Key word is diffraction, from edge/termination.
I was just thinking that the designer should learn to know what (edge) dimensions/discontinuities/angulations and inter-driver distances create diffraction/bending/interference at certain wavelength. With a multiway loudspeaker one has more means to overcome these problems by choosing smart driver sizes, baffle shape and xo points. Positioning, edge roundover, waveguide etc. can be used to alleviate these problems.
ps. a dipole doesn't have edge diffraction ripples below certain F, because front/back wave summation overrules them! This gets very difficult to achieve above 5-6kHz...
I was just thinking that the designer should learn to know what (edge) dimensions/discontinuities/angulations and inter-driver distances create diffraction/bending/interference at certain wavelength. With a multiway loudspeaker one has more means to overcome these problems by choosing smart driver sizes, baffle shape and xo points. Positioning, edge roundover, waveguide etc. can be used to alleviate these problems.
ps. a dipole doesn't have edge diffraction ripples below certain F, because front/back wave summation overrules them! This gets very difficult to achieve above 5-6kHz...
Last edited:
You didn't understand my comment correctly. All front facing closed box speakers have narrowing dispersion, its just a matter of how it narrows.
Here is an amusing solution to diffraction and similar issues from our cousin web site audio asylum. This is, er, a padded cell so to speak
The ole felt tweak - emailtim - Tweakers' Asylum
Impressive. I'm guessing photoshop, but I wouldn't be surprised if someone actually did that.
A rounded surface will get rid of 100% diffraction. Another thing to consider is if you have EQ then position drivers so that the off axis is consistent. That way you can equalize to all axes. This off axis is a major problem for box spekers.
But as many people know too much direct sound can kill ambience and you might as well use head phones.
And too much direct sound is why certain kinds of distortion can sound pleasant.
Last edited:
can you supply evidence of that?
sorry but that sounds like something you just made up so please provide some corroboration.
and since when does amplitude correction (EQ) change off axis response that's also new to me.
Last edited:
This is not strictly true. It's more so true for direct radiating drivers than horn loaded drivers.
Sorry, this is not true. Even if an object is round it still diffracts.
It's not possible to do this because diffraction around any solid physical object is frequency dependent. At lower frequencies the object is too small compared to the wavelength and there is no diffraction. At high frequencies the object appears very large compared to the wavelength, as if it were a wall or plane. In between you get complex transition behavior that depends on the shape of the physical object. It's only possible to influence this somewhat by "positioning of drivers" on the loudspeaker, e.g. off-center or near an edge. The driver's inherent radiation pattern (direct radiating, horn loaded) and the radiation pattern of the driver-in-loudspeaker (monopole, dipole, etc.) have much influence over the on- and off-axis radiation patterns versus frequency. EQ can only boost or cut the magnitude of the radiation on ALL AXES at the same time, so it cannot influence the pattern itself. This is one reason why you often see multi-axis frequency response plots normalized to the on-axis frequency response. This presentation make it easier to see the radiation pattern, and to recognize resonances and other artifacts independent of the on-axis frequency response.
Siegfried Linkwitz described this problem for boxed loudspeakers in his talk last year at Burning Amp. Here is a pdf of his presentation.
http://www.linkwitzlab.com/BA-2017/Radiation-BA2017.pdf
Pages 8 and 9 and will help to illustrate that EQ cannot improve a bad pattern.
Nope, that is NOT CORRECT. Alas, doubling down on your statement does not make it true. That only seems to work if you are a certain high government official...
I suggest you try Google search. Try "diffraction from a sphere acoustic" and start reading. Diffraction is a form of scattering so you can also search using the terms "acoustics scattering". There are some images under these terms that might help to demonstrate the effect.
He listens to classical music which is very ambient. I'm not sure how much of a golden ear he is if he thinks an aurasound whisper sounds as good as a nice tweeter. It doesn't, even with equalization.
As far as diffraction see my explanation. I clearly say "relative to frequency".
You seem to also misunderstand what I said about equalization. I clearly imply that it effects the axes. What I said is that the drivers can be positioned so that they have a consistent off axis pattern.
Meaning that if you lay out the baffle so that the same bumps and dips appear at different axes, then equalizing one axis will help another.
And yes you can do this, but creating prominent dips and peaks is usually the opposite of what most designers do in a passive set up.
WRONG, You're making a classic "out of context" argument. In this context we are talking about baffle edge diffraction which refers to a specific phenomenon where the sound wave creates an out of phase reflection due to the impedance change.
Understanding Cabinet Diffraction – Audioblog
Why don't you explain how equalization doesn't affect more than one axis as implied by your statement?
When visiting the grocery store, I came up with this:
To minimize diffractions, use a narrow(ing) baffle and 3-4 drivers. With tweeter use a shallow wide waveguide.
The problem is that many like more directivity than the principle above gives. Then one must use maximally big drivers with great care in their narrowing band and/or deep horns. In my view a synergy horn is close to ideal highly directive loudspeaker.
A dipole actually has very little directivity until above 30¤ off-axis. Then there is the back wave creating it's own reflections in the room.
So, one must just choose her poison! Breaking/mixing these "rules" makes a lousy loudspeaker!
To minimize diffractions, use a narrow(ing) baffle and 3-4 drivers. With tweeter use a shallow wide waveguide.
The problem is that many like more directivity than the principle above gives. Then one must use maximally big drivers with great care in their narrowing band and/or deep horns. In my view a synergy horn is close to ideal highly directive loudspeaker.
