really? would you mind explaining to me why? in my experience if i didnt point them at the listening position i never knew they were on.
We do use HF to localize, because the shorter wavelengths create larger phase differences between our ears due to our head (or HRTF to be exact).
But that does not mean we need to hear all the HF. Our highest sensitivity is midband, somewhere up to around 3k, where it begins to fall off (towards 20k upper limit, lower for some). By pointing your speaker away, you may reduce the very HF, but not the 1-3k range, so you can still localize it quite well. What you
have done, it possibly create more short delay reflections, which will diffuse some of the localization.
horns do not beam as much as unloaded tweeters, they have a much wider dispersion area.
Incorrect. One of the functions of the horn is to
increase directivity, to maintain pattern control where an non-horn would have none, i.e. wider dispersion. At 2-3k, the typical 1" dome on a baffle has 180 deg dispersion. In a (large enough) horn, that will be reduced by the horn angle. Where a horn
can increase dispersion is at higher frequency,
if it diffracts the wavefront produced by the dome/diaphragm. Otherwise, the beaming is simply dictated by the diameter of the transducer, just like without the horn.
With any setup, some adjustment is needed. So you tweak, listen for a while, tweak, listen, etc... until you have what sounds good to you.
For those who cannot understand theory or measure, that is the
only option. Trial and error. Or at least error....until you have convinced yourself that things are "properly" set up

.
Hardly repeatable or consistent, that "trial and error".
Surrounds are supposed to be more diffused in general, so they should not be pointed directly at the listener.
That is in contradiction to your claimed requirements for MCH music, where one is supposed to
clearly hear instruments, like synthesizers, horns and singing (Roxy music, etc.). Why would you want those to be diffuse when they were clearly not mixed to be heard that way?
Why would you use a monopole for diffusion/spaciousness when a multi-polar is superior? Oh wait...because j_garcia prefers it that way...sometimes...I guess

.
With bipoles, the listener is supposed to sit in the null area of the speaker so you hear the reflected sounds.
Wrong.
There is no null with a bipole, only a dipole.
You clearly have no idea what you are talking about, so rather than continue...
A PDH eh???