Popular Post Blackmorec Posted December 5, 2019 Popular Post Share Posted December 5, 2019 Hmmm, unfortunately I’m not particularly technical when it comes to electronics and noise. I just rely on my ears and plain old logic. Now one thing I do know is that if I want to compare 2 cars I don’t do it by parking them next to each other, starting them up, allow them to idle, stick a sound level meter up each exhaust and when I get the same 83dB from each, conclude that they’re the same. But that seems to be what’s going on here, or am I mistaken? In terms of testing noise and jitter I don’t see how you can do that without a signal. What are you actually measuring? Quiescent noise? The only conclusion I personally reach from this evaluation is that there’s not a lot of noise present. Am i wrong? I would have thought that to test the efficacy of an Ethernet switch in its ability to remove noise you’d have to test them in the presence of noise, preferably a known quantity, then look at how well one switch removes it vs the other. Again, am i wrong? And then there’s the question of noise. If there is no noise to start with, then why do you need an Etherregen? But where does the noise and jitter come from? Again in my very limited knowledge I understood its generated in the first place by demanding loads on the power supplies, which in turn interact with the systems’ oscillators to produce jitter/phase noise. So again, a signal needs to be present for the system to generate the noise for the Etherregen to remove. No signal, no noise, no difference, no sense. The last piece of my homely logic relates to people and statistics. There‘s a saying that goes; “ you can fool all the people some of the time, and some of the people all the time, but you can’t fool all the people all the time” Yet that appears to be exactly what Uptone Audio has done with the Etherregen if I’m mistaken and these measurements actually do have merit. That’s quite an achievement....A piece of electronics that allows almost 100% of the people who hear it to IMAGINE the exact same thing. Wow, that really takes some doing....far more impressive than just building a better network switch 🤯. John Swenson, the mind-control engineer Finally, I’ve already freely admitted that I’m not an expert in these matters and the above conclusions are reached by simply applying logic. If in fact I am wrong, i would appreciate being corrected with a good technical explanation as opposed to the utterly useless sarcasm that seems so prevelant in these parts. At least then some of us learn something. Thanks mansr, thyname, Teresa and 5 others 7 1 Link to comment
Popular Post Blackmorec Posted December 5, 2019 Popular Post Share Posted December 5, 2019 I’ve just looked at the full review, and frankly I’m no less confused. Essentially the EtherRegen was compared to a TPLink switch and the comparison revealed NO, I repeat NO difference. Now one of the things I’ve learned through almost 40 years of scientific testing, is that when a comparison of 2 different things show up NO differences, you might not be testing what you think you’re testing. So I looked a little closer at the testing protocol.....and again I’m no expert in digital electronics.....I only look at the logic of the testing. The results showed very low level noise and were indeed identical, so identical that they look for all the world like they were measuring the SAME THING. Looking a little further at the results, there was also NO difference between the ‘A’ and ‘B’ ports of the Etherregen, so now we’re testing 3 (!) inputs and all are giving the exact same results. That’s beyond unusual in terms of any sensitive measurements, its practically unheard of. So I took another look at the test protocol and what we have is 3 inputs to a DAC (Etherregen A, B and TP Link) with the measurement at the DACs output. Logically then, we’re not measuring the TP Link, Etherregen A and B, what we’re actually measuring is TP Link + DAC, Etherregen A + DAC and Etherregen B + DAC. So what do all those measurements have in common? The DAC. And what would happen if the noise we’re measuring all came from the DAC...remember the DAC is an active device. All the measurements would be the same, identical. Which they are. So, either I’m missing something important due to my ignorance and I’m happy to get technical input in that regard, or the test is flawed. Generally speaking a good scientific test has a ‘control’ or a ‘blank’ element. In the case of this test that would be an ethernet feed with no switch. At least then you’d at least know if both switches were removing noise or not. In this case, you’ve no idea what noise is coming into the DAC vs what’s being output and you’ve no idea what the switches are doing vs what the DAC is adding BigAlMc, RickyV, LowMidHigh and 6 others 5 1 3 Link to comment
