'Beaming' is, I think, a function of the way the top (in particular) works at higher frequencies. In the low range the lower bout of the top pretty much works like a loud speaker cone, moving in and out as a unit: a 'monopole'. This starts to break up somewhere above the pitch of the open G string (usually) when the top 'cross dipole' mode gets going. The bridge is rocking from left to right around the center line. Half of the top area will be in phase with the monopole (which is still working, even if it's not at as high an amplitude) and half out of phase, so the sound starts to be projected in a different direction. This shifts suddenly as you actually pass the dipole resonant pitch, and the motion of the dipole goes from being in phase with the string to out of phase, while the monopole motion stays the same. This shows up quite clearly if you measure the sound output of the guitar in different directions as you go up in frequency.
As you go higher in pitch the top breaks up into more and more smaller areas, each of which is out of phase with it's neighbors. What you hear coming off the top is a sort of sum of the sound coming off all the different areas, and it adds up differently depending on which direction you are looking at. You can, of course, think of the sound hole as just another top area. The top with the hole is a 'phased array'. Close up a lot of the sound simply cancels out, but as you go farther away one or another direction predominates, and the sound is 'beamed' out.
I remember one Classical 'shoot out' at a GAL convention that was held in a class room at the college. Folks in the back of the room said that one guitar was a 'cannon'. I was near the front of the audience and thought it was nothing special. The player said he could barely hear it, and Dana Bourgeois, who was running the test and stood off the player's right elbow, said he could not hear it at all.
FWIW, I have noticed that it's often the 'better' guitars that do this, although seldom to such a degree. In a hall that is properly set up for music (the classroom was, of course, designed for speech) there is enough mixing to give all of the audience a good sound. As I mentioned before, a well made and 'balanced' top can still project well at high frequencies, even if it was designed and built to accentuate the lows. It's hard to do that, though: usually it seems to be the thicker tops ,or, at least, those that are very stiff for the weight, that do it.