Unless the effects are repeatable in proper blindtest there is no basis for general claims. It's rather
illusions then science.
If I get a DDOS attack om my home router or switch it is because I have rotten children attacking from within.
Could be, however sometimes the term is used more loosely to refer to disruptively large packet volumes for any reason. I once worked in a building where a PhD physicist who didn't know any better plugged a CAT-6 cable between two network segments. As a result broadcast traffic started going around in circles which pretty much took the whole building network down for a couple of hours until the IT guys found the problem. They described it as packet flooding (and it seemed pretty clear what they meant by that).
With all due respect, "modulate the power supply in a DAC" sounds extremely improbable to me.
It would have to be an incredibly shoddy power supply - so shoddy that the circuit probably wouldn't power up, either. Since a DAC is intrinsically a mixed-mode circuit, I can't imagine what sort of designer wouldn't bother to put a decent amount of power supply filtering on the analogue half of things, not to mention on the digital side as well.
Yes, it's possible for a person to be too skeptical. But at the other extreme, we don't want to "keep our minds so open that our brains fall out", either.
It alarms me that so many people are so quick to accept incredibly absurd propositions, particularly if they come from famous narcissists. Elon Musk's "hyperloop" is snake oil just as much as CD demagnetizers.
Hyperloop - what a brilliant idea. Make a gigantic vacuum chamber, then stick large numbers of vulnerable human beings inside it. With no way to get them out if their "train" stops midway between two exits.
Not to mention several other serious problems, including the fundamental one that the cost of building and pumping out a vacuum system of this size far exceeds the cost of fuel needed to run a conventional bullet train through air at atmospheric pressure.
The idea of numbers of people living on Mars is equally absurd. Really? People are going to live indefinitely on a planet with no soil, no breathable air, no biosphere, no infrastructure? It would be easier to put a colony on the ocean floor - at least help would be only a few miles away. (But that's a stupid idea, too.)
Maybe we should be calling it "Musk oil" rather than "snake oil"?
-Gnobuddy
It would have to be an incredibly shoddy power supply - so shoddy that the circuit probably wouldn't power up, either. Since a DAC is intrinsically a mixed-mode circuit, I can't imagine what sort of designer wouldn't bother to put a decent amount of power supply filtering on the analogue half of things, not to mention on the digital side as well.
Yes, it's possible for a person to be too skeptical. But at the other extreme, we don't want to "keep our minds so open that our brains fall out", either.
It alarms me that so many people are so quick to accept incredibly absurd propositions, particularly if they come from famous narcissists. Elon Musk's "hyperloop" is snake oil just as much as CD demagnetizers.
Hyperloop - what a brilliant idea. Make a gigantic vacuum chamber, then stick large numbers of vulnerable human beings inside it. With no way to get them out if their "train" stops midway between two exits.
Not to mention several other serious problems, including the fundamental one that the cost of building and pumping out a vacuum system of this size far exceeds the cost of fuel needed to run a conventional bullet train through air at atmospheric pressure.
The idea of numbers of people living on Mars is equally absurd. Really? People are going to live indefinitely on a planet with no soil, no breathable air, no biosphere, no infrastructure? It would be easier to put a colony on the ocean floor - at least help would be only a few miles away. (But that's a stupid idea, too.)
Maybe we should be calling it "Musk oil" rather than "snake oil"?
-Gnobuddy
You don't know that Vref/AVCC (including any slight noise on it) is convolved with the audio output? It becomes audio signal-correlated noise which even ESS admits the ear is 'exquisitely sensitive" to.
Oh, I understand the mechanism. But I'll bet that the effect is orders of magnitude below the threshold of perceptibility.
None of our human senses is really all that exquisitely sensitive. Our senses can't differentiate between a 100-gram weight and a 101 gram weight; we can't tell a 100 cm metre-stick from a 101 cm stick without laying them side by side; we can't tell by touch the difference between an object at a temperature of 50 C and one at 50.5 C; we can't taste the difference between 10 gm of sugar dissolved into a litre of water, and 10.1 grams of sugar dissolved into the same volume of water.
All these stimulii have differences are at the 1% level - and we can't detect them with our senses.
And in controlled, double-blind listening tests, I've never heard of a credible result where significant numbers of people could detect less than 1% THD with any statistical significance.
These days, cheap equipment bought from Amazon has far better than 1% resolution. A while ago, I bought a digital scale that can weigh up to 3000 grams, with a resolution of 0.1 gram, for around a hundred bucks IIRC.
That's a resolution of 0.0033% - three hundred times better resolution than 1%. And our human senses can't even detect a 1% change in most cases, so the cheap digital scale is probably at least a thousand times more sensitive than our own ability to sense weight.
