Of course there is refraction in audio.
It can be heard at beach in summer with the hot sand, it's exactly the same thing with optic but with a different angle.
An externally hosted image should be here but it was not working when we last tested it.
For exemple in this picture if we change the eye by a mouth and if we have a ear in place of "Normal", the ear hears nothing.
If the device you are referring to is that large white washbowl looking one, then judging by the picture it seems to have a convex then concave thence convex flare.
Used in r.f. applications this is known as a Gaussian profile horn, it being used because it has minimal side lobe production.
If this is so then it would also have a very narrow tapering beam, and the hole in the middle is just caused by this fact.
rcw.
Used in r.f. applications this is known as a Gaussian profile horn, it being used because it has minimal side lobe production.
If this is so then it would also have a very narrow tapering beam, and the hole in the middle is just caused by this fact.
rcw.
Actually no – contour is monotony.
The contour of the "large white washbowl looking one" is pretty similar to what you can see in :
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-67.html#post1930266
The polar performance should also be comparable with some up shift in frequency as the mouth diameter is slightly less than shown there.
But - you never know about the exact wave front delivered by a compression driver.
So auditioning is the final judge I'm afraid.
There were plenty of "big horn" 2-way concepts for movie theatres out there, like the biiig JBL bi-radial's and comparable ones from EV – but mostly combined with 15" or dual 15".
Their sound isn't any bad in the voice region *if* you stay with decent SPL levels and reasonable listening distance – they can almost fool you. They were not as (smoothly) extended in the heights as we demand for a good home speaker of course.
Cranking the volume at complex sounds pushes them into intermodulation distortion easily.
Summing up - a two way could possibly be done for home like levels – but given a *sweet* bandwidth of the min phase horn to be roughly one decade (as pretty usual for horns) we end up at a XO around 1kHz as a starting point for such designs.
For the current make of the min phase horn, a lower XO point of ~700 – 800Hz plus a super tweeter coming in at 8-12 kHz would be my personal preference.
My guess is that because the min phase horn has such smooth roll off at the top, a single capacitor "XO" for the super tweeter would do just fine.
Michael
Last edited:
It must be a trick of the light Micheal but it looks like the sort of horn I mentioned to me, (my poor old bespectacled eyes must be to blame).
Going back in history the two way seems to have began with the Shearer system and this was intended to reproduce the 50-8kHz. Bandwidth optical sound tracks of the time.
Later systems with equalisation went up to around 15kHz.
The two way is simple and can be made with inaudible group delay but with a 15-18inch driver getting bandwidth much higher than 10kHz. Is a problem.
Reckonings that I have done indicate that a two and a half way that covers a 30Hz. 20KHz. bandwidth, with constant directivity from about 1.5k. and minimum phase needs a pair of ten inch drivers, the top one crossing over in this region to the “oskugel” horn I have described, driven by a specially made annular compression driver.
Bi-amped with a suitable digital crossover, and d.s.p this should give a very good replication of the electrical input signal over at least a sixty degree wide twenty high window.
All it needs is someone with a few million bucks to throw at it, although a diy version with a shallow waveguide driven by a seas 27 series tweeter should also be worthwhile.
Rcw.
Going back in history the two way seems to have began with the Shearer system and this was intended to reproduce the 50-8kHz. Bandwidth optical sound tracks of the time.
Later systems with equalisation went up to around 15kHz.
The two way is simple and can be made with inaudible group delay but with a 15-18inch driver getting bandwidth much higher than 10kHz. Is a problem.
Reckonings that I have done indicate that a two and a half way that covers a 30Hz. 20KHz. bandwidth, with constant directivity from about 1.5k. and minimum phase needs a pair of ten inch drivers, the top one crossing over in this region to the “oskugel” horn I have described, driven by a specially made annular compression driver.
Bi-amped with a suitable digital crossover, and d.s.p this should give a very good replication of the electrical input signal over at least a sixty degree wide twenty high window.
All it needs is someone with a few million bucks to throw at it, although a diy version with a shallow waveguide driven by a seas 27 series tweeter should also be worthwhile.
Rcw.
I know one that is darn close. My buddy John's "Big Boys." P-Audio 18" coax on open baffle. He is still working out the crossover kinks.
Dynamic, big, live, and with great depth and imaging. One of the best 2-ways I've heard. We had it at the 2008 RMAF and it even in its early form it easily trounced 90% of what was there.
It can be done.
This one?
taken from : http://www.audioheritage.org/html/profiles/lmco/shearer.htm
Keep us updated please – drawings, measurements, simus etc....
I missed that – are there any pix or further comments on the web on this one?
Michael
When the idea was expressed in terms of "elastic sphere" I assume there to be deformation of the sphere as the energy is transferred through it. That deformation, then return to shape occurs over a period of time, a delay.
