Min Phase Horn with Faital Drivers at German DIY Show in Gelsenkirchen

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I start this split thread on the topic of “Min Phase Horns” to no longer occupy Earl Geddes ones with what he isn’t directly involved in.

An externally hosted image should be here but it was not working when we last tested it.



Roots of “Min Phase Horns” go back some month in discussion here on diyaudio in several threads – where I – considering myself a noob in horns developed towards the idea and the term “min phase horn” which finally got reality ready to purchase for the DIY’er from German Strassäcker company due to the effort “jzagaja” has taken as a manufacurer.

The whole min phase philosophy originally stems from John Kreskovsky who repeatedly was teaching the very basic fact that speakers actually *can* be seen as min phase device.
Meaning that by advanced equalization a close to perfect behaviour can be achieved.

My contribution here was to look at horns as to be seen as pure “diffraction alignment device” meaning that if we want to achieve min phase behaviour over an as wide as possible room angle we have to concentrate on the effects of diffraction in a way to hit exactly that goal.

The basic philosophy behind my min phase horns is presented in my paper down from the capture “Second step of Optimisation”

http://www.kinotechnik.edis.at/pages/diyaudio/DDCD/DDCD_dipole_horn.html


There was a lot of contribution of others to this concept evolving in several threads on diyaudio especially “soongsc” and “rcw” which are somewhere on the same line. It may be worth to go back some month of this discussion.


The first public appearance of a Min Phase Horn took place at German DIY show 2009 in Gelsenkirchen – though it’s wrongly have been announced or anticipated as Kugelwellenhorn there.

Some pretty pictures can be seen in a German DIY forum:
http://www.roehren-und-hoeren.de/phpBB/viewtopic.php?t=7528

It’s a huge min phase horn for which I have calculated the contour, but have not been any further involved in the design process.


It may be helpful to read through following ( not only ! ) postings too:

http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-67.html#post1930197
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-67.html#post1930266
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-71.html#post1988358
http://www.diyaudio.com/forums/multi-way/103872-geddes-waveguides-72.html#post1990489

There is a lot more to read about though to get the whole picture – anyway – I hope to having made a useful starting point to a good discussion about min phase horns - in particular the very *first one in public* as well as any other ones to come (sooncs and rcw - do you hear me? )

Michael
 
Last edited:
Hello Earl,

Where have you seen the profile of the min-phase horn/waveguide with this discontinuity?

Did it was published on DiyAudio?

Best regards from Paris, France

Jean-Michel Le Cléac'h

Exactly - there isn't any discontinuity in min phase horn - but Earl has already noticed he referred to the wrong pix

Michael
 
Its Narrow CD without diffraction that is difficult. To repeat, getting CD with the lowest amount of diffraction is the goal - OS does that, period.

Wrong – diffraction - meant as bending around the corner (the most basic and very historic meaning of that term) is *always* the same.
For any piston like source – moving back and forth – we have 180 deg of diffraction each side to fill 4 PI space.

What really counts is the *alignment of diffraction in space* to get the lowest possible sound field defects *in combination* with controlled directivity over a as wide as possible room angle and a as wide as possible frequency range.

This is actually what’s min phase horn philosophy is all about.

Michael
 
Actually my sentence:

“What really counts is the *alignment of diffraction in space* to get the lowest possible sound field defects *in combination* with controlled directivity over a as wide as possible room angle and a as wide as possible frequency range.”

…is nothing less than my statement regarding a radically new paradigm of how to look at any speaker design and *especially* at horn speaker design.

1. I set the *main* goal in speaker design to be a smooth and consistent resulting sound field
2. I set the goal in speaker design that this sound field has to be of min phase behaviour to an as widely as possible range in space and frequency
3. I claim diffraction to be seen as an “alignment tool” – in fact the only one we actually have (!) – in the process to achieve the goal of 1) and 2)
4. I’m aware that to some degree the basis to benefit from 1) – 3) is nowadays availability of *advanced equalizing* (response shaping at will) – which was not the case at any reasonable price until now


One can argue that neither 1) nor 2) is any new at all, and many good speakers already have been optimized to come close to this goal in the end.
I certainly agree – though the combination of all three points to a stringent philosophy (with quite some impact) *is* new – to my honest knowledge.