Quite a long time ago, it became easy to build audio amplifiers with less than 1% THD, and ruler-flat frequency response over a wider range of frequencies than the human ear could perceive. The audibly perfect amplifier was already here.
But how do you continue to sell new models of amplifiers when all of them are audibly perfect? So the goal-posts were moved to 0.1% THD, a level far below anyone's ability to detect at a statistically significant rate. Good. Now there is no doubt the amplifier is audibly perfect.
But no. The "spec inflation" didn't stop there. Pretty soon we had 0.01% THD, then 0.001% THD. Then we had claims that clock jitter of the order of one part in a million was sufficient to audibly degrade digital audio from a DAC. We have people claiming they can hear the difference between polyester and mica capacitor dielectrics, from a capacitor in the internal tone control circuitry of an electric guitar.
Heck, we have people claiming they can hear the difference between red and blue Teflon insulation on a signal-carrying wire.
Now it's supply voltage ripple on DACs. I have no doubt the background acoustic noise from the nearest 'fridge is orders of magnitude louder than any such effect. I have no doubt the "sticktion" from loudspeaker spider and surround causes more distortion than this.
What's sticktion? At the microscopic level, a graph of displacement vs force will show tiny stair-steps in most materials. Stick, then slip, then stick again. Someone coined the term "sticktion" to describe this.
The human mind is restless, and like a dog uselessly chewing a bone, cannot help but imagine an endless stream of ways in which things might go wrong. That's why empirical measurements and the application of the scientific method are so crucial. Without those reality checks, we not only dream up nonsense, we also start to believe it.
The nonsense we are willing to believe knows no bounds. For instance, billions of people believe they have an invisible, intangible, undetectable super-human lord and master in the sky, who takes a great deal of interest in being praised, and who believes in meting out incredibly cruel eternal punishments.
Compared to that, CD demagnetizers seem sane and sensible.
-Gnobuddy
Interesting claim. Is there some logic used in arriving at that conclusion?
Regarding HD perception, PMA once did a double blind listening test of non-inverting unity gain audio opamp buffers to see if anyone could reliably hear a difference. Before he announced which wav file was of which opamp, he gave me credit for sorting the opamps in order of of measured distortion (except one which I skipped because of feeling too burned out to sort yet another one; each one was hard and took a lot of sustained concentration). Since the test files were of music there was no way for listeners to know which was either more or less distorted as compared to another file except purely by ear. Part of being able to sort them by ear is to learn what to listen for. IME low level HD does not sound like what most people think of as distortion but it still has a sound. In some cases for odd order HD, at very low levels it can sound like increased clarity, but its a false type of clarity. You know all this stuff too?
Last edited:
One of them is the one I quit bothering to sort at the end. Had to bubble sort them one by one, comparing them to each other one at time. Remember this was probably with Benchmark DAC-1, definitely with a noisy 20-year old Bryston 4-B power amp, and some dreaded Yamaha NS-10 speakers. Not ideal at all. It was a difficult process. Figured someone else would nail it easier than I could, so why torture myself. But nobody did nail it. In retrospect guess I should have kept going. Anyway I got some credit for what I did do even if it wasn't a completely perfect score.
Point is they did sound different, even if only very slightly so. Think it illustrates its possible to do better than 1% distortion perception if you know what to listen for and are willing to do the work.
Point is they did sound different, even if only very slightly so. Think it illustrates its possible to do better than 1% distortion perception if you know what to listen for and are willing to do the work.
He bridged two LANs. And yes this can be a real problem when the untrained are let loose with complex networks. This cannot happen in the home in most scenarios. If it could then a magic low noise switch will not make a blind bit of difference. If it did then music servers would have a seperate NIC to go to the player or at the very least offer VLAN support. A quick check shows that, even when you spend $15,000 on your music server it only has one ethernet port. Funny that.
And yet I don't see this mentioned in any reviews.
You did, but that once again has NOTHING to do with the overpriced switch we are laughing at.
ESS has been pretty circumspect about what they are willing to say. Looks like they took down some Martin Mallinson youtube lectures. Last time I checked, their comments about what audiophiles hear from dacs are still in the powerpoint at:
https://www.yumpu.com/en/document/read/23182504/noise-shaping-sigma-delta-dacs-ess-technology-inc
The part about what audiophiles hear starts at page 28 of the slides. Somewhere in there are the signal correlated noise comments. Regarding convolution of Vref/AVCC with the audio output its well known among dac guys. MarcelvdG has talked about it and bohrok has measured some of it.