When the idea is expressed in terms of a linear spring, there is still some time required for the energy to travel from one end of the spring to the other. This is easier to see in long springs as resonance or expressed as a period. For the transfer to be instantaneous, there can be no deformation. The spring, sphere or transfer medium would have to be rigid, not elastic
Hi Ed
I find this topic extremely interesting as I think there is a veeeery basic mistake in that theory, and yes I was referring to an ideal spring – mass chain system or – if you like ideal "elastic spheres" (of a given mass).
The first question is *if* it is correct to decompose a bunch of "ideal spheres" (of a given mass) into a chain of ideal springs and masses. My answer would be yes, as anything else doesn't make sense to me.
It is certainly true that it takes time to *fully* return when you say:
"That deformation, then return to shape occurs over a period of time...."
– same as for an "ideal spring – mass chain" it takes some time until the last mass is FULLY moved - but the point is, the movements STARTS immediately – hence *no delay* - hence no explanation for the time of flight.
So your conclusion
"...., a delay."
isn't correct.
Very strange indeed.
If in doubt you can go to that very simple math that describes how the acceleration of a mass is defined by a force created by a pressed spring.
You will find NO term of *time* in this !
To state that:
" there is still some time required for the energy to travel from one end of the spring to the other. "
is certainly correct as this exactly defines the time of flight. but the point here is you refer to a "linear" real world spring – whereas I refer to an "ideal spring".
The problem at hand is - that we can NOT explain the time of flight in a "linear" real world spring (air, water, rubber, steel, or any material at all) with the concept of ideal spring mass chains.
At least *I* see no chance here! do you?
Michael
I agree, it's the dephasing due to the reactance of spring, so with frequency crossover effect.
The dephasing is variable with frequency or, if you prefer, the speed of the impulse force.
There was info about it on my website, which is down now. I'll try to get you something. The 18" coax on OB is not a commercial product, just a DIY project at the moment
This is just silly. We know from observation that there is a delay in transmitting the force so any theory that disagrees is just wrong.
More specifically, any math involving a spring uses the spring constant k, usually expressed in newtons/meter which reduces to 9.8*kilograms/second^2. So time is intimately involved in any calcs involving a spring. Water has the equivalent of a higher 'spring constant' (stiffer spring) than air so sound travels faster through water. This is basic freshman-level dynamics and there is no mystery at all to any of it.
Edit: in fluids, the bulk modulus replaces the spring constant. Same idea though except in 3 dimensions.
http://en.wikipedia.org/wiki/Speed_of_sound
Last edited:
" The transmission of sound can be illustrated by using a toy model consisting of an array of balls interconnected by springs. For real material the balls represent molecules and the springs represent the bonds between them. Sound passes through the model by compressing and expanding the springs, transmitting energy to neighboring balls, which transmit energy to their springs, and so on. The speed of sound through the model depends on the stiffness of the springs (stiffer springs transmit energy more quickly). Effects like dispersion and reflection can also be understood using this model."
Wiki is beautiful - but its been written by people too - there are several severe faults in that explanation as already outlined.
If you try to follow my line of argument you immediately see the flaws.
Thend has picked up my point already.
whats that???
Spring constant "k" tells us what force is needed to change spring length (or vice versa)
F = k * delta L
– nothing else - and also no time involved in particular.
Michael
Last edited:
Exactly what balls and springs have to do with sound transmission in air I am not sure.
As a simple model you can look at air as a collection of elastic particles in random motion.
The “mean free path” between these particles is such that they cannot travel very far before banging into another particle.
In this condition the Boltzmann principle of equipartition applies, i.e. for a solid sphere each degree of freedom has the same energy.
In a monotomic gas for instance you can say that the particles have three degrees of freedom, diatomic gasses have five because they have two rotational degrees, this is where Gama comes from, for diatomic gasses it is 1.4, for monotomic 1.6.
From this you can correctly predict that sound travels slower in air because it is largely diatomic and any energy gained by the particles is partitioned to the spin degrees as well as the three spatial ones.
Sound is then an ordered motion superimposed upon this random thermal motion, and when a sufficient number of particles is considered then the ordered motion can be treated statistically as a wave moving in a smooth continuous medium.
Balls and springs can be used as a model if a matrix structure exists that can transmit shear waves, such as in solids.
All of this depends upon the initial assumption that the particles are perfectly elastic and all energy of collision is conserved as motion etc., for air at normal temperatures and pressures this is a reasonable assumption, and the classical theory of sound based upon it works well enough for most practical purposes in audio.
As to there being no time dependence in acceleration, all the laws of motion are “dt”, with respect to time, acceleration being...dv/dt.
In quantum mechanics this is a problem because this implies that a particle can be at an infinite number of points between two points, all quantum mechanical descriptions are then first order in time, acceleration is second order, i.e. d^2x/dt^2.
