That is kinda funny..That's a good question. Maybe they were not designed for audio environments, just for data.
Ok, so is there any consensus yet on the best sounding ethernet switches?
Ok, so is there any consensus yet on the best sounding ethernet switches?
Well its a start, measuring port to port and port to power leakage with injected synthesised noise.
But first they've got to find that noise level out in the wild, and half a volt of wide band rfi ain't happening anywhere. He's a good 10-100x out of whack there.
If they go on to use that injected noise level in the listening tests I'd suggest it makes a nonesense of the whole thing. First they have to prove it's a suitable environmental noise replacement, then do the listening.
Great to see someone actually looking at the numbers. But all that counts is does it change the output of the dac, pre, power amps? That noise exists doesn't matter, the effect is all.
Regardless of the device, switch, PSU, amp, etc anything injecting noise that reaches the DAC output and is audible is bad, the key point is whether it is audible and at the moment the measurable noise and IMPs look down in the noise.I disagree. They should first prove the possibility of noise reaching DAC outputs at an obvious level and then find the realistic limits. It would remove any room for doubt.
Regardless of the device, switch, PSU, amp, etc anything injecting noise that reaches the DAC output and is audible is bad, the key point is whether it is audible and at the moment the measurable noise and IMPs look down in the noise.
Injecting non-real world levels of noise into a device, measuring the impact indicates that there is a link, but once the injected noise are at real world levels then the link is still there just inaudible, being so low.
EMC radiated immunity testing for consumer equipment is at 3V/m up to 1 GHz and at some spot frequencies due to WiFi up to 6 GHz. These levels are achievable with a cellphone or access point nearbyBut first they've got to find that noise level out in the wild, and half a volt of wide band rfi ain't happening anywhere. He's a good 10-100x out of whack there.
But I think it was established, via measurement, that real world noise was at a level to be inaudible, regardless of where it came from. And not just established for DACs. Any noise if at a level to be audible is audible, the root here is that the measured noise at a switch is at a level to be inaudible. And the measurements say no, not audible, unless the human has evolved a new, super-duper earI understand, but one of the biggest hypotheses (and doubts!) expressed in this thread is that noise from the network reaches the DAC (and that a switch can attenuate this). So step 1 is a convincing demonstration that the noise does/does not reach the DAC. Then establish that the switch attenuates this. Then check for linearity if the effect (ie does noise only get through if it's if a sufficient amplitude to "swamp" things or do low levels also great through)? Only then start looking into audibility. Skip all those first steps and we'd still have circular arguments about whether noise is actually reaching the DAC. There's a reason scientific papers have multiple tests/results presented. One needs to build a solid foundation to support one's argument.
I surprised that they use exactly the same methods to protect against noise that most other electrical data networking solutions use. It is after all a hybrid system with optical and electrical connections, hardly seems odd they should use tried and tested methods to protect the electrical side of things.
In fact the only noteworthy bit is they got a patent.
But still just handwaving and no measures of electrical interference from networks affecting dacs.
What level is deemed to be inaudible?But I think it was established, via measurement, that real world noise was at a level to be inaudible, regardless of where it came from. And not just established for DACs. Any noise if at a level to be audible is audible, the root here is that the measured noise at a switch is at a level to be inaudible. And the measurements say no, not audible, unless the human has evolved a new, super-duper ear