*Diffraction*, on how I learned to look at things (during the last several month) is key.
By many seen as some crude mix of “action / reaction” - diffraction got bad reputation as an ill effect in loudspeaker design to be avoided or suppressed at any price - with Earl Geddes to be the most famous and outspoken proponent of this thesis around here.

In my point of view I set *diffraction* or even more precisely *alignment of diffraction* to be the ultimate tool to achieve considerable progress in speaker design.

Quite some difference!

######

To have a semantically precise handle on the subject, I strictly restrict “diffraction” to be a “term of cause” in the historically meaning of

“bending around the corner”

In German its maybe more intuitively to do so, as “diffraction” in a technically understanding here is closely connected to “Beugung” = “bending”

Anyway - to my knowledge “diffraction” as a word of science, prominently entered public awareness with the most famous photos of solar eclipse being taken to prove Einstein’s theory a long time ago.

The most important benefit to stay clear about that sharp definition is that now we immediately see the big bunch of diffraction *effects* as a pure result of alignment.

1.) Any bending of a wave front causes a “second source” – no matter how.
2.) To fill 4 PI space or 2 PI space (in case of infinite baffle) a “one vector” wave front - from a piston moving back and forth - has obviously to be bent by 180deg (4 PI space) respectively by 90deg (2 PI space) either side – no matter how

The alignment of the inevitable “second source” points in space is what I call “diffraction alignment” in short.

Whether we look at a speaker mounted in closed box or at a horn – the outcome in sound field smoothness and consistency depends *only* on “diffraction alignment”


#####

Undertaking a short survey on dampening measures in the light of what I said above, we immediately see that applying dampening materials to “avoid” diffraction is a double sided sword at best and a “nonsense claim” at a closer look.

For one - there simply is *no* way to avoid diffraction. All space available gets filled with sound – no matter how – so its nonsense to claim “avoiding diffraction” by dampening measures.

On the other hand – if we look at how dampening materials actually work like regarding diffraction *effects*, we might get a better feeling in what dampening measures can do for us at most.

For example - let’s take an extreme case of an ideal dampening material and apply that magic dampening material at the *whole surface* of a horn contour.

In its sound field outcome one would have a hard time to distinguish that experimental setup from the compression driver playing without any horn.


The reason behind is, that any ideal dampening material applied at a surface makes that surface virtually disappear and also creates a sharp discontinuity – building for a stoooong second source – comparable to intentionally bending the wave front all around 360deg at a knife like edge.

Obviously – current dampening materials (foam etc.) are far from being ideal absorbers. So the effects taking place will be that there is gradual diffraction / bending at the point where this foam is applied to a surface – meaning it will be a matter of try and error what possibly more easily and more predictably can be achieved by a clever designed solid contour.

As for a dampening foam *plug* inside a horn - as is done by Earl in his OS – IMO, the positive effect is limited to dampen the specific cavity resonance (horn honk) caused by sub optimal contour and mouth reflections.
IMO – basically a “end of pipe” measure and something not even being in *direct* relation to “avoiding ( controlling – at best) diffraction by dampening measures”


Michael
 
Last edited:
To have a semantically precise handle on the subject, I strictly restrict “diffraction” to be a “term of cause” in the historically meaning of

“bending around the corner”

In German its maybe more intuitively to do so, as “diffraction” in a technically understanding here is closely connected to “Beugung” = “bending”

Anyway - to my knowledge “diffraction” as a word of science, prominently entered public awareness with the most famous photos of solar eclipse being taken to prove Einstein’s theory a long time ago.

For me it's not diffraction, it's refraction, synonym of "bending".
Refraction is unidirectional depending of environment, diffraction is omnidirectional depending of geometry.

For me:
Reflection: return or rebound
Refraction: bending
Difraction: decomposing or separating (due to the interferences from reflection for one source)

So I don't know, because I don't understand everything, if your concept is based on an interesting global point of view but developed with some false details.
 
In researching this problem I have come to certain conclusions outlined bellow.....

You can define a general minimum phase system as one that is positive definite.(ref. available).

The usual definition used in signal processing is that in the transfer function there are no zeros in the right half of the complex plane, this breaks down however for spatial systems since pure time delays have this property and yet are non minimum phase.

