I started reverse engineering the brain some time ago. Although I have not yet reached the stage where I can replicate it in silicon, I know enough to be able to report to you that what you say here is only a half truth.
The brain is hard wired for face recognition, google fusiform face area. The result of this is that we see faces in many things that aren't, like car fronts. Of course, this piece of hardware has to be operationalized and optimized by learning. But, without the hard wired bits in your brain, you could not learn, and you cannot learn beyond the capabilities of your brain hardware and your senses that feed it. Your brain is far from a freely programmable computer. All subroutines are based on dedicated neural networks, all located in specific areas in the brain. That goes for anything. Lose your Broca area and there goes the active use of your mother tongue. Damage to the fusiform area can lead to the inability to recognize faces. There is some redundancy and plasticity within the brain, for example between the left and right hemispheres, but as brain damage reports show, this is fairly limited.
Much of what is alleged about the abilities to differentiate between DACs appears to go beyond innate human capabilities to either pick up differences or to discriminate between them in further processing. So I can only lend credence to it if it is tested. Sorry to have to come back this point. If you can learn to differentiate by ear only between different DACs, you could certainly also learn to do so in a controlled environment.
I only said "suggests." And I might be wrong, especially given that research also shows that people behave differently, and usually worse, in internet forums than they tend to in real life. And because I always might be wrong about anything. But, still, yes, I think there is some reason to think so. That being said, I do remember that the concept of personality type is only a simplified model of reality.
Agree about facial recognition in general. However, for recognition of individual people, there is a racial effect, and there is learning. If you see a face in a cloud, you probably don't recognize a whole lot of details in it. Of course, I can't anticipate every objection to what I say, and I can't write a book in response to every post. What I did say in this case is true. It is also true that there is much more I left out.
Regarding DACs, if you want proof, that's fine. So do I. I think some people can ABX DACs reliably, at least to some extent. Apparently they do that at Benchmark. They use an ABX test setup because they want to know the truth. And I still want to know if blind people can be trained to hear the differences. They tend to have more acute hearing than sighted people.
Why not test it? Of course the tester must know how to set up a proper or representative test for this unique case. Problem is, nobody know what the proper test should be.
CORRECT.
I didn't say never make it. I was talking about statistical characteristics of people who tend to make good scientists and/or good superforcasters.
A good part of the reason I made the original comment was because I knew about the freshly minted part and thought somebody should say it to him before he goes too much further to get him thinking about it. It was intended as more of a wake up call, and not a put down. Probably, almost certainly, I could have phrased it better. Maybe I didn't wait long enough before replying, as I think now the word magical probably primed me for that particular form of response.
Also, I am posting way too much. I will take a break.
Last edited:
Scatter of all kinds is the base reason I find, david griesinger also.
Our hearing takes note of dynamic envelopes (affects/effects intelligibility strongly), and scatter serves to alter perceived dynamic envelopes in a reverberant environment.
In closed headphone environment these room alterations/exaggerations are of course removed, but on close HP listen the system scatter caused exaggerations (and HP scatter caused/excited exaggerations) are still evident.
System scatter (signal dependent system excess noise of all kinds and spectrums), plus signal embedded noise of all kinds and spectrums can interact to cause subjective masking of lower level sound.
Force reduce, set a slope and set a rigid order to the system noise, and the ear happily PLL cancels the aberrations and the performers are magically/holographically 'Live In Da House"
.Welcome to cranking sound 'Live In Da House'.....Damn, how did you stand those pansy electrostatics for so long !!!.Yes, it is subtle. But a positive step forward IMO. And my compressed sound was still there.
Except, NOW the sound isn't compressed any more with the change to the new speaker/amps.
Dan.
I find your forum inputs to be perfectly polite and perfectly reasoned.
I usually find this to be a nice reflection of the person 'in person'.
Dan.
Talking about noise in general, I took a look at the Benchmark video in YouTube that compared SMPS to standard linear PSU.
What I struggle to reconcile is the amount of mains hash around nowadays. Sure, most of it is above the audio band, but I cannot help think that this must have some effect at in band audio signals. If I clip my scope ground lead to the tip and then touch anything plugged into the same outlet, I get 2-10mV of wideband hash. BW limit it to 20 MHz and there is a noticeable drop in the level. I can trigger on certain noise sources, and in general there is stuff from about 50 kHz all the way through to 30 MHz.
Its on all three of my scopes - TEK, Rogol and B&K Precision - although as reported a month or two back, the TEK seems particularly bothersome.
On the Rigol, if I average 256 times, I get a clean trace on the 2mV/Div scale. . . but that's not reality is it?
I remember designing T/C and RTD conditioners amongst other things 25 years ago and getting absolutely zip on the scope trace (we used TEK scopes almost exclusively, although I think there were a few Gould's and Hameg's floating about in the lab as well). T/C's are similar to MM amps in terms of signal levels. The only time we had a problem was with fluorescent lamps so they were turned off when making sensitive measurements. I stepped out of that activity and onto a different role at about the time everyone was getting ready to comply with the EU EMC directive, but got involved again with it on the automotive side a year or two later in '93-'95.
The other thing, is that most of these SMPSU's have 'X' and 'Y' connected caps to the equipment earth. So even though the loop area on the actual PSU is small, you still have residual garbage that ends up on the line including the earth (ground) - and mains cabling inductances/capacitances don't make things any better.
In the case of the Benchmark, its a pretty tough test since the other end of the cable was plugged into a mic preamp. Personally, I would have thought a flux band (and inter-winding screen) would be in order - I am assuming they were not in use on the unit used in the demo.
