Ethernet cables ...is there any difference in SQ?

[Bonky]

Following on from another thread...

...well, the title says it all.

Thanks to anyone who responds, and, yes, I know there have been previous mentions on the subject- but is there any consensus?

Thanks,

Bw

Bonky

[Stratmangler]

Why don't you try it and report back?

[StanleyB]

Going back to the days when I had to design network cables, and later test the production versions of the proto types, there were wild variations in transmission speeds between different wrapping configurations. There were also wild variations in speed between stiff and flexible leads. And the same applied to the RJ45 connectors.
On a side note: I am the guy who invented the multi-coloured RJ45 patch panel, but failed to make anything out of it. I didn't think that it was possible to patent an idea like that at the time.

[Radford Revival]

Even given differing electrical characteristics of two different ethernet cables, there are so many layers that sit between any of these and the output of the DAC in the device playing music. The ethernet chipset in whatever device receives the raw packets over the wire, which if error free are then handed over to the kernel of whatever operating system is running the device via a bus completely different than ethernet.

The OS, likely some kind of linux if on a streaming box, parses the packets, generally moves things around in memory and abstracts them down to a continuous stream of data for whatever process or program is dealing with receiving the audio data over the network. It will also transparently perform error correction for the most common type of network stream. The process/program dealing will be receiving these chunks of data piecewise in differing sizes mainly depending on network traffic and how the kernel decides to pass along the data, but this isn't an issue as it will employ some kind of buffering.

Any decompression / decoding of any particular file format is then performed and converted to audio samples that sit in memory. Then perhaps some processing is done to the audio stream. Then this audio data is passed onto the sound driver via the kernel, which then may perform further manipulation on the audio stream (likely nothing to the waveform itself but the way it's formatted in memory). This data is then transferred to whatever audio chipset is in the device. During this process, the system is doing countless other things so the various memory areas and buffers are being filled at a fairly jittery (but on average, constant) rate.

So the notion that an ethernet cable can make a difference, or two functioning ethernet cables can sound different is absolutely absurd!

[ALRAINBOW]

Buy cat 6A shielded. Ready made patch cables. This is what I use for interconnects. But honestly I do not hear any difference but atleast I am not spending hundreds on just one cable.
Al

[Bonky]

Going back to the days when I had to design network cables, and later test the production versions of the proto types, there were wild variations in transmission speeds between different wrapping configurations. There were also wild variations in speed between stiff and flexible leads. And the same applied to the RJ45 connectors.
On a side note: I am the guy who invented the multi-coloured RJ45 patch panel, but failed to make anything out of it. I didn't think that it was possible to patent an idea like that at the time.

I tried to reply to Stan's response last night but it doesn't appear here so I'll try again...

...thanks Stan. Do you mind my asking which of the various configurations you preferred?
If shielding is an issue here, is it better for the earth connection to be linked at one or both ends?

I've just read the detailed response from RadfordRevival for which I'm grateful if somewhat overawed by. As there seems to be so much BlackBox gubbins in the way, is wireless a 'purer' option after all? I get drop-outs at present from using wireless and was willing to experiment with Ethernet. (My 'thunderbolt to Ethernet adaptor only arrived yesterday so haven't been able to test it. (The only Ethernet cables I have are the 'bog-standard ones that are given away with various items.))

Thanks to all,

Bonky

[Radford Revival]

Ethernet will likely result in far fewer dropouts than wireless if you're experiencing issues so it's definitely worth a try.

Just for the record, the processing that takes place between the cable and what ultimately arrives at the DAC doesn't matter, I was just simply demonstrating the sheer number of steps that occur between the two (wireless has a few more steps). As long as the buffers are being filled on time, it simply doesn't matter how it happens, as long as the DAC gets an uninterrupted data stream.

The actual data on the ethernet wires, or wireless, does not resemble a pure audio stream or anything close to it. The data is chopped up into packets that may or may not even be in the correct order, and may even have pieces missing (which thanks to TCP/IP or other streaming protocol will quickly get re-sent). Various other protocols are occasionally chattering back and forth at the same time on the same cable. The layers of hardware and software between the server and playing software handle the data stream with utter indifference toward the fact that it contains audio, they only care if it's error free. As I said, this doesn't matter at all, it's just how it works.

Very similar principles also apply to things like USB or firewire cables. If an ethernet cable was so 'special' it was somehow affected the data on it, it wouldn't work. For all intents and purposes it's literally impossible for data cables to make a difference in sound quality, barring things like dropouts from very poor (ie broken) cables with a high error rate.






I tried to reply to Stan's response last night but it doesn't appear here so I'll try again...

...thanks Stan. Do you mind my asking which of the various configurations you preferred?
If shielding is an issue here, is it better for the earth connection to be linked at one or both ends?

I've just read the detailed response from RadfordRevival for which I'm grateful if somewhat overawed by. As there seems to be so much BlackBox gubbins in the way, is wireless a 'purer' option after all? I get drop-outs at present from using wireless and was willing to experiment with Ethernet. (My 'thunderbolt to Ethernet adaptor only arrived yesterday so haven't been able to test it. (The only Ethernet cables I have are the 'bog-standard ones that are given away with various items.))