The formulation of sound that I prefer is one that is based upon first order in time equations because it gives ready access to other first order in time equations. These allow you to make genralisations about the likely and desired solutions to the classical wave equation that are very difficult to do if you just look at it alone.
Rcw.
As a simple model you can look at air as a collection of elastic particles in random motion.
The “mean free path” between these particles is such that they cannot travel very far before banging into another particle.
In this condition the Boltzmann principle of equipartition applies, i.e. for a solid sphere each degree of freedom has the same energy.
In a monotomic gas for instance you can say that the particles have three degrees of freedom, diatomic gasses have five because they have two rotational degrees, this is where Gama comes from, for diatomic gasses it is 1.4, for monotomic 1.6.
From this you can correctly predict that sound travels slower in air because it is largely diatomic and any energy gained by the particles is partitioned to the spin degrees as well as the three spatial ones.
Sound is then an ordered motion superimposed upon this random thermal motion, and when a sufficient number of particles is considered then the ordered motion can be treated statistically as a wave moving in a smooth continuous medium.
Balls and springs can be used as a model if a matrix structure exists that can transmit shear waves, such as in solids.
All of this depends upon the initial assumption that the particles are perfectly elastic and all energy of collision is conserved as motion etc., for air at normal temperatures and pressures this is a reasonable assumption, and the classical theory of sound based upon it works well enough for most practical purposes in audio.
As to there being no time dependence in acceleration, all the laws of motion are “dt”, with respect to time, acceleration being...dv/dt.
In quantum mechanics this is a problem because this implies that a particle can be at an infinite number of points between two points, all quantum mechanical descriptions are then first order in time, acceleration is second order, i.e. d^2x/dt^2.
The formulation of sound that I prefer is one that is based upon first order in time equations because it gives ready access to other first order in time equations. These allow you to make genralisations about the likely and desired solutions to the classical wave equation that are very difficult to do if you just look at it alone.
Rcw.
as usual, I only understand a fraction of what you lay out.
" As to there being no time dependence in acceleration, all the laws of motion are “dt”, with respect to time, acceleration being...dv/dt."
The way you put this seems to tell us that there *is* time (delay) involved in acceleration. And exactly this is the flaw I see.
To go into statistics doesn't help IMO either, as there is no additional time *delay* term generated nor defined there ?
Michael
" As to there being no time dependence in acceleration, all the laws of motion are “dt”, with respect to time, acceleration being...dv/dt."
The way you put this seems to tell us that there *is* time (delay) involved in acceleration. And exactly this is the flaw I see.
To go into statistics doesn't help IMO either, as there is no additional time *delay* term generated nor defined there ?
Michael
In classical mechanics there must be time delay in acceleration.
if you write f=ma, as f=m [dv/dt], if there is no passage of time acceleration becomes infinite, and so does the force needed to cause the acceleration.
This is based upon the assumption that time is a continuous directed quantity.
One of the great ironies of physics is that just about every thing is formulated with respect to time and yet nobody has the slightest idea what time is, or indeed if it has any sort of definable existence at all.
When you talk of time then what are you actually talking about? nobody really knows, it is nevertheless a useful concept.
rcw.
if you write f=ma, as f=m [dv/dt], if there is no passage of time acceleration becomes infinite, and so does the force needed to cause the acceleration.
This is based upon the assumption that time is a continuous directed quantity.
One of the great ironies of physics is that just about every thing is formulated with respect to time and yet nobody has the slightest idea what time is, or indeed if it has any sort of definable existence at all.
When you talk of time then what are you actually talking about? nobody really knows, it is nevertheless a useful concept.
rcw.
Michael, I can see why Earl gets frustrated trying to discuss (argue) wave theory with you.
Force/length is useful in statics. But, if you convert the units, k also means mass/time-squared. That's the "else" used in dynamics. Here in the US, freshman engineering students learn the difference between static behavior such as you describe and dynamic behavior which depends on time. If you had actually read and understood the Wikipedia article on speed of sound, beyond the first couple of sentences that mention the spring and balls, you would have seen the dynamic behavior of fluids and solids explained well enough (with the math) to understand that there is clearly a time component in the dynamic behavior and that the 'speed of sound' is quite real. The equations presented there are (from my fuzzy memory) identical to the ones in my engineering textbooks. And more important, they describe exactly the measured 'time of flight.' Their theory works, yours doesn't.
Try this. Drop a pebble in a pond and watch the waves spread outward. Do the waves reach the shore instantly? No. Does it take time for them to get there? Yes. And before RCW jumps all over me, just like the spring thing, the water wave thing is just a simplified example you can see with your eyes to get an intuitive grasp of what is happening.
Last edited:
Sure - of *paramount* interest even - as its the reason for this thread after all!
(all other stuff is only to spend some quality time...)
I'm curious - what's been your impressions? have you heard yourself ? is it you on the picture ? have any spec details ? ....
Michael
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.