We can then say that if we have a class of spatial phenomena that can be described as minimum phase, then they must have a symmetrical matrix which has all positive eigenvalues, it is usual to call these Hermitian, but any other solution to a square matrix that also yields all positive eigenvalues is by this definition minimum phase.

We also don't need to consider all solutions to wave equations in a sphere, all we need is a solution for a unidirectional beam, and there are a class of solutions that are called the parabolic approximations that are suitable for this purpose.

Those that use a Pade series approximation are close to exact over +- 40, degrees from the axis, and useful over a full hemisphere.

A convenient feature of this equation is that it is also the Schrodinger equation, and since all solutions to this are matrices with all positive eigenvalues, then all solutions to the Schrodinger equation are minimum phase.

To be minimum phase the field from the radiating element, and the field caused by its interface with the outside space, via the duct, must be potential fields that sum to a third potential field that then maps conformally to the listening area.

If you can design a device that conforms to all of these criteria then it should be possible to get an exact replication of the input electrical waveform at a useful set of points defining a listening area.

I will not ask for comments because I am sure there will be plenty so I leave it there for now.
rcw.
 
In researching this problem I have come to certain conclusions outlined bellow.....

You can define a general minimum phase system as one that is positive definite.(ref. available).

The usual definition used in signal processing is that in the transfer function there are no zeros in the right half of the complex plane, this breaks down however for spatial systems since pure time delays have this property and yet are non minimum phase.

We can then say that if we have a class of spatial phenomena that can be described as minimum phase, then they must have a symmetrical matrix which has all positive eigenvalues, it is usual to call these Hermitian, but any other solution to a square matrix that also yields all positive eigenvalues is by this definition minimum phase.

We also don't need to consider all solutions to wave equations in a sphere, all we need is a solution for a unidirectional beam, and there are a class of solutions that are called the parabolic approximations that are suitable for this purpose.

Those that use a Pade series approximation are close to exact over +- 40, degrees from the axis, and useful over a full hemisphere.

A convenient feature of this equation is that it is also the Schrodinger equation, and since all solutions to this are matrices with all positive eigenvalues, then all solutions to the Schrodinger equation are minimum phase.

To be minimum phase the field from the radiating element, and the field caused by its interface with the outside space, via the duct, must be potential fields that sum to a third potential field that then maps conformally to the listening area.

If you can design a device that conforms to all of these criteria then it should be possible to get an exact replication of the input electrical waveform at a useful set of points defining a listening area.

I will not ask for comments because I am sure there will be plenty so I leave it there for now.
rcw.
 
For me it's not diffraction, it's refraction, synonym of "bending".
Refraction is unidirectional depending of environment, diffraction is omnidirectional depending of geometry.

For me:
Reflection: return or rebound
Refraction: bending
Difraction: decomposing or separating (due to the interferences from reflection for one source)

So I don't know, because I don't understand everything, if your concept is based on an interesting global point of view but developed with some false details.


Refraction ('"Brechung" in German) - at least as I understand it - is a variation in the vector (in-angle versus out-angle) when a ray changes from one medium to another one *and* does not pass the border at *exactly* perpendicular angle.

So light going through a lens at any non perpendicular angle gets "bent" or "refracted" - whereas the "bending" of light rays at the very rim of the lens (aperture) is called "diffraction" - basically *this* bending / diffraction would also happen if there is no lens mounted and there would only be a hole of a certain diameter.
Its the physics experiment in school where they show us how slits widen the parallel light beam going through.
In all photographers all days experience its the fact that pix do *not* get any sharper below a aperture of ~8 – 16 quite in contrary.


Also - in my reference example of the historic photos captured to prove Einstein's hypothesis its *diffraction* which was the underlaying mechanism of bending the sun's light around the moon - not refraction.
These crazy guys form England traveled around halve the world to take a picture of the total eclipse of 29 May 1919, that in the end made Einstein famous even in his days.

So, I can't exactly see how "refraction" enters the picture in audio - no?



Michael
 
Last edited:
In researching this problem I have come to certain conclusions outlined bellow.....

...I will not ask for comments because I am sure there will be plenty so I leave it there for now.
rcw.

LOL - I'm really happy you added that last line ! - as I was listening your words like coming from outer space - not having understood a single sentence.
:)

Michael
 
The observations you mentioned were a confirmation of the general relativity prediction that the presence of mass warps space-time and causes light to travel in a geodesic, i.e. the path of least distance between two points, that is not a Euclidean straight line.

