john k... said:
Hi John
The Falstad apps are quite useful, I use them in my physics class all the time. The "ripple tank" applet actually solves the wave equation in 2D, its not a water wave at all (which is different from a sound wave). He just uses that name because it sounds "catchy" - the sims are quite accurate for what they are. But you are quite corect that 2D can sometimes be misleading. For example a 2D cross section of a 3D axisymmetric waveguide WILL NOT be accurate. There is no way to do an accurate 3D problem in 2D, its always an approximation.
But heck, who cares about "reality" its just too much fun playing with the software.
Re: what's the difference??
What would you say about the shape of a wave front in one plain - isn't it *exactly* "a two dimensional at best"?
So I don't see why to complain.
You are mixing things and again are simply talking down any attempt to bring in some light to diffraction impacts other than your beloved math.
But as you have already joined the topic may I ask you the same question:
- Where do you see the main differencies - I mean as a visualisation tool as I used it for acoustic wave front propagation / wave front bending / interference pattern?
Michael
auplater said:
For one thing it's more or less two dimensional at best...
John L.
What would you say about the shape of a wave front in one plain - isn't it *exactly* "a two dimensional at best"?
So I don't see why to complain.
gedlee said:
You are mixing things and again are simply talking down any attempt to bring in some light to diffraction impacts other than your beloved math.
But as you have already joined the topic may I ask you the same question:
- Where do you see the main differencies - I mean as a visualisation tool as I used it for acoustic wave front propagation / wave front bending / interference pattern?
Michael
Actually, as long as it is recognized that the solutions are 2-D I would wirthdraw my objection. It can be shown that the shallow water equations yield the same wave linear equation as those for a compressible gas. The only questions would be what the details of the simulations are; dimensions, wave speed, wave number, etc.
[edit] But I would also agree with Earl that there is danger in assuming what you do in 2-d has any connections to what happens in 3-d. As a demonstation that edges and boundaries mess things up they are find, but the degree to which diffraction and reflection enter the pictide is another thing. These appletes are basically educational toys. If they help gain insight, fine.
[edit] But I would also agree with Earl that there is danger in assuming what you do in 2-d has any connections to what happens in 3-d. As a demonstation that edges and boundaries mess things up they are find, but the degree to which diffraction and reflection enter the pictide is another thing. These appletes are basically educational toys. If they help gain insight, fine.
john k... said:
Thanks John for your clarification.
Yes, its first hand do gain insight / understanding and second about the "degree to which diffraction and reflection enter the pictide" in terms of comparable numbers
Until now I haven't found a tool that does this "degree to which diffraction and reflection enter the pictide" correct - CARA is mere ray tracing (though with smashing results when working around with some tricks) and BEM doesnt show what's going on either.
On the other hand there is a lot of confusion going on about the very mechanisms involved and how they translate.
Earl's statements are of little help as he is confused the most when it comes to this very topic - no "real" understanding IMO - just giant general knowledge and some impressive math juggling for the sake of a "rigor single spec optimisation".
Hence I do not understand what he wants to tell me and he does not understand what I'm after.
Michael
Re: Re: what's the difference??
...yes, it may be a 2D representation of wavefront behavor, but (by necessity) it will ignore any confounding from the Z axis solutions, hence it does not represent 3D space, and as John K says the danger is in misinterpreting this as reality for sound waves
I don't see Earl talking down at all... just stating reality...
as a visualization tool for diffraction of 2D waves ok... but we don't listen to 2D waves, sound waves are 3D, and since you seem to want to predict audio performance using sims, I don't see the application as representing anything accurate in doing so...
John L.
mige0 said:
What would you say about the shape of a wave front in one plain - isn't it *exactly* "a two dimensional at best"?
So I don't see why to complain.
...yes, it may be a 2D representation of wavefront behavor, but (by necessity) it will ignore any confounding from the Z axis solutions, hence it does not represent 3D space, and as John K says the danger is in misinterpreting this as reality for sound waves
mige0 said:
I don't see Earl talking down at all... just stating reality...
as a visualization tool for diffraction of 2D waves ok... but we don't listen to 2D waves, sound waves are 3D, and since you seem to want to predict audio performance using sims, I don't see the application as representing anything accurate in doing so...
John L.
Re: Re: Re: Re: what's the difference??
yes... equations have limitations as well... but until you demonstrate some understanding of the mathematical approach to modelling sound transmission, rather than dismiss it, your approach has less credibility...
John L.
mige0 said:
there is no "reality" in *any* allegory or equation - lets face it...
Michael
yes... equations have limitations as well... but until you demonstrate some understanding of the mathematical approach to modelling sound transmission, rather than dismiss it, your approach has less credibility...
John L.
1 +1 = 2
need more?
Its not exactly "equations have limitations" that I'm concerned about with Earl's math - he is for sure brilliant in this - but - he has had some severe misconcept about the terms of "soundfield" and "wave front" (*and* to some other topics we went through in the past) that lead me to question what exactly he set up as the ancillary conditions to solve the equations.
This is where we need some more "real" understanding in the first place - no?
