Thank you for showing them here, what I would be more curious though are the vertical data as due to the large mid driver they should be quite hard to optimise, would enjoy also hearing your experiences about that task.
This is a good question but I think that it's also important to understand the audibility/significance of vertical polar response errors.
There are actually some good threads here on ASR about this (the important or lack thereof of the vertical response of the a speaker in a home listening room). Certainly in professional sound reinforcement vertical directivity is super important and guys like Don Keele and line array folks have made great designs for this.
In his CBT paper (an arrangement with constant vertical directivity), you can see that on page 94, even a well designed tower speaker like the revel salon has very inconstant overall response (including boundary interactions with the floor etc).
https://www.xlrtechs.com/dbkeele.co...(AES-ASA-BAS presentation, Jan. 14, 2010).pdf
Thankfully, we have ears on the sides of our hears and the sound quality of speakers in a home environment are primarily driver by the horizontal directivity, otherwise we would see different baffle layouts or coaxial designs be favored a lot more. Yes, you want try to have floor and ceiling early reflections be as balanced as you can but this is one of the cases where the Nelson Pass saying of "ears are not microphones" could likely be used, though there doesn't seem to be a ton of scientific study (that I've seen at least) on this particular aspect of performance.
As far as directivity goes, as sound wavelengths approach the size of the cone or in this case the planar membrane, they start becoming directional because of phase cancellation. There is a dimensionless measure of this called “ka” - circumference of the cone divided by the wavelength of sound it producing. Relatively smooth off-axis response is maintained until ka=2. In the case of our mid, the effective circumference of the vertical length would be about ~62 cm, so ka=2 is just over 1 kHz. A 6.5" woofer puts this at about 1.65 kHz.
Our planar has the benefit of no breakup or resonances in this area, which is normally a big issue in the useful upper midrange bandwidth of cone drivers but they do start to narrow vertically sooner. I put the tweeter crossover as low as I could to get the best balance of power handling, tonal balance and distortion etc. and this involved a lot of measurements and listening etc. We did target the listening window and estimated in-room curves to be quite smooth, generally, and I think that obsessing about very small details of these may or may not correlate with better or worse overall because I think that a lot of the nested design decisions to get there (like the use of very small midranges or coaxials) have their own sound quality implications.
Depending on how the speakers are set up and listening axis, if a customer feels like there isn't enough energy in the 3 kHz range, the speaker could be raked vertically with the spikes a bit but I prefer the sound as-is.