A dipole actually has very little directivity until above 30¤ off-axis. Then there is the back wave creating it's own reflections in the room.
So, one must just choose her poison! Breaking/mixing these "rules" makes a lousy loudspeaker!
Last edited:
No, certainly not for any boxed loudspeaker and for any appreciable frequency and angle range. It is just not possible. Also, it's important that both the on and off axis are very similar except for the SPL level, not just a similar "off axis pattern" (meaning the on axis might be different still).
Could you share one of these magical driver positioning arrangements so we can all see what you are talking about?
Thanks, I think I have a pretty good grasp of diffraction...
Just like a sphere in empty space the curved edge WILL SCATTER and cause diffraction. So, no, what I said applies perfectly well to a rounded cabinet edge, too. In fact, you can make the edge radius as large as half the cabinet width if you want. Then you will have... a sphere! It will still diffract. Sorry to keep driving this home, but it's true. The only thing that is really happening as you make the edge radius larger and larger is that the diffraction "signature" gets smoother. Olsen showed this in the 1950s I think.
Generally speaking, even the largest practical cabinet roundovers are only influencing the diffraction signature of the tweeter, e.g. above 1-2kHz. They do nothing for lower frequencies.
Not possible? Try putting a single driver in the middle of a circular flat baffle and you will create a more consistent pattern. Linkwitz pluto is similar to this (just a round driver), and uses equalization.
A square box can't be perfectly consistent, BUT I never said that it could, just more consistent at certain positions. You seem to have trouble understanding what's actually being said here.
Again, I am not not saying that the sound doesn't technically "diffract" around a round baffle. This is what diffraction is. But the context of the discussion here is the diffraction signature or ripples created by corners. Context matters. When people talk about controlling diffraction in a speaker that's what they mean.
And yes Olsen documented that a sphere has a smooth diffraction signature, in fact the diffraction of a sphere is the baffle step which is in this context full diffraction control.
A square box can't be perfectly consistent, BUT I never said that it could, just more consistent at certain positions. You seem to have trouble understanding what's actually being said here.
Again, I am not not saying that the sound doesn't technically "diffract" around a round baffle. This is what diffraction is. But the context of the discussion here is the diffraction signature or ripples created by corners. Context matters. When people talk about controlling diffraction in a speaker that's what they mean.
And yes Olsen documented that a sphere has a smooth diffraction signature, in fact the diffraction of a sphere is the baffle step which is in this context full diffraction control.
Last edited:
That's a good design. I have some speakers that are like that and I love the sound, but they need a lot of space around them because of the mid range dispersion.
One of the things that holds back wave guides though is the hump in response. Also they benefit from steeper slopes. All in all I'd say waveguides are better suited to active systems. But they sound smoother IME.
I think that Siegfried (rest in peace) and I differ in what we are after. He is after ambience and spaciousness in almost exclusively large venue recordings. I am after imaging in almost exclusively studio recordings. I have come to the conclusion that the two approaches are not compatible, especially in a small room. You must choose one or the other.
This was brought home to me more than ever before just a few days ago when I went to the symphony. The sound was massive, ambient and spacious but there simply was no image. There was no singular location from which the sound of the violin section was emitted. Not like one would get in a studio recording of a single violin. The single violin can be reproduced in my listening room with almost perfect fidelity. A large room recording can never sound like the original space in a small room - too many conflicting aspects of room/recording.
Siegfried relied far too much on his personal impressions of sound, doing virtually no blind large subject testing, ala Toole. Hence I have always taken his conclusions with a grain of salt, unlike Floyd's, whom I almost never have disagreed with (there are some small things regarding directivity in small rooms. He prefers a lower DI that I design for, but then he is also exclusively a large venue recording person, and I am not.)
I have listened to many of Siegfried's designs, none have impressed me very much other than the bass in his larger dipoles was very good. But I achieve comparable bass with multiple subs. I found his high end sound sadly lacking however.
One point that I should make on baffle diffraction is that just because a source has a high far-field directivity does not mean that it has this same directivity in the near field - on the baffle. The sound will always propagate across the baffle to a certain degree.
And how much diffraction is audible has only ever been studied by myself, that I know of. The answer is, of course, it depends. Since its audibility increases with SPL, any amount will always be audible above some SPL, just as any amount will always be INaudible below some SPL. How can one make any claims about audibility in that situation. All that one could say, and I DO make this claim, is that, in my system diffraction is not audible even at exceedingly high SPLs.
And how much diffraction is audible has only ever been studied by myself, that I know of. The answer is, of course, it depends. Since its audibility increases with SPL, any amount will always be audible above some SPL, just as any amount will always be INaudible below some SPL. How can one make any claims about audibility in that situation. All that one could say, and I DO make this claim, is that, in my system diffraction is not audible even at exceedingly high SPLs.
Last edited:
I don't know what you mean by "consistent", but the directivity will most certainly not be constant. It will beam as the frequency goes up.
Thank you, that makes some sense to me.
I visit concerts of philharmonic orchestra several times a year and I know what you are speaking about. However, in a recording of a big orchestra, number of microphones, their positioning and mixing makes the resulting sound and imaging and localization in the recorded sound. Imaging from a recording of classical music is usually much better than that you may ever hear in a concert hall (yes because of reflections of all kinds and long sound paths to the listener's ears), though the sound is never that "massive" as in a concert hall. I have some experience from Supraphon classical music recording studios, 40 years ago, during my university studies, microphone placement is a key.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.