Popular Post Blackmorec Posted December 5, 2019 Popular Post Share Posted December 5, 2019 1 hour ago, firedog said: Don't agree at all. The ER and all similar devices make the claim that they improve SQ by reducing "noise" going into the DAC which then improves the audible output at the DAC. If the output at the DAC isn't improved, it's irrelevant what the device "does", as it isn't making any difference to SQ, at least not in any audible way ("improvements at -130db don't count). Amir's test is correct in concept - testing how the ER effects the output of the DAC, which is what it claims to do. If you test the ER on it's own and it "reduces noise", that doesn't prove what we need to know. What if the existence or non existence of that "noise" has no effect on the functioning of the DAC, in spite of what Uptone and others claim? What if, as many claim, all well designed modern DACs are basically immune to the effects of the noise Uptone claims the ER reduces? Note Amir also tested a second DAC which he knows to have issues and that even he says was helped a bit by the USB Regen; this DAC also showed no difference in output with the ER attached in the chain. If what you are saying is true, how come manufacturers of such devices can't find DACs that show improvement when these devices are in line? If the devices do what they claim, it should be easy to find multiple DACs that have an output improved by the device. How does the ER affect the output of the DAC if its the DAC that’s creating the noise? The ER is upstream of the DAC so can only improve what comes into the DAC. If what comes into the DAC is nothing and the DAC then adds this low level noise, how does Amir’s test detect that? As far as I can see, it doesn’t. For Amir’s test, the ER and the TPLink could have been removing loads of noise, but the input to the switches was never checked. Based on Amir’s tests, as far as i can see, he actually has no idea where the noise he’s measuring is actually coming from. Again I’m no expert in these matters but if I were looking to see if a device removed noise, i would inject a known quantity of noise into the device and look what comes out. I would also check to see what the noise contribution is of other components downstream of the device I want to measure. Without knowing much about the electronics, there’s still a logic to testing in order to rule out confounding variables. In the case of Amir’s test, if it was the DAC that was introducing noise, then the noise from the DAC’s output would be identical regardless of what was connected upstream and that’s exactly what the results show. Uncannily so. Urs, RickyV, jaynyc and 2 others 3 2 Link to comment
Popular Post Blackmorec Posted December 5, 2019 Popular Post Share Posted December 5, 2019 Let me simplify the points I was making. Regarding Amir’s tests, he has no idea what’s going into the switches, no idea what’s coming out and no idea what the DAC is adding. All he knows is that what’s coming out of the DAC, where ever it came from, is the same, regardless of what’s connected upstream. If Amir tried to publish these results in a peer reviewed journal he wouldn’t make it past the initial editorial review. The paper would simply be rejected for lack of experimental rigour. sandyk and Teresa 2 Link to comment
Blackmorec Posted December 5, 2019 Share Posted December 5, 2019 2 minutes ago, bobflood said: No one will ever win the "Sounds Good vs Measures Well" debate. That's very true. But I don’t think that’s quite the debate here. I’m more inclined to suggest, “Sounds Good vs. Measured Badly” 😉 Teresa 1 Link to comment
Popular Post Blackmorec Posted December 5, 2019 Popular Post Share Posted December 5, 2019 4 minutes ago, thumb5 said: What I understood from Amir's description of the test is that he used Roon on a PC to generate a tone which is streamed to a DAC via the switch under test. Bearing that in mind, I am still not clear what you mean by "what's going into the switches, ... what's coming out" - what specifically do you think is not known that is important to the end result? Putting my point more simply: isn't the etherREGEN supposed to make a noticeable (audible) difference, compared to a generic Ethernet switch, in a typical audio streaming configuration? And isn't that what the test method tests for? I'm not claiming that the test has scientific rigour, but it is surely better than nothing. Sometimes ‘nothing’ is far better than ‘something’ when that ‘something’ is completed flawed. Firstly, the results published were based on no signal and a test tone. Neither of these 2 conditions are likely to actually create any noise in the signal chain. So a relatively clean signal enters both the TPLink and EtherRegen, challenging neither. Both output the stream to the DAC. The DAC processes the stream, adds its own low level noise and outputs its signal. Amir then measures the DACs noise on both streams and concludes that EtherRegen is no better than the TPLink, whereas the real problem lies with the test protocol and the fact that it actually isn’t testing the switches, rather its measuring the DAC. Frankly the chances of a `TPLInk and an EtherRegen measuring exactly the same must be vanishingly small, given the vastly different technologies, whereas if you were measuring the DAC’s noise, the results would be identical. Go figure. Teresa and Ralf11 1 1 Link to comment
Recommended Posts