Thanks to all,

Bonky

[Mark Pope]

The only Ethernet cables I have are the 'bog-standard ones that are given away with various items....



Stick with them and spend the money you would otherwise spend on fancy data cables on music instead.

[Clive]

I have heard small differences between CAT5 , 6 and 7 when running a dual laptop setup. These were just cheap patch cables, not audiophile exotica. It's hard to credit that digital transmission is the issue, my gut feeling is that this is to do with shielding, RFI and grounding which finds its way into the analogue side. The specs of these cables vary mainly in grounding and shielding as well as the quality of the plugs which in the better specs are metal. It seems that to achieve higher network transmission rates better cable specs are needed so noise can be an issue. It may therefore can the capability to impact sound quality in some way.

[Stratmangler]

I have heard small differences between CAT5 , 6 and 7 when running a dual laptop setup.

How is this dual laptop setup configured, and is it standalone?
You've mentioned it in the past ....

[AlanS]

If you get WiFi drop outs have you got/tried a dual band router or maybe have a lot of devices that use the WiFi router to connect to your network. WiFi isn't purfict but can be good when their isn't too much load due to devices. On the other hand the purist will be happier with a well wired network (I have cables as stand by and rarely have the need)

[Clive]

How is this dual laptop setup configured, and is it standalone?
You've mentioned it in the past ....
It was when I used jplay with the laptop connected to dac operating effectively as a massive buffer. Now that I use MQn I only need one laptop and no ethernet.

[Stratmangler]

It was when I used jplay with the laptop connected to dac operating effectively as a massive buffer. Now that I use MQn I only need one laptop and no ethernet.

Thanks Clive.
I just wondered how the 2 laptop setup was all connected up. I'm trying to work out what could have an effect on sound quality.

[Clive]

Thanks Clive.
I just wondered how the 2 laptop setup was all connected up. I'm trying to work out what could have an effect on sound quality.
The only logic I can put to it is as I mentioned...grounding / noise having an effect on my DAC (which has GI).

[Stratmangler]

Cat7 cabling has lots of shielding in it.
Each of the 4 pairs has its own foil screen, and then there's an overall foil screen around them.

You can be certain that the shielding was the reason for the difference you heard.
My guess is that it made for a common ground plane between the 2 machines, and anything connected off the back of them, such as a DAC.
Your two laptops will have had floating ground, probably your DAC too.
Tie all the grounds up (as the shielding in the Cat7 cable will have done) and they're all floating ground together - their reference ground is common between the 3 devices.
You could probably have done the same thing with a wire jumper between the devices, but that would no doubt mean opening up the various bits of kit to find the ground points - I'd take an easier route, as you did.

[Reffc]

Will (Radford Revival) is absolutely spot on here.

I've heard countless queries on whether ethernet cables make a difference and also the somewhat mythical idea that the differences that some people claim to here are all down to the shield/grounding. That's utterly wrong. Ethernet cables do not pass an analogue signal and are not subject to anything like the S/N issues of a single ended analogue system where S/N can and is affected by shield effectiveness and impedance. Will has it right. The data packets arrive somewhat randomly and as long as they arrive without the redundancy on the packets being used up (ie the signal doesn't drop out) then there can be no difference between ethernet cables. I dont make, or otherwise offer them for sale in a hifi context as I feel that would be a dishonest misrepresentation to claim a "hifi ethernet" sounds better than a £1.99 one from Tandys. It simply wouldn't and couldn't make a difference, but if folks feel that they need to part with hard earned to have that "hifi" peace of mind tag, knock yourselves out ;). You'd be better advised putting the cash into music folks. Now analogue signals are a completely different story, as to a lesser extent, are S/PDIF where impedance and shielding are more important.

[Clive]

Will (Radford Revival) is absolutely spot on here.

I've heard countless queries on whether ethernet cables make a difference and also the somewhat mythical idea that the differences that some people claim to here are all down to the shield/grounding. That's utterly wrong. Ethernet cables do not pass an analogue signal and are not subject to anything like the S/N issues of a single ended analogue system where S/N can and is affected by shield effectiveness and impedance. Will has it right. The data packets arrive somewhat randomly and as long as they arrive without the redundancy on the packets being used up (ie the signal doesn't drop out) then there can be no difference between ethernet cables. I dont make, or otherwise offer them for sale in a hifi context as I feel that would be a dishonest misrepresentation to claim a "hifi ethernet" sounds better than a £1.99 one from Tandys. It simply wouldn't and couldn't make a difference, but if folks feel that they need to part with hard earned to have that "hifi" peace of mind tag, knock yourselves out ;). You'd be better advised putting the cash into music folks. Now analogue signals are a completely different story, as to a lesser extent, are S/PDIF where impedance and shielding are more important.
This is not about data packets. What I'm saying is that grounding / noise is to do with the DAC and quite probably the rest of the system - I'm meaning the analogue part of the DAC and the preamp / power amp. To dismiss this aspect is to suggest that grounding and noise is not important across a hifi system. That would be wrong. Very wrong.

