It's about how to design a crossover for drivers in a box which has some residual diffraction which can't be eliminated. Specifically, what method of measurement and design is best when there is diffraction if you measure the drivers in the box.
After diffraction is eliminated mechanically as much as possible without causing new significant problems:
1) Measure drivers in final box or prototype. Hor and ver planes (-180..)0..+180 deg with 10 deg angle step, dual channel gear and mode, rotation center in the center of radiating area on baffle surface or mouth etc... Measure impedance responses. See measurement preparations for simulation with VituixCAD for more information. Load to XO simulator and create initial XO.
2) Initial crossover points to frequency or a bit above where next upper frequency radiator has clear dip in DI-response.
3) Accurate phase match to range where DI-response of upper frequency radiator has clear dip. Shallow HP slope for upper frequency radiator or possible phase mismatch to range where directivity of lower frequency radiator is excessive.
4) Set listening window average to +/-20 deg hor and +/-5 deg ver. Optimize crossover by both listening window average as axial response and power response. Weight axial/power 40%/60% if room is reflecting and/or typical listening distance is long (>2.8 m). Weight axial/power 60%/40% if listening distance is short (<2.2 m) or room acoustics is very damped. 50%/50% if cannot decide.
Flat horizontal target for axial. Tilting straight target for power, about -1 dB/oct from 200 to 12k (depending on directivity concept).
This is normal XO design approach whether speaker has separate diffraction issue or not because directivity spectrum of different radiators is not flat and equal compared to each other. Proper design approach is to use crossover as one directivity component always in multi-way as much as possible without causing some other significant problem.