My bet is that they are phased normal, firing toward each other: If the midbass is a resistively damped cardioid (much different than simple aperiodic. See Liknkwitz's pages - extremely calculated), along with a waveguide tweeter and Dipole woofer there would essentially no rearward radiation, which enables placement against the front wall . . . which matches what is going on in the photo!
The bass nulls in the listener zone would be pretty narrow, and only experienced when sitting directly on axis of either speaker, well off the sweet spot. as you move off-center into the null of one speaker, you are also moving closer to the sidewall/front corner --which is on the axis of strongest radiation-- AND are moving slightly more on-axis to the farther speaker. The nulls could be very hard to hear in this configuration!
The catch is that the rest of the speaker, from a few hundred Hz on up better be directional to keep from sounding colored & smeared like a regular box speaker against the front wall.
I've experimented with U-frame woofers along the front wall, firing toward each other. It does 'work' as long as you are sitting not too far off the sweet spot. The biggest downside is that it puts the bass well 'behind' the rest of my dipole system - 'starts to sound slow like the familiar corner box sub.
Gradient may have solved that problem here, at the expense of the dipole "sound"/reflections that some listeners prefer.
Has anybody seen literature or talked with the vendor?
If they recommend placement close to the front wall, and specify a range from the sidewall, I'm on track. I can't see how this arrangement would work very well if placed 5 feet from the front wall . . . or in a room that is unusually wide.
I could be totally wrong here . . . If so, at least we have some new ideas to try!
-- Tubamark