There seems to be a hobbyist community around using TAD drivers with wooden bi-radial horns...
They sure look amazing.
But are there any technical merits / de-merits to these kind of horns versus other shapes?
Eyeballing those TAD horns, my guess is that their radiation pattern width is a full 90 degrees in the horizontal plane over most if not all of their spectrum, with the little "vanes" helping out in this regard at high frequencies. In the vertical plane I'd guess that the radiation pattern is considerably wider at the bottom end of the spectrum than at the top end. The part of the horn deep inside, back near the throat, is what controls the vertical pattern at high frequencies, and it looks to me like the top and bottom of the horn are virtually parallel back there, implying a tight vertical pattern. Then as we go down in frequency (as the wavelengths become longer), the pattern starts to be controlled more by the mouth-end of the horn, so it widens, at least in the vertical. But I could be wrong; there could be other things going on which I haven't taken into account.
As for potential sources of coloration, I like the big round-overs of the "lips" but am not a fan of the relatively sharp edges at the "corners" of the mouth. I'm not sure what the coloration implications of the "vanes" are; in general reflections within a horn are undesirable in my opinion.
But they sure do look awesome!!
Speaking of which....
Is there a primer somewhere on horn geometry?
The
Wikipedia article looks pretty good to me.
On another forum Jack Bouska examined some of the qualitative issues around horn design. These two posts are imo the most instructional. Note that the abbreviation "CD" herein stands for "constant directivity", not "compression driver" or "compact disc":
http://www.audioheritage.org/vbulle...amp-Waveguides&p=132741&viewfull=1#post132741
http://www.audioheritage.org/vbulle...amp-Waveguides&p=132743&viewfull=1#post132743
Getting back to the TAD horns for a minute, they are designed for "old school" compression drivers which have a long internal throat with a relatively narrow flare angle. Such compression drivers generally do not work well on more modern waveguide-style horn shapes, such as the Quadratic Throat and Oblate Spheroid, nor on the more recent generation of JBL waveguide-style horns (which came along after Bouska's posts). These more modern, low-coloration, constant-directivity waveguide-style horns work best with modern compression drivers having a much shorter internal throat. Imo ideally the exit angle of the compression driver matches the entry angle of the horn, to avoid a diffractive discontinuity where they meet.
This website has useful polar maps of some of their horns:
http://horns-diy.pl/. Let's look at a couple of radiation patterns.
The
SEOS 12 (which was used in one of my thumbnailed designs in Post #350) is an approximately constant-directivity design. You can see that in the (narrower) vertical plane it starts losing pattern control around 2 kHz, whereas in the (wider) horizontal plane its pattern control is good for another octave or so. The pattern widening in the vertical plane is because the rectangular horn is physically too small in that dimension to maintain good pattern control below 2 kHz or so. However it's close enough that the off-axis sound has approximately the same spectral balance as the on-axis sound, out to the pattern's -6 dB limits (90 degrees horizontal, and ballpark 45 degrees vertical over most of the spectrum). When using this type of horn, I like to dip the on-axis response just a little bit south of 2 kHz to offset that excess vertical off-axis energy.
The
J. M. LeCleac'h horn is clearly not a constant-directivity design; as you can see, the pattern becomes increasingly narrow as we go up in frequency. This simplifies crossover design AND results in considerably higher on-axis efficiency, as the top octave is usually the limiting factor, so by concentrating the energy into a narrow angle the sound PRESSURE level is higher within that narrow pattern. (Think of a variable nozzle garden hose: The same amount of water comes out regardless of the pattern width, but the PRESSURE where it hits is higher with the narrow pattern.) The downside is, that the off-axis sound has a significantly different spectral balance from the on-axis sound. So we are unlikely to see JBL using LeCleac'h horns any time soon. Note that JBL went to great lengths in the M2 waveguide to get an exceptionally wide radiation pattern (especially for a 1.5" throat compression driver) all the way up the spectrum, but that the broadband efficiency of the M2 is not particularly high for a large horn loudspeaker (92 dB).
The closest thing to a TRUE constant-directivity device on that website is probably the 15" Oblate Spheroid, but no measurements are shown on its page.
So if you want to use a flea-powered tube amp, you probably either want a horn which concentrates the top-octave energy into a narrow angle, or you want an additional small horn (tweeter or super-tweeter) to handle the very top end. If you don't mind having more modest efficiency, then you can use a constant-directivity horn to cover the entire upper portion of the spectrum. Typically 1" throat compression drivers have more top-octave energy than large-format (1.4", 1,5" and 2") compression drivers, so it is usually more cost-effective to run the midwoofer up a bit higher and cross over to a 1" throat compression driver, the
GedLee Summa being an arguably advanced example of such (which to the best of my knowledge was never tested by Harman). But there are also arguments for going a bit lower and covering everything from ballpark 800 Hz on up with a single driver, hence designs like the JBL M2 and the beautiful DIY TADs in your photos.
Sorry for the rambling post, I'm winging it.