As I have posted previously, if you look at air as an assemblage of elastic spheres in random motion, then given certain conditions that can be shown to be true of air, we can describe the behavior of these statistically as waves, this is the Rayleigh theory of sound.

This is a theory of classical statistical mechanics, the mentioned Schrodinger equation is of quantum mechanics.

In this single particles can behave statistically like waves, in classical mechanics only large assemblages of particles can.

In this single particles can exhibit the various phenomena such as refraction interference that are restricted to waves in the classical case, for particles these phenomena are in the general category of scattering, and by using the concept of scattering many things become conceptually easier and mathematically simpler.

This is all standard physics, as many looking at this forum will no doubt observe, but I put it forth to illustrate how we might clear a great deal of tree out of the way the better to see the wood, and come up with a formulation that deals a set of pre-conditions that our device must meet in order to do what we want.

By this means we don't have to make a numerical frontal assault upon the Helmholtz equation, using an arbitrary set of parameters if we want to find something out, (many people post results that do this and usually find out that what they model dosen't work, and don't know why).

What I set out to do was to know before hand what the required solution will look like and just use number crunching to optimize it, this is after all its intention.

As someone once said if you give an infinite number of monkeys a typewriter each then after a particular time they will write the complete works of Shakespeare.
I have neither the time nor inclination for such an exercise, and if you have no theoretical basis to provide you with initial parameters and indicate where a solution might lie then that is what you are doing.
Rcw.
 
Ex-Moderator R.I.P.
Joined 2005
If remember correct there have been a picture of compression driver
Looked to me like the model with multiple small neos, and the most expencive 2" Faital

I honestly dont understand this choise
Is it chosen just because its the most expencive
It clearly shows a peak around 15khz
Same problem with the smaller 1.5" variant
Problems with sibilant rough highs may be partly due to this

The two other 1.5" and 2" with neo disc magnets doesnt appear to have this peak
Do these people know what they are doing
Or do they choose by "expencive sounds better", or just the fancy look fore marketing
Is priority higher profit
Or do they simply not know better
 
Last edited:
The observations you mentioned were a confirmation of the general relativity prediction that the presence of mass warps space-time and causes light to travel in a geodesic, i.e. the path of least distance between two points, that is not a Euclidean straight line.

As I have posted previously, if you look at air as an assemblage of elastic spheres in random motion, then given certain conditions that can be shown to be true of air, we can describe the behavior of these statistically as waves, this is the Rayleigh theory of sound.

This is a theory of classical statistical mechanics, the mentioned Schrodinger equation is of quantum mechanics.

In this single particles can behave statistically like waves, in classical mechanics only large assemblages of particles can.

In this single particles can exhibit the various phenomena such as refraction interference that are restricted to waves in the classical case, for particles these phenomena are in the general category of scattering, and by using the concept of scattering many things become conceptually easier and mathematically simpler.

This is all standard physics, as many looking at this forum will no doubt observe, but I put it forth to illustrate how we might clear a great deal of tree out of the way the better to see the wood, and come up with a formulation that deals a set of pre-conditions that our device must meet in order to do what we want.

By this means we don't have to make a numerical frontal assault upon the Helmholtz equation, using an arbitrary set of parameters if we want to find something out, (many people post results that do this and usually find out that what they model dosen't work, and don't know why).

What I set out to do was to know before hand what the required solution will look like and just use number crunching to optimize it, this is after all its intention.

As someone once said if you give an infinite number of monkeys a typewriter each then after a particular time they will write the complete works of Shakespeare.
I have neither the time nor inclination for such an exercise, and if you have no theoretical basis to provide you with initial parameters and indicate where a solution might lie then that is what you are doing.
Rcw.

What if you understand all that and still get the idea it's best to listen to your tweeters through several inches of foam.
 
If remember correct there have been a picture of compression driver
Looked to me like the model with multiple small neos, and the most expencive 2" Faital

I honestly dont understand this choise
Is it chosen just because its the most expencive
It clearly shows a peak around 15khz
Same problem with the smaller 1.5" variant
Problems with sibilant rough highs may be partly due to this

The two other 1.5" and 2" with neo disc magnets doesnt appear to have this peak
Do these people know what they are doing
Or do they choose by "expencive sounds better", or just the fancy look fore marketing
Is priority higher profit
Or do they simply not know better


The neo disks are actually a cheap way out of a large ring magnet. It's also a good way to generate several flux shorts. My favorite is when all kinds of complex analysis is used to arrive at a pole piece with large unsaturated volume.
 