BTW my cables are patch cables sourced from Amazon, mostly 99p to a couple of quid.

[Reffc]

You have mis interpreted what Will and I both have tried pointing out. We were responding wrt the OP. Shielding types on an ethernet dont vary SQ and no-where HAVE I dismissed grounding as unimportant. Re-read and you'll see I have alluded to it being important in analogue but the OP was specifically "do all ethernet cables sound the same" (in effect) to which the answer is "ethernet cables don't have sound (sound of anything!) nor contribute to analogue SQ provising that there is no data drop out"...that is all that we're trying to say. To mis-represent what was said is the absurdity, as nowhere have Will or I suggested that grounding isn't important in a hifi system ;)

[Clive]

You have mis interpreted what Will and I both have tried pointing out. We were responding wrt the OP. Shielding types on an ethernet dont vary SQ and no-where HAVE I dismissed grounding as unimportant. Re-read and you'll see I have alluded to it being important in analogue but the OP was specifically "do all ethernet cables sound the same" (in effect) to which the answer is "ethernet cables don't have sound (sound of anything!) nor contribute to analogue SQ provising that there is no data drop out"...that is all that we're trying to say. To mis-represent what was said is the absurdity, as nowhere have Will or I suggested that grounding isn't important in a hifi system ;)
Likewise I'm not saying that cables have an intrinsic sound and BTW the same applies to analogue cables. Any SQ differences are down to how the equipment used reacts to the difference between cables. I would also say that a better shielded cable (eg CAT6 or CAT7 vs CAT6) stands a more chance of not negatively impacting SQ by picking up less noise. The better shielding is there to reduce noise pickup.

[Reffc]

Likewise I'm not saying that cables have an intrinsic sound and BTW the same applies to analogue cables. Any SQ differences are down to how the equipment used reacts to the difference between cables. I would also say that a better shielded cable (eg CAT6 or CAT7 vs CAT6) stands a more chance of not negatively impacting SQ by picking up less noise. The better shielding is there to reduce noise pickup.

The mechanisms Clive for noise pickup to affect S/N in ethernet cables are very different from analogue, hence I hope you'll forgive my scepticism. In analogue, the noise is via inductive coupling with the signal direct to the analogue signal from the shield, so reducing (effectively) S/N ratio. The same is not true of digital. Indeed, many ethernet cables aren't shielded at all (twisted pairs are used) and even those that carry shielding often have no connector mechanism for common grounding along the signal path to the rest of the chain...if they did, some ground loop issues would become problematical. Pick-up of noise is more a common issue with digital switching creating noise picked up further down in the analogue chain, particularly with poorly shielded, or unshielded analogue connections.

[NRG]

But Clive, the CAT7 shielding is there to enable higher data rates, 40Gb, you seem to be inferring that the shielding is affecting sound quality by rejecting noise on the data pairs of the cable as if this where an issue at the data rate (100Mbs?) run between the two computers. Its a non issue.

[Clive]

Paul, I can appreciate you skepticism but given that to achieve higher transmission speeds reliably it is thought necessary by the standards bodies to require improved shielding and metal plugs, I feel we can't say digital cables do not pickup noise, whatever the mechanism is.

[Clive]

But Clive, the CAT7 shielding is there to enable higher data rates, 40Gb, you seem to be inferring that the shielding is affecting sound quality by rejecting noise on the data pairs of the cable as if this where an issue at the data rate (100Mbs?) run between the two computers. Its a non issue.
That's not what I'm saying Neal. I'm simply saying the grounding connections may be having some downstream implications on the analogue kit. I'm making no claim about the data packets.

[brucew268]

I'm simply saying the grounding connections may be having some downstream implications on the analogue kit. I'm making no claim about the data packets.

That fits my experience as noted 2 days ago on another post... It may also be about the noise on the voltage/ground wires affecting the overall circuit, not just the effect on digital information passed. I was surprised to find that different ethernet cables made a difference to the sound even though my music files are connected via USB not ethernet! This obviously could not be affecting the digital packets. My only reasonable theory is that the electrical noise was passed along the ethernet cable into the computer, probably onto the ground plane, and from there passed into the DAC. So perhaps some USB cable differences are due to noise passing between the ground plane of the computer and DAC.

[StanleyB]

...thanks Stan. Do you mind my asking which of the various configurations you preferred?
If shielding is an issue here, is it better for the earth connection to be linked at one or both ends?
I have switched to using a dedicated wireless-N connection on 5.4GHz for my audio and video streams around the house. I bought a couple of Zyxel N300 repeaters to do that.
But when I ran a wired connection, I had the luxury then of being able to take leads, patch panels etc., from my old workplace and try them out at home till I was satisfied. I didn't really spend too much time in listening out for the absolute difference. Far too much to do then.

[brucew268]

is wireless a 'purer' option after all?

I found a significant SQ improvement and lowered noise when I turned off my computer's wireless antenna, switching to ethernet connection for internet and JRMC Gizmo.