What I struggle to reconcile is the amount of mains hash around nowadays. Sure, most of it is above the audio band, but I cannot help think that this must have some effect at in band audio signals. If I clip my scope ground lead to the tip and then touch anything plugged into the same outlet, I get 2-10mV of wideband hash. BW limit it to 20 MHz and there is a noticeable drop in the level. I can trigger on certain noise sources, and in general there is stuff from about 50 kHz all the way through to 30 MHz.
Its on all three of my scopes - TEK, Rogol and B&K Precision - although as reported a month or two back, the TEK seems particularly bothersome.
On the Rigol, if I average 256 times, I get a clean trace on the 2mV/Div scale. . . but that's not reality is it?
I remember designing T/C and RTD conditioners amongst other things 25 years ago and getting absolutely zip on the scope trace (we used TEK scopes almost exclusively, although I think there were a few Gould's and Hameg's floating about in the lab as well). T/C's are similar to MM amps in terms of signal levels. The only time we had a problem was with fluorescent lamps so they were turned off when making sensitive measurements. I stepped out of that activity and onto a different role at about the time everyone was getting ready to comply with the EU EMC directive, but got involved again with it on the automotive side a year or two later in '93-'95.
The other thing, is that most of these SMPSU's have 'X' and 'Y' connected caps to the equipment earth. So even though the loop area on the actual PSU is small, you still have residual garbage that ends up on the line including the earth (ground) - and mains cabling inductances/capacitances don't make things any better.
In the case of the Benchmark, its a pretty tough test since the other end of the cable was plugged into a mic preamp. Personally, I would have thought a flux band (and inter-winding screen) would be in order - I am assuming they were not in use on the unit used in the demo.
Last edited:
Last edited:
Yes Dave found the anisotropy of the air in a typical venue caused jitter-like effects orders of magnitude larger than a good CD player.
How about measurements for the first 250 ms?
My question about this clock settling time is how this can make a real problem if it does indeed settle within this time frame. Once settled is anyone saying that the clock will drift around some central locking a sort of ranging of the clock or once settled is it actually stable and unchanging? From what I understand the master clocks are more than capable of very precise timing function without drift over time and especially over very short periods that could be detected in sound reproduction as an equivalent to wow or flutter, so what gives with this new phenomena being brought into the picture? Once the music is playing what is supposed to be the problem, it seems to be looking for ghosts and goblins to find problems with digital. Why aren't the haters of digital hating on their analog systems for similar and obviously much greater timing errors?
My question about this clock settling time is how this can make a real problem if it does indeed settle within this time frame. Once settled is anyone saying that the clock will drift around some central locking a sort of ranging of the clock or once settled is it actually stable and unchanging? From what I understand the master clocks are more than capable of very precise timing function without drift over time and especially over very short periods that could be detected in sound reproduction as an equivalent to wow or flutter, so what gives with this new phenomena being brought into the picture? Once the music is playing what is supposed to be the problem, it seems to be looking for ghosts and goblins to find problems with digital. Why aren't the haters of digital hating on their analog systems for similar and obviously much greater timing errors?
I hope everyone is sitting VERY still and all heat and/or AC is off and all the windows and doors are sealed tight. Remember we are talking nsec jitter.
Last edited:
I'm okay with that--I'll make due some other way. And I'm another different internet random no less.
I think other posters have covered it, but my recent posts are in the line of, "extraordinary claims necessitate extraordinary measures". And, do your homework first, which to me is to ask and roughly guesstimate the magnitude of the issue at hand. Mark, perhaps I just approach the problem from a different perspective? The issues being discussed are electrically small or in regions of our hearing that are marginal at best, and thus are going to have small effects (if any) on audibility. Which certainly raises the bar quite high as to the rigor of tests needed to tease out audibility (small effects are easily lost in the noise of experimentation).Not even moderate listening tests have been demonstrated but people have moved well past that point. So let me stop for a second: have we tested the audibility of DACs AND have differences been shown that are unique to DACs (i.e. not phenomenon otherwise known)? I.e. are the effects being talked about REAL and, if real, are they already understood, albeit not clearly elucidated? If the latter, then there is still good scholarship in the pursuit of teasing out these relationships and what makes us hear certain things.
If not, then this is all a huge amount of scabbing about without the remotest of a hypothesis. A lot of electrons have been inconvenienced with the conclusion forethought and mechanisms trying to be placed to it. I'm not saying things are impossible, I'm saying that plausibility is low.
If that mindset makes me a bad scientist, so be it. I usually consider that scientific sobriety and discipline. The methods I've known for discovery usually consist of inching just past our present understanding (as to minimize the number of known-unknowns and unknown-unknowns), but what do I know? I'm more of a D guy in R&D anyhow, much preferring to work just off the leading edge and make stuff with that new understanding. Which, uh, tends to require keeping a lot of different, often confounding factors, in flight at once. Not a visionary, which, within the worlds I traverse, are only the few atoms on the tip of the spear--the rest of us are there to support those few individuals and keep the whole sucker stably in flight with enough heft to punch an effective hole in the intended target.
And, lastly, no, not all of us seem to think something is clearly wrong. I use magic liberally and not altogether disparagingly, in the case that Mark seems to take exception, feel free to replace it with "major".
Why? It's a total non-problem. For a quarter of a second after switching a source, there's an anomaly. Answer: after you switch a source, which is not exactly something that occurs continuously, wait a quarter of a second.
- Status
- Not open for further replies.