Okay so no one can see me and my username is not my actual name so where's the risk of looking stupid?
Here goes. To me it seems counter intuitive to place ESL panels in a convex curve/array. Dynamic cones are placed concave side out when they could just as easily be built with convex side out. Since beaming is the issue wouldn't a concave ESL panel not send waves cross court so to speak the way dynamic speakers do? Thereby spreading rather than focusing? It seems the convex array sends the sound waves away from your ears without the mitigating nature a concave array presents. It's only that thin line or single panel directly facing you that is on axis.
My experience in general has been that the 'head in a vice' scenario is the position you take regardless. In fact, the better the speakers are, the more this is required to realize there full capability...the better the system, the smaller the sweet spot. They may sound great off axis, even better than other speakers on axis, but rest assured you gotta put your head in that vice to hear them at their best.
Here goes. To me it seems counter intuitive to place ESL panels in a convex curve/array. Dynamic cones are placed concave side out when they could just as easily be built with convex side out. Since beaming is the issue wouldn't a concave ESL panel not send waves cross court so to speak the way dynamic speakers do? Thereby spreading rather than focusing? It seems the convex array sends the sound waves away from your ears without the mitigating nature a concave array presents. It's only that thin line or single panel directly facing you that is on axis.
My experience in general has been that the 'head in a vice' scenario is the position you take regardless. In fact, the better the speakers are, the more this is required to realize there full capability...the better the system, the smaller the sweet spot. They may sound great off axis, even better than other speakers on axis, but rest assured you gotta put your head in that vice to hear them at their best.
The included angle of a concave cone surface (or array) restricts the dispersion,
compared to a flat disc shape. A convex surface (like a dome tweeter) does not.
The best systems do have a large sweet spot, especially a coaxial design like the KEF LS50.
An approximation to a point source is best, though room acoustics then become more critical.
Last edited:
I don't have any technical explanations for you, your question just jarred my memory of an article in Speaker Builder Magazine circa 1998.
Matter of fact, I think I still have the full issue around here somewhere, although I'd have to do some digging.
Anyway, first page of the article and picture of Bill Waslo's focused array is here:
http://libinst.com/Article_Pg_1.pdf
I'm in the process of building an electrically segmented ESL that eliminates the 'head in vice' syndrome, for the most part..
My ESL history started with a set of R. Sanders' flat panels back in '99, and two years ago, I build a set of curved panels similar to a large ML panel, that build thread is around here somewhere.
Cheers
-Steve
Matter of fact, I think I still have the full issue around here somewhere, although I'd have to do some digging.
Anyway, first page of the article and picture of Bill Waslo's focused array is here:
http://libinst.com/Article_Pg_1.pdf
I'm in the process of building an electrically segmented ESL that eliminates the 'head in vice' syndrome, for the most part..
My ESL history started with a set of R. Sanders' flat panels back in '99, and two years ago, I build a set of curved panels similar to a large ML panel, that build thread is around here somewhere.
Cheers
-Steve
Last edited:
curved ESL do radiate curved wavefronts at frequencies where the wavelengths are much longer than the flat strip width (for segmented, curved ESL)
so they can approximate line sources
voice coil drivers only move along one axis - the cone shape doesn't change that fact - they are all called "piston drivers"
so they can approximate line sources
voice coil drivers only move along one axis - the cone shape doesn't change that fact - they are all called "piston drivers"
Since I know from your other posts that you are experienced with earlier Acoustat ESL's, you should do yourself a favor and try to audition one of the later Acoustat Spectra models. These models use vertically-segmented panels, whereby only a half panel-width is played full range, a larger area on either side plays mids and lows only, and finally, an area on the outer edge plays lows only. (Yes, the speakers were produced in mirror-image pairs.)
This solution addressed the age-old compromise of the narrow models having the best imaging but limited dynamic capacity and bass response, compared with the wider models that had only so-so imaging but quite acceptable dynamic capacity and bass response. In the Spectra models, the speaker is essentially varying its width with frequency, offering the best of both worlds. In my opinion, the Spectra models have the best imaging of any of the Acoustat speakers, and although they do still have a "sweet spot" like any speaker, it is considerably less finicky that previous designs. I find that listening off-axis is hardly a compromise at all, although of course for critical listening, I do prefer to be smack-dab in the middle.
