The other night in bed came to think about a slope variant called eleptic, and after a check today think it can look M105/M106/M16 use some variant of this filter type, there is signs below for that into impedance curve and directivity index (DI) and it looks woofer is bleeding that on axis spike around 5kHz and cause DI take a blib for M105 and M106, plots below from upper left is M105 / eleptic 3-way example / M106 / M16..
Took me together set dimensions on baffle and model baffle step loss/diffraction for M105, suggest
@5th element and
@GelbeMusik should lookup below graphs of printed data before their next discussions of M105 is good enough build or not
not that i have this little beauty physical but its possible use a ruler on a flat panel because Amir is so kind take his picture from front without much distortion in angle, printed XO slope is not any high precision calculation but probably thereabout, its created from Amir's nearfield measurement times baffle step loss/diffraction and in CAD software one is told for a 107mm piston max microphone distance is 5,9mm and max frequency is 1023Hz, for 25mm piston numbers are max 1,4mm mic distance and max frequency is 4380Hz but good enough it looks circa a 6th order roll off slope and practical view overlaid to DI and impedance..
Below plot was for fun and own curiousity and maybe some can enjoy it too but notice its not the 100% reality because Amir's nearfield measurement is not a precision of slope for woofer and tweeter, exercise was try calgulate a inverting curve for sum (not printed) of the two passbands red/green minus orange anechoic on axis then we get the port curve and any other interference that when inverted and summed to red/green curves output the orange curve, the lower red curve is DI and the difference to the black one symbolize two summing omni transducers spaced 143mm apart using below particular slopes so had M105 had no vertical spacing as a coxial DI had looked like the black curve. In exercise was using linear phase for all curves, and woofer plus tweeter passbands was in same polarity, in reality it can be tweeter is flipped polarity and acoustic center is offset relative to woofer, also using real world minimum phase will change a bit on how for example port curve looks but in Amir's spindata is linear phase used that default.