Beyond the Ariel

[gedlee]

Originally Posted by John_E_Janowitz

Secondly you want nonparallel surfaces and unequal distances to help avoid standing waves from developing.
John

John

This is simply not true. It can be proven that any fixed volume, despite its shape (reasoanably low aspect ratios assumed) will have the same modal density and the same frequency of its first resonance. The modes shapes will differ in every case, but the number of modes in any fixed frequency interval is virtually independent of shape. This is not true of panels, of course, but it is true of the acoustic interior - they are entirely different problems.

[ScottG]

Originally Posted by gedlee

John

This is simply not true. It can be proven that any fixed volume, despite its shape (reasoanably low aspect ratios assumed) will have the same modal density and the same frequency of its first resonance. The modes shapes will differ in every case, but the number of modes in any fixed frequency interval is virtually independent of shape. This is not true of panels, of course, but it is true of the acoustic interior - they are entirely different problems.

..but you can *add* non-parallel surfaces (adjusting for added displacement) with the same volume.

[sendler]

Originally Posted by gedlee

John

This is simply not true. It can be proven that any fixed volume, despite its shape (reasoanably low aspect ratios assumed) will have the same modal density and the same frequency of its first resonance. The modes shapes will differ in every case, but the number of modes in any fixed frequency interval is virtually independent of shape. This is not true of panels, of course, but it is true of the acoustic interior - they are entirely different problems.

So, a floor standing cabinet with internal measurements 12" x 12" x 48" will have the same internal box modes as a 19" cube?

[ScottG]

Originally Posted by sendler

So, a floor standing cabinet with internal measurements 12" x 12" x 48" will have the same internal box modes as a 19" cube?

Originally Posted by gedlee

..The modes shapes will differ in every case..

[gedlee]

Originally Posted by sendler

So, a floor standing cabinet with internal measurements 12" x 12" x 48" will have the same internal box modes as a 19" cube?

Thats a very large change in aspect ratio so it deviates at the very lowest frequency, but as the frequency goes up the "modal density" of the two becomes the same, yes. If you reread my claim, I said that the aspect ratio needed to stay approximately the same. For example; take your cube and tilt every wall 20 degerss of so, there will be no change in the frequency of the lowest mode. Start stretching it and yes, the lowest mode lowers, but the density stays constant.

[sendler]

Originally Posted by gedlee

Thats a very large change in aspect ratio so it deviates at the very lowest frequency, but as the frequency goes up the "modal density" of the two becomes the same, yes. If you reread my claim, I said that the aspect ratio needed to stay approximately the same. For example; take your cube and tilt every wall 20 degerss of so, there will be no change in the frequency of the lowest mode. Start stretching it and yes, the lowest mode lowers, but the density stays constant.

I see what you are getting at now. Making the walls non-parallel or adding internal diffusion will change the distribution and amplitude of the internal modes very little?

[chrismercurio]

Gary,

I hope the opera was wonderful. It must be quite a rush, even though it's a lot of work' to conduct one of those.

Did you spring for the Apollo option for your TD drivers?

Thanks,

Chris

[John_E_Janowitz]

Yes, if you take a cube and then skew the sides to make a parallelogram there won't be that much difference. However if you go and compare a cube to something the shape of an axe head, there will be a noticeable difference even even if they are the same volume.

John

[mige0]

Thanks guys for correcting me.
Yes I was wrong – well – partly wrong – or in other words : not completely right

*But* you are not completely right either.


Where I made a mistake is that „consecutively min phase *in general* can not be corrected“, well I agree it *can* be corrected – if possibly not in any case.
But as thats after a common agreement on „consecutively min phase“ as such taking place – I leave that for the moment (guess it will be a lot of fun to discuss correct-ability by its own).

To get a better handle on the concept of „consecutively min phase“ I've set up some basic simus to give a picture.

It simulates the behaviour of two point sources placed at 1ms delay time distance to make it easy to grasp the effect in a qualitatively manner at first hand.

The traces shown are :
the original signal in RED
the overlay of the two point sources WITHOUT any low pass filtering of the second part of the doublet in BROWN
the overlay of the two point sources WITH an (arbitrary) low pass filter for the second part of the doublet in BLUE



To put things into perspective, different sine bursts are shown at peaks and nulls and also at frequencies well below „first baffle peak“
Assumed here is an ideal speaker in the sense that there is no LP or HP for the point source in front


2000 Hz :


1500 Hz :


1000 Hz :


750 Hz :


500 Hz :


250 Hz :


125 Hz :


75 Hz :


25 Hz :



The BROWN trace is the trace of an „idealized“ OB dipole showing perfect nulls and peaks – or any other speaker with discrete time delay (and equally applies to room effects) as mentioned above.

We see the point of „changing gear“ is at 1ms – befor that we only hear the first half of the doublet (monopole presentation), after that, the second half of the doublet is overlayed and we actually listen to dipole.
In the simplified simu at hand, the two consecutively min phase parts are sharply seperated from each other.
So, at 1ms there is a considerable kink in the „SPL pressure“

With frequencies going up the low pass filter (taking somehow care of frequency dependent diffraction / beaming effects affecting the second part of the doublet) becomes more and more effective, what leads to a flattenig of the peaks and nulls that would happen otherwise.

Now – that we know what to watch out, real world measurements would be fine for confirmation.

The challenge here is, that usually the two min phase parts are not as sharply separated as in the simu – so, looking for a sharp and distinct kink would probably be not the best idea – anyway I guess the effect is real and even if not happening in the pronounced way shown by the simus, further investigation seems interesting.



Michael

[mige0]

Originally Posted by dlr

...
One can conceptualize whatever one wants to try to present it, this changes nothing. There actually is no "doublet" or second impulse in reality, since diffraction of any baffle and more especially with any real non-point source driver is a distributed issue. The concept of "consecutively min phase behaviour" of which you speak might be a better way to describe it as it is the integrated response from the diffracting points.
...

Dave

As much as I agree that the effect of „consecutively min phase“behaviour will often be masked, I nevertheless find it worth the effort to bring down the discussion to a point where common ground could possibly be achieved

This is out of two reasons
1.) the consequence of the concept of CMPB is a pretty unique form of distortion that isn't exactly easy to deal with and I doubt many are aware of its shere existence (hence Lynn is referencing to the "multi path" issue in the HF department - simply no naming or consciousness here in audio)
2.) I haven't seen this specific form of distortion to be characterized and outlined elsewhere ( thouigh I possibly could have overlooked it easily )

And again - I'm not discussion 3-D here nor does John nor Jean-Michel - though the issue of course is highly 3-D connected.

The core issue *for the listener* though lyes in the time domain as I tried to make clear with my simus above.
My posts are in continuation to the same - pretty long - discussion about that subject, that started out way back in this thread - spread to other threads - and comes back here again.


Michael