One related experiment is appropriate here: play a 5kHz sine on your system and ask listeners to estimate the frequency. Almost to no exception they guess too high. In one test I did most participants (audiophiles) estimates were between 8kHz and 15 (!) kHz.
Jan
Sony is currently producing affordable speakers that claim 40kHz BW and headphones claiming up to 100kHz.
For the deep pocketed and status seeking "audiophile" they also have expensive models up to 120kHz.
P.S. as for the amps failing with SACD at that show. Not all of them failed. Some worked just fine, the only thing that could be noticed was an increase in power consumption and a slightly higher chassis temperature. Some designers are honest.
You are overlooking the fact that those were the modal frequencies. The impact is not covered there. And in the end it's about accepting that Fourier Heisenberg Gabor applies to everything or not. It was nice too see, for a change, that it was the engineers showing incredulity this round. After so many "Fourier denier" trophies they awarded in the past, it now looks like poetic justice.
Last edited:
Okay, but not necessarily confined to integrated circuit opamps? There are high power opamps, such as those offered by Apex. Many audio power amplifiers are much like opamps, with excess open loop gain, and feedback. I would agree that too much digression into musical instruments may not be necessary. But we were talking about signal sources including HD 24/192 audio, which could carry ultrasonic information from musical instruments. In the event of the presence of some natural ultrasonic information from musical instruments, the question arose as to how op-amp like power amplifiers might respond. Maybe integrated opamps too, although we are assured there should in theory be no problems in well designed circuits.
Am I? Please explain how a comment on the possible values for n, the integer which counts the modal frequencies, can be construed as "overlooking" the modal frequencies.Kirchhoff said:
Is it? I thought it was about where simple models of tuning forks break down, so we can't naively assume that a tuning form has a meaningful spectrum up to MHz.
Where? Who are "the engineers"?
Actually, Fourier denier trophies are mainly awarded by scientists to people who exhibit Fourier denial (in some cases these are badly-taught engineers).
You're correct. Opamps and how they actually work (through feedback). Can we go a few pages back now, where it was "demonstrated" through an analogy that "feedback works". "Bycicles work through feedback", "We can ride bycicles" thus "feedback works". Can someone please explain the role of the helmet in this?
Please. There is some math that is commonly used to describe control systems (LaPlace Transforms) and opamp feedback is commonly treated as a subset of control in school. Riding a bicycle is also a control problem, as is the cruise control in a car, or a flying ball speed regulator on an old steam engine. The math describing the dynamics of the all such systems is essentially the same, so long as the systems are linear, or close enough to linear for practical purposes. Belaboring this stuff is probably not time well spent. If there are questions, or some desire to better understand in what way these things could be the same, or use the same math, then maybe that would be of some use going forward.
DF, please let's not discuss this "Argument Clinic" style. The gist of it is that:
- people self identifying as EEs came very hard against people puzzled by Fourier and how it is applicable and sufficient in signal / amp analysis
- when placed in front of the same Fourier theory via the Gabor limit, the EEs have shown incredulity too
The fork was just a small detour of the discussion as we went through musical instruments spectrum. And the question wasn't the accuracy of the article about forks. The question was whether Fourier / Gabor apply to musical instruments (fork included).
We can ignore that mechanical device and keep it strictly within our circuit - the theory indicates that we will have signals of infinite BW at both input and output of the amp, since we are switching a signal on and off instead of playing a stationary one. In the electrical circuits realm, is this conclusion acceptable or not?
- people self identifying as EEs came very hard against people puzzled by Fourier and how it is applicable and sufficient in signal / amp analysis
- when placed in front of the same Fourier theory via the Gabor limit, the EEs have shown incredulity too
The fork was just a small detour of the discussion as we went through musical instruments spectrum. And the question wasn't the accuracy of the article about forks. The question was whether Fourier / Gabor apply to musical instruments (fork included).
We can ignore that mechanical device and keep it strictly within our circuit - the theory indicates that we will have signals of infinite BW at both input and output of the amp, since we are switching a signal on and off instead of playing a stationary one. In the electrical circuits realm, is this conclusion acceptable or not?
Last edited:
Mark - agreed with the mention that I believe that the components are sufficiently non-liniar as to warrant non-linear analysis techniques.