Last edited:
As I have posted previously, if you look at air as an assemblage of elastic spheres in random motion, then given certain conditions that can be shown to be true of air, we can describe the behavior of these statistically as waves, this is the Rayleigh theory of sound.

This is a theory of classical statistical mechanics, ....
Rcw.

You really seem to be the right person to ask a OT question I have for a long time now.

Given the picture of that " air as an assemblage of elastic spheres in random motion " – where actually stems the "time of flight" in air from?

I mean – if the elastic spheres would have *no distance* in between of each other and therefor no time would be needed to travel before they bump the next one to transfer the energy of a wave – well – then we would end up with infinite speed of sound in that media.

On the other hand – if we consider the elastic spheres having some free room around each other and therefor some time would be needed to travel before they bump the next one to transfer the energy of a wave – well – then we would end up with speed of sound to be dependant on the initial impact.

Obviously, both isn't true.

Not sure if I have made clear my point of dilemma – what are your thoughts on that?

Michael
 
Last edited:
If remember correct there have been a picture of compression driver
Looked to me like the model with multiple small neos, and the most expencive 2" Faital



The neo disks are actually a cheap way out of a large ring magnet. It's also a good way to generate several flux shorts. My favorite is when all kinds of complex analysis is used to arrive at a pole piece with large unsaturated volume.

I noticed that picture of the Faital driver on the German DIY forum as well – but didn't do further investigation.

Thanks for pointing this out !

In this case it seems they have mounted the 1.4" Faital HF 140

http://www.faitalpro.com/products/schede/cd.php?id=502020150
An externally hosted image should be here but it was not working when we last tested it.




on a horn I actually calculated for the 2" Faital HF30AT

http://www.faitalpro.com/products/schede/cd.php?id=502030100
An externally hosted image should be here but it was not working when we last tested it.



Ui ui ui - no good!


Michael
 
Last edited:
Administrator
Joined 2004
Paid Member
In this case it seems they have mounted the 1.4" Faital HF 140
on a horn I actually calculated for the 2" Faital HF30AT
\

No, that would not be too good!

But here is a stupid question:
Given that the criticism was of no central image - were the drivers out of phase - polarity flipped? It's an easy mistake to make, has happened to me often enough.

It could have been something as simple and silly as that. Tho I doubt you'd ever find out, at this point.
 
Infinite speed would be approached if the spheres were rigid. ;)

I doubt - let me explain :

To construct for a simple experimental setup - assume a chain of - say 3 ideal masses coupled by say 2 ideal springs (which equals a set of three elastic spheres).

Then we apply force to the first mass only ans watch the other ones - where do you see any possibility to have a delay?

Sure - the third mass in the chain will not move with the same *initial speed* but for certain even the third mass *must* move the very same moment where at which the first mass in the chaing gets moved by the force we apply.

So - by definition - there is *no* "delay" - IMO.


Michael
 
\

No, that would not be too good!

But here is a stupid question:
Given that the criticism was of no central image - were the drivers out of phase - polarity flipped? It's an easy mistake to make, has happened to me often enough.

It could have been something as simple and silly as that. Tho I doubt you'd ever find out, at this point.

From all the impressions I've read of, I would conclude that there were more than a single point of improvement to be done to tell us more about the sonics of that very min phase horn contour.

I agree that the problem with the wrong imaging most certainly may have been caused by a polarity flip - not discovered at the stress of the show. If so - no big deal.

The sizzling heights may be a combination - as a second pair of speakers showed the same problem - its been either a undiscovered problem in the electronics and/or the personal "preference" of the person who did the voicing.

I really hope for good polar measurements and possibly for a second set up with the driver intended.
Not sure if a two way is possible at all, if we are after top notch sound. I mean - are there any examples for really good working combinations of a 18" plus a 1,5"? If *that* could be made sounding smooth and live like throughout - it would be an *exceptional* design IMO


Given what we must assume for now - I'm still happy that (male ?) voices seemingly had good tonality and ease of presentation.

Michael
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.