I was responding to a post that more or less seemed to be saying humans can't hear much at all. I think there is good reason to disagree.
Hey there, I take my snake oil very seriously and allow myself one snake oil component per tube amp build. For my last build, I created the super amp of dreams by purchasing that Sozo 1000v 500uF silver mica cap from Mojotone. $10 bucks, when I could have spent 70 cents on a perfectly good EQ treble cap. But, we all know that now, the high frequencies in that amp will have a magic aura to them that is so sublime, and so magically subtle that only the few and gifted will be able to hear it. I hope I'm around to meet that person because I guess it wasn't me. But, its big, its bright and yellow, and it stands tall and proud over the other run of the mill Mallories and Spragues, and I hope it makes all those unremarkable components look to find their higher audio spirit. Hey. Components are people too --- but not really.
Of course there's logic. It's the opposite position that's illogical - i's illogical to assume that an audible effect will exist, which is the position you're taking.
Please not that you only quoted an excerpt from my sentence, and a misleading one. I wrote "I'll bet that the effect is orders of magnitude below", while you quoted only "...the effect is orders of magnitude below". By leaving out the first three words of my sentence, the excerpt you quoted makes it sound as though I was making a definite statement of fact.
Which I clearly wasn't. "I'll bet" means exactly that: my educated GUESS is that the effect you're discussing is far, far below audibility.
I won't waste my time attempting to prove this. I leave it to the people who make extraordinary claims, to provide the plausible proof for them. You believe ethernet packet data can audibly affect the sound from a DAC? Prove it.
Mark, I know lots of stuff, and so do you.
You believe lots of stuff you don't actually know. And so do I.
(For instance, I don't know that the "multiverse" concept is wrong, in the sense that I can't prove that. But I believe it's utter nonsense, and I believe whose who think otherwise have lost their grip on reality.)
No need for a "measuring contest" as to how much you know, or how much I know. It won't settle anything.
The issue here is fairly simple: your personal bias is to err on the side of charity when it comes to subjective audio-related claims. If someone says they can hear the sound of two electrons rubbing together inside a PN junction, you might be willing to consider that claim.
My personal bias is to firmly reject claims that contradict data collected by decades of carefully conducted tests. I don't trust subjective claims unsupported by objective data, most particularly when those claims deviate a long way from facts painstakingly gathered by decades of carefully conducted science.
I trust the scientific method, because it's proven to be the only way in which fallible human beings can arrive at correct conclusions. Cultures that haven't embraced the scientific method continue to live in the stone age.
Wanna prove that floods of Ethernet packet data can translate to audible distortion? Well, the ball is in your court. Set up a large-scale double-blind listening test, and show that the majority of listeners can detect the claimed distortion at least 70% of the time.
Without data to substantiate it, the claim is worthless. It's like claiming the earth really is flat, because, look, I can walk for an hour in Florida (America's flattest state), and never climb a hill or descend into a valley.
-Gnobuddy
It's a broadcast storm. Any decent Layer 3 switch will detect this and shut down the offending port. A Layer 2 switch like Nordost is defenseless against this.
And, to be fair, this happens a LOT, when attempting to differentiate phenomena that are right at the threshold of detectability. In fact, it happens 50% of the time!
This is exactly why "I got it right once!" is NOT significant. "I got it right more than 70% of the time during multiple trials" may be significant, if all the other ducks are also in a row (properly conducted double-blind test, sufficient number of trials, other sources of error excluded, et cetera.)
I once threw a basketball awkwardly in the general direction of the hoop, and scored a 3-pointer from a considerable distance. That doesn't mean I have the ability to score 3-pointers from that distance - I don't! It only means that even a low-probability event can happen from time to time.
-Gnobuddy
No. No. No. A bit will be -2.5V - +2.5 volts, or vice versa. For 24 bits, it's that x 24. It not 1 bit divided into 24 levels.
A decent Layer3 device ( aka router) will not detect this, it will never forward broadcasts!
Level2 devices may have additional sw that disable a port that is suspect ( any layer device might have this function too)
The broadcast storm mentioned here ( someone interconnected 2 branches of a network) would not have
happened if the devices used had enabled spanning-tree protocol or any of it's cousins.
Of course, anything that works on layer 3 can be considered a router. Then again, if a switch breaks down at work and I send one of our technicians out to "replace the router" I'm just asking for trouble 😉
All I meant to say is that most switches with layer 3 capabilities also come with broadcast storm detection. I did not mean to say that layer 3 capabilities are required for this feature, that is not the case.
Setting up some variant of Spanning Tree is indeed good practice.
- Status
- Not open for further replies.