You are falling into the common trap of assuming that waves follow the walls of a horn simply because you want them to. They don't. For instance at frequencies where the wavelength is comparable to the horn mouth the directivity is substantially smaller than the wall angles. It is only when the mouth is much larger than a wavelength that the device even begins to get CD. Thus for your mid horn to be CD it would have to be gigantic - about 8 ft. across at 250 Hz. Then this places the HF horn many wavelengths away - unless they are coaxial, which has a whole other set of problems - and there will be a great deal of polar lobing at 1.6 kHz - not CD. SO from where I sit this design is anything but CD.

Unless you have data to support your contentions I will have to stick with my opinion that your hypothesis is unlikely.

Mine claims, on the other hand, have substantial supporting measurements which I freely show on my web site.

Sorry to disagree with you like this, but we don't see this issue eye to eye.