I've been summoned! Unfortunately, we don't have NFS results for the Salon2 yet, so I can't do the most accurate comparison, but
Soundstage Network's anechoic results should suffice as their directivity measurements have correlated closely with Amir's before. 60 degrees horizontal is the widest overlapping angle (the NRC uses 15-degree increments), plus it's a typical angle for sidewall reflections, so it's what I'm using for this comparison. Keep in mind the NRC also uses a different listening window calculation (+/- 15 degrees for both horizontal and vertical), but that shouldn't affect these results much.
Below I've plotted the Listening Window, 60 degree measurement, and a "60-degrees directivity index" for each speaker. (Ignore the tall spikes at the ends of the graph, that's a small glitch)
View attachment 151028
It seems
@echopraxia's ears were not lying =] The Salon2 is, as expected, much wider than the Gelenec in the range where it counts. The F328Be and the Salon2 are actually very similar throughout most of the important range, from roughly 1-10kHz, but the Salon2 ultimately ends up being a little wider due to the bit of a directivity mismatch at 1kHz with the F328Be. What goes on beyond 9-10kHz is unlikely to be of major importance for spatial presentation, but we see the Salon2 does drop off highly in this region.
That said, this is just one angle. I've noticed that sometimes speakes have a quicker or slower 'directivity roll-off' