Most Important, Unloved Cable...

When playing back audio and you pull the Ethernet cable, and of course it will still play back (with most systems) for a few seconds, DOES THE SOUND IMPROVE? It’s really a simple question and yet for some reason...

I don’t have an ethernet connection in my audio system, so I can’t answer that based on experience. However, assuming (as I do) that the several highly experienced and widely respected audiophiles who have reported realizing significant sonic benefits by changing from one inexpensive ethernet cable to a different inexpensive ethernet cable are correct, and if the explanation of those benefits that I hypothesized in my post in this thread dated 3-27-2017 is correct, the sound may or may not improve depending on the specific system.

As you will realize in reading that post, and if my hypothesis is correct, whether or not the sound improves would depend on the path(s) by which, and the degree to which, electrical noise and/or RFI from the signals in the cable reach and affect downstream circuit points that are ostensibly unrelated to the ethernet interface. It would also depend on how the content of the signal sent into the cable by the source component changes when the cable is disconnected, as a result of that component having nothing to talk to at the other end.

While what I stated in my 3-27-2017 post is only a hypothesis, as an EE experienced in the design of high speed digital circuits that operate in close proximity to sensitive analog circuits and D/A converter circuits it is the only means I can envision that would account for the reported differences. And I don’t think that most of those who have similar circuit design experience and are also reasonably open minded would rule out the possibility I have stated.

Regards,
-- Al

I don’t have an ethernet connection in my audio system, so I can’t answer that based on experience. However, assuming (as I do) that the several highly experienced and widely respected audiophiles who have reported realizing significant sonic benefits by changing from one inexpensive ethernet cable to a different inexpensive ethernet cable are correct, and if the explanation of those benefits that I hypothesized in my post in this thread dated 3-27-2017 is correct, the sound may or may not improve depending on the specific system. 

As you will realize in reading that post, and **if** my hypothesis is correct, whether or not the sound improves would depend on the path(s) by which, and the degree to which, the signals in the cable reach and affect downstream circuit points that are ostensibly unrelated to the ethernet interface. It would also depend on how the content of the signal sent into the cable by the source component changes when the cable is disconnected, as a result of that component having nothing to talk to at the other end.

So flipping this on it's head: It's not the cable, it's the device. 

I'm able to show via FFT that an $18 Intel NIC allows for 3 separate DACs to have un-altered output regardless of if I'm maxing out the driving voltage with a 315 foot generic cable or a 12 foot boutique cable.

Ethernet spec has 328 ft / 100m single segment length. We are talking lengths that are, I would guess, that are typically 10 foot or less in in 95% of the consumer  installations. 

So if really expensive equipment is actually susceptible to this then the EE's that are designing it don't know what they are actually doing. They may want to re-read Ott. 

It would also depend on how the content of the signal sent into the cable by the source component changes when the cable is disconnected, as a result of that component having nothing to talk to at the other end.

Why would it matter what the source is sending if you have disconnected it at the client? These are buffered (FIFO) systems and the audio is still playing.

The only change in 'content' is the 100% lack thereof. 

Once again: What sonically changed about the audio when the cable was disconnected. What does the cable, what does mixed signal systems wrt to analog and digital ground plane mixing, have to do with the DAC playing out of buffer (that isn't going to get refilled unless the cable is plugged back in)?

A good litmus test for these multi-thousand dollar pieces of audio gear is to evaluate it blinded and start playback and see if you can hear when a friend, without cueing you in on it, has disconnected the cable. If I / you hear a clear difference I would take a pass on it personally. 

So if really expensive equipment is actually susceptible to this then the EE’s that are designing it don’t know what they are actually doing. They may want to re-read Ott.

It wouldn’t surprise me if many audio designers have not read Ott. Or studied the aptly named book "High-Speed Digital Design: A Handbook of Black Magic." (My background, btw, is in defense electronics, and I have taken Mr. Johnson’s related course).

Why would it matter what the source is sending if you have disconnected it at the client? These are buffered (FIFO) systems and the audio is still playing.

Because as I indicated in my 3-27-2017 post possible pathways by which RFI may find its way from the cable to circuit points that are downstream of the ethernet interface include radiation into power wiring, or into other cables, or directly into various circuit points within the DAC or other components. And the degree to which that may occur may be affected by the bandwidth, shielding, and other characteristics of the cable being used. Also, perhaps a difference would occur because noise generated by the ethernet interface circuitry at the receiving end (i.e., in the DAC) may change as a result of having nothing connected to it.

As far as I can recall, everything that has been said in this thread by those who deny that the reported sonic differences are real has focused on the robustness and accuracy of ethernet communications. While ignoring or discounting what I would consider, based on my experience, to be the very real possibility of interactions between signals and circuits that are ostensibly unrelated. Designs should not be assumed to behave in a manner that is theoretically ideal, IMO, and signals should not be assumed to only follow their intended pathways.

Regards,
-- Al

Because as I indicated in my 3-27-2017 post possible pathways by which RFI may find its way from the cable to circuit points that are downstream of the ethernet interface include radiation into power wiring, or into other cables, or directly into various circuit points within the DAC or other components. And the degree to which that may occur may be affected by the bandwidth, shielding, and other characteristics of the cable being used. And perhaps also because noise generated by the ethernet interface circuitry at the receiving end may change as a result of having nothing connected to it.  

Then I would suggest as T.I. does in "Reducing Radiated Emissions in 10/100 Ethernet LAN Applications" one uses a balanced power supply for the benefits of CMNR. 

Bottom line is that selecting well engineered componentry is what is needed for fidelity. 

I'm already able to produce ADC'd tracks, that given the natural losses of generational copy, are extremely close to the source PCM when compared. So that means, by fact, my $250 is indeed very transparent and high resolution, and that an $18 NIC, a $69 mainboard, $24 stick of 4GB DDR3, $55 240GB SSD, $25 PICO PSU, $60 LPS is impervious to changes in cable vs some that claim their $8000 device is capable of 'resolving' when it's really a failure for it to protect the output from something as simple as a change in one short run of cable vs another.