That's how I earn my "Fourier denier" trophies. From Lyapunov deniers 🙂
That's how I earn my "Fourier denier" trophies. From Lyapunov deniers 🙂
Last edited:
A large part of how engineers operate is they start with some science, usually physics, and simplify it enough to make it tractable by the human brain, and thus practical for humans do deal with. When working on circuits with multiple loops and nodes, and when limited to hand calculations on paper, only so much is practically possible.
With SPICE modeling, especially transient modeling, it has become possible to handle more complicated problems with much more accuracy. Many nonlinearities can be handled with pretty accurate results, so long as the models are carefully prepared.
When we talk on forums, since we can't do simulations for each other very readily, we tend to revert to manageable, idealized theory as much as possible. Occasionally, we may over-simplify, and we may need to switch gears a little and take a different approach to deal with some problems.
It would be nice if we could talk in terms of where and how we may be over-simplifying, and perhaps discus the tradeoffs or merits of taking a different approach for some problem.
However, it would probably be unwise and impractical to treat every issue at a maximum level of depth. Often, it may not be particularly useful or of practical help. In other cases in may be crucial, but we need to find some friendly way managing disagreements about when it is necessary to dig deeper. Too often, statements of what the problem is are not well stated or clearly posed. Perhaps out of frustration, sometimes the response from others is name calling instead of conversational education. It tends to turn into adversarial debate instead of a cooperative search for agreement and mutual understanding. Unfortunately, the human mind tends to quickly switch into win-at-any-cost-mode. (for those interested, I like to point to pages 75-76 of the book, The Righteous Mind, for a well formulated statement of how people think in such situations. Too much for me to explain here.)
Last edited:
I posted links on another thread last week showing how difficult it is when air has sufficient non-linearity. Several classes of wind instruments, for instance, internally produce shock waves. There is a body of literature on this but in the end very high frequencies propagate with a lot of loss and there is little left in a normal listening situation.
As Mark says you rapidly get to models and maths that only a few (mostly Russians) can deal with.
No real switch can carry out its task instantaneously, so even before the wiring has added a low pass filter we do not have infinite bandwidth. Theoretical games can be useful (in the right hands) for exposing nonsense, but can be misused to propagate nonsense too. Hence theory does not indicate that "we will have signals of infinite BW at both input and output of the amp" - that would be a naive misapplication of theory.Kirchoff said:
You seem to keep tossing in hand grenades into the discussion, then get narked when someone picks one up and defuses it.
Agreed, nonlinear analysis can be quite hard to deal with, and doing so is usually avoided unless it is really necessary. For another example involving air, modeling of turbulence and air-flow over airplane wings using sophisticated computer codes has produced increased fuel efficiency and longer range for the most modern airplanes, but it has been a difficult undertaking.
Not if you want to talk about real physical instruments and music. Take the example of hitting a metal rod with another one. You have a mechanical problem of applying stress over a small area for a finite time and initiating mechanical waves that couple to the air. I see no reason for a switch analogy (Gabor or whatever) to be invoked. Certainly the rod is not moving at relativistic speeds and the problem with geometries, boundary conditions, and mechanical properties like Young's modulus and density specified is probably solvable. At finite speeds the force is not infinite and the deformation will happen at some finite rate. Any limit on how fast you can deform the metal will limit the frequency of acoustical energy coupled to the air.
One of the things which was often said to us physics students is that if you want to run a 'gedanken' experiment then you have to ensure that you don't cheat; the experiment must follow the laws of physics (even if in somewhat idealized form). Mathematicians can have inextensible strings and elephants whose mass can be neglected, but physicists cannot.
Bang two pieces of metal together and you don't get a step change in air pressure.
Bang two pieces of metal together and you don't get a step change in air pressure.
Agreed, but their complexity should not detract from the fact that there are aspects of musical playback (and amplifier internals, including opamps) that cannot be elucidated exclusively through linear approximation and analysis.
So, as a plea to the group, can we please have a break with those "denier" awards? And allow discussions to lead into non-linear (where warranted) instead of having them preemptively labeled as "not applicable", "nonsense" etc?
Okay. Good, things are starting to look better already. Maybe I should mention that this point that any attempts to assign blame for whatever may have transpired in the past will probably not be useful. I would suggest starting out from where we are now. However, if anyone feels like taking some blame, self-acknowledgement of anything that might have been handled differently in the past would be okay. Let's just not reignite any remaining hot spots from before.
- Status
- Not open for further replies.