Across the range of Spectra models, the imaging characteristic is virtually identical. What you get as the total array size goes up is increasing dynamic capacity and more robust bass response without any loss of imaging precision.
BTW, Spectra is an acronym for Symmetric Pair Electrically Curved TRAnsducer. It may interest you to know that the panels are NOT arranged in an arc like all previous full-area models, because the "curve" is introduced in the time domain via the delay introduced by the simple RC filters that progressively roll off the highs and mids to the flanking segments. Therefore, the speaker is a convex curve when viewed from either front or back! This of course, is not possible when arranging the panels in a mechanical arc.
Last edited:
Conventional electromaganetic (cone) drivers) typically operate at in a regime where the dimensions of the cone are always smaller than a wavelength. In this regime, the shape of the cone is irrelevant from the sound production point of view. The conical shape is there only to stiffen the cone and make it behave like a piston - all parts of the cone moving together.
These operating conditions largely ensure that sound is radiated in all directions.
You might notice in some full-range drivers that the cone is not strictly conical - it is curved and flatter near the surround. At low frequencies it behaves as a piston, then as the frequency increases, the outer parts of the cone move less because they are not so stiff. At the highest frequencies, more or less only the button in the middle of the cone moves.
In contrast, ESLs mostly operate in a regime where the membrane width is larger than a wavelength, and this is what makes them highly directional. The segmentation reduces this problem by reducing the width of the radiating surface at high frequencies.
Very crudely, the sweet spot for a non-segmented ESL is often less than two degrees, about the width of the listeners head at 3 m - hence the head in a vice problem. With segmented ESLs, the sweet spot expands to about 20 degrees.
These operating conditions largely ensure that sound is radiated in all directions.
You might notice in some full-range drivers that the cone is not strictly conical - it is curved and flatter near the surround. At low frequencies it behaves as a piston, then as the frequency increases, the outer parts of the cone move less because they are not so stiff. At the highest frequencies, more or less only the button in the middle of the cone moves.
In contrast, ESLs mostly operate in a regime where the membrane width is larger than a wavelength, and this is what makes them highly directional. The segmentation reduces this problem by reducing the width of the radiating surface at high frequencies.
Very crudely, the sweet spot for a non-segmented ESL is often less than two degrees, about the width of the listeners head at 3 m - hence the head in a vice problem. With segmented ESLs, the sweet spot expands to about 20 degrees.
There are many consequences related to dispersion in ESL speakers but mostly fixed by room arrangements.
For many us who have only one sweet spot seat, virtually no real problem related to dispersion that isn't trivially repaired by speaker placement and DSP EQ.
For the record, I use a 2x3 array (6 cell) of flat cells in a slightly cylindrical panel.
Dayton-Wright had a 2-high x 4-wide array inside each speaker on the surface of a sphere! His theory was that it would be virtually like a point source (meaning the point at the origin of the sphere). Think about that.
Ben
For many us who have only one sweet spot seat, virtually no real problem related to dispersion that isn't trivially repaired by speaker placement and DSP EQ.
For the record, I use a 2x3 array (6 cell) of flat cells in a slightly cylindrical panel.
Dayton-Wright had a 2-high x 4-wide array inside each speaker on the surface of a sphere! His theory was that it would be virtually like a point source (meaning the point at the origin of the sphere). Think about that.
Ben
Last edited:
I'm not saying Wright's theory is correct.
His panels were in a one-meter-square box and I'd guess the origin of the spheric shape was maybe 10 feet back.
As for flat panels, yes there is beaming. But if you are sitting in the beam, what's the problem?
Geddes presents a good rationale about dispersion angle and tweeter horns. Yes, dispersion matters (but can be fixed as I said previously). It matters most when, like Geddes, you are manufacturing a speaker that has to work even in the homes of uninformed users.
B.
Honestly I don't experience any more than with my Kappa 9s. Transition is more pronounced maybe, but so what? My Kef 105s have the smoothest most gradual transition but again, when it's gone, it's gone. I will stand by my position that sweet spots are always narrow as I define the term, to lose nothing.
I don't have a clue what you mean by any of those sentences. Perhaps the fault is mine.
B.
Last edited:
I was referring to the beaming and the sharper transition from the sweet spot.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.