Michael
need more?
Its not exactly "equations have limitations" that I'm concerned about with Earl's math - he is for sure brilliant in this - but - he has had some severe misconcept about the terms of "soundfield" and "wave front" (*and* to some other topics we went through in the past) that lead me to question what exactly he set up as the ancillary conditions to solve the equations.
This is where we need some more "real" understanding in the first place - no?
Michael
I think this thread is long overdue. Mostly internal mirror looking. Like this last write up with claims not at all documented as to what is estimated or how. But maybe blueprints can be estimated ?
This can be seen as tragic or just as a case among a lot of others when nothing is to be achieved but more to provide a playground. And yes of course, it's just a hobby !
/Erling
This can be seen as tragic or just as a case among a lot of others when nothing is to be achieved but more to provide a playground.
And yes of course, it's just a hobby !/Erling
mige0 said:
Although BEM generates numerical solutions to the linear wave equation this is insufficient to show diffraction and reflection of acoustic waves?
mige0 said:
Indeed.
But are people happy with their illusions (e.g. the cable thread), happy to pursue a cut-and-try approach and/or happy to banter away where neither party has a good grasp of the principles involved? It is a hobby after all and not everybody finds learning physics and maths fun (or easy).
andy19191 said:
stange, isn't it ?
If I think about why this confusion of wave front versus sound field may come from I have a feeling that its simply because we do not distinguish clearly enough between a wave entering a "virgin space" (wave front) and waves that already have woven a pattern (sound field).
Meaning you could be interested in how the wave front enters virgin space only and not calculate for the interference that build up later on.
This way you get kind a ("virgin space") sound field that isn't worth the name - and I sometimes use the term "established sound field" to pronounce that the sound field *I* mean is with interference included.
To calculate a sound field with interference included you need "additional time" as by definition interference happens later on (at least if you calculate for a single "first" source)
Does this make sense?
Michael
Also if we "meditate" a little on the term diffraction.
There is some ongoing confusion as well – and I have the feeling we do not distinguish clearly enough between diffraction describing *mere bending* and diffraction used to describe a whole bunch of *effects* as a result of that bending
Michael
There is some ongoing confusion as well – and I have the feeling we do not distinguish clearly enough between diffraction describing *mere bending* and diffraction used to describe a whole bunch of *effects* as a result of that bending
Michael
mige0 said:
Yes it would be if a case was made but otherwise...
mige0 said:
Not to someone who considers the linear wave equation to be an approximation to the generally accepted laws of mass, momentum and energy conservation and their closure for a fluid. Such a person would like to know which of the following was the problem:
1. The assumptions made to derive the linear wave equation from the full equations for fluid motion are not holding (e.g. significantly non-isentropic expansion, significant generation of vorticity).
2. The conventionally accepted full governing equations are not holding (e.g. like in expensive audiophile wire).
3. The numerical error for BEM cannot be reduced to insignificant levels by refining the grid.
4. Valid boundary conditions cannot be imposed for BEM.
The models used by the DIYers here, the models used by the low-tech loudspeakers industry or the models used by the medium/high-tech engineering industry?soongsc said:
Agreed but people derive their fun, and in some cases their incomes, from a wide range of interests. I tend to derive more pleasure from solving problems than from, say, making things with my hands. Others are most certainly the other way round. It takes all sorts.soongsc said:
Complexity is not really a function of frequency. Modern methods with automatic grid generation will only require the user to specify a solid model, boundary/initial conditions plus a few controlling parameters. Often a solid model already exists because of other activities like CAD/CAM. The process is straightforward using good quality software.soongsc said:
What resolving the higher frequencies does require is large number of elements and using appropriate numerical methods. The former becomes cheaper year after year as the cost of computing plummets. BEM uses full matrices and is not appropriate for high resolutions, FEM could be made to scale linearly with number of elements but not when calculating modes, other numerical approaches are likely to be more efficient for highly resolved simulations particularly if nonlinear motion is to be included in some regions.
soongsc said:
I won't hold my breath while waiting.
You have no need to hold your breath. The subject area is broadly called Computational AeroAcoustics (CAA):
http://en.wikipedia.org/wiki/Computational_Aeroacoustics
and there is plenty of material about it on the web.
In the same way as the calculator replaced the slide rule, computational approaches have/are in the process of replacing analytical approaches in most areas of engineering. Nobody uses analytical approaches anymore to address structural engineering problems, very few would use them to solve fluid problems but the use of analytical/semi-analytical approaches is still dominant in acoustics. There are good reasons for this related to the size of the relevant physical scales and the fact that the assumption of linearity is a good one for much of acoustics. Nonetheless, the cost of computing is now at a level where the move from analytical/semi-analytical towards computational has begun in high/medium tech industries.
There are lots of examples of BEM/FEM in the low tech loudspeaker industry:
http://www.vibroacoustics.co.uk/audio/fsaudio1.htm
but I am unaware of more complete computational approaches. However, I have not looked and so there is a good chance there are several by now. If you make the effort to browse for some I would be interested to hear back.