@johnk --
"The Jubilee uses DSP since the Celestion Axi2050 doesn't have much upper range unless EQed and processed. I use the same driver as Jubilee in a better non-CD horn design and with a proper tweeter set up so I need no DSP and can just let the Axi2050 roll-off gently mating it to a proper tweeter. The sound quality once done is one of the best in compression drivers."
What's your general issue with Constant Directivity horns - the frequent use of diffraction slots here, abrupt mouth termination (straight-sided horns), falling HF response, other? And what (older?) profile are you referring to as "a better non-CD horn design"?
Certainly a CD horn/driver combo can be in need of some EQ to ameliorate falling HF response, which would be less ideally achieved with a passive cross-over due to the likelihood of filter complexity. Active XO's/DSP's are the better choice here, but such a configuration is not your cup of tea, I know. Above a certain HF frequency the horn becomes more irrelevant, and here the driver itself will be a co-determining factor with regard to upper extension and the roll-off rate here. In that regard the axi2050 doesn't help.
In my own actively configured CD horn/driver combo I use no shelf filter to attenuate HF response, but rather a few filter notches, a peak suppression, and a mild wide-Q PEQ gain at close to 20kHz. Filter notches and peak suppression were eventually derived from my own measurements and, at an early juncture, more crudely from the manufacturer's supplied measurements of the same horn/driver combo I use, to which there was a rather close correspondence with my own measurements. Determining the exact frequencies for these mild filter corrections was essential for them to have the most sufficient effect; indeed, what quickly stood out from that outset was that "less is more," and that achieving the most smooth and natural sound with these corrections was the primary concern.
"Flat" HF response, more easily achieved with non-CD horns, sounds unnaturally tipped-up to my ears, and moreover the increasing directivity here can lead to HF-beaming and more prevalent issues with finding the correct placement. What may more readily persist with CD horns is diffraction issues, but theoretically at least those emanating from the diffraction slot area of my own CD horns (what some perceive as the "frying bacon" phenomenon) are not a concern until well beyond 120dB's - if they are at all.
It comes down to actual listening, and I've never been able to provoke any issues with my speaker setup here, although I've never reached +120dB's - certainly never felt inclined to. Question is whether diffraction issues would present themselves sonically at much lower volumes, and how that would imprint as a "signature" of the sound. Comparing with quite a few other horns and waveguides I've heard - i.e.: a variety of non-CD horns, OSWG's and other - I can only surmise that it's a non-issue at any SPL's I've cared to try out. If anything the sheer size of a large horn like my EV's is an advantage.
Begging the question: when is a problem a problem, and how does it apply to a given context. I guess it all depends.