the same sound more when you plug in a more expensive ethernet cable, but the sound is exactly the same.

And your increased enjoyment that results from a more expensive cable makes your purchase worth it; there's nothing fake about the placebo effect. I have no problem with your approach; I do the same exact stupid thing with beautiful expensive cables (carrying digital data) that cannot make any difference in the resulting sound. Of course I enjoy the sound more when the cables are prettier; who wouldn't?!?


It's even less noisy than that. As it is passed around and processed, digital data is never "stored" or "held" in a way that degrades, no matter how many trillions of hops it takes, and no matter how long it is held. Whether it's in a register in a CPU, or in a DRAM cell, that 0 is "any charge kind of close to 0", and that 1 is "any charge that isn't around zero". And many times every second, an automatic process occurs (called a refresh, e.g. see "DRAM refresh rate") that "tests" the data, and then "sets" the data to a "zero" or "one", based on whether the voltage was around zero or not.

So basically, a digital device like an audio player is juggling 100,000,000s of "hot potatoes" (bits) many many times every second, and never, ever, ever drops one of them. And every time it juggles one, it "re-cleans up" the bit.

(OK: Technically, a cosmic ray might randomly flip a single bit in your computer once every couple of months. See https://ieeexplore.ieee.org/abstract/document/658626 and https://www.science.org/doi/10.1126/science.206.4420.776 ... this can be a design issue in building hardware for outer space, for example.)


There are two different issues here.

First, there is the data being transmitted and received, and it is always received bit perfect. (Because if it is not received bit perfect, it is not received at all. That's why TCP exists: any packet that is not 100% bit perfect is just tossed, which forces TCP to re-send the tossed packet until it is received bit perfect. 99.99% of packets come through bit perfect on the first try. Do the math: 99.999999% of packets are bit perfect on the first retry, and 99.99999999999999% on the second retry, and so on.)


I'm not really sure what the key factors are on the hardware side, and I work with this stuff. Each generation has tighter and tighter tolerances; just read the specs and tolerances for CAT5 vs 5e vs 6 vs 6e vs 7 vs 8 vs 8e.

The main thing, for this particular conversation, is that the data that comes through is digital and it is transferred bit perfect. There are no errors. There is no noise. No matter how bad the power supplies are. No matter how crappy the cables are. The data either comes through bit perfect, or not at all.

So the very first step in the chain that can affect audio quality is the DAC, because up to that point the data is digitally communicated in a bit perfect manner, whether that is Ethernet or USB or coax or fiber or whatever.

I did not see any first-hand experience related in your post?

Telling me that I have not heard what I have heard is not being pedantic - it is something else.

The DAC is not the first thing in the chain that can affect audio quality as it is not the first thing in the audio chain.

Again, please post your first-hand experience only - that is what the thread is about. As has been suggested by moderators on several occasions, if you want to discuss theory then there is a sound science forum for that.

 

[moriez]

I don't mean to be pedantic, but you have not found a noticeable difference in the sound produced when using different ethernet cables.

You may happen to like the same sound more when you plug in a more expensive ethernet cable, but the sound is exactly the same.

And your increased enjoyment that results from a more expensive cable makes your purchase worth it; there's nothing fake about the placebo effect. I have no problem with your approach; I do the same exact stupid thing with beautiful expensive cables (carrying digital data) that cannot make any difference in the resulting sound. Of course I enjoy the sound more when the cables are prettier; who wouldn't?!?

He said he did hear a difference, and you're saying he isn't. Okay man.. How person A precisely knows what person B is experiencing is a trip. Possibly a big help to humanity!

 

[griff500]

I see you have the innous phoenixnet. Any good? heard a difference from what you previously used?

I was not able to assess this as I would usually like to do as I bought and installed it a week before I moved to a new home in a different country. I seem to recall that my initial impressions were that I felt there was a difference but it was not as obvious as the ethernet cable, but I then broke the system down and shipped it to the new place and then simply set everything up again, so I hadn't really assessed it properly and I've been too busy to do so since moving.

I was previously using a cheap ethernet 'hub', so the Sean Jacobs power supply in the PhoenixNET might be a major factor in any differences that people notice and the rest is perhaps better quality connectors, assuming that's the case. Just speculating...

I am now used to the sound of my system in it's new location so if I can motivate myself to do so I'll remove it from the chain and see if I notice a difference.

I am getting internet over a wireless (mobile) system in the new location, whereas it was over the usual telephone cables in my previous location, although it was not connected to a telephone and I don't think this should be a factor. It's perhaps worth noting this change though.

So to conclude, I cannot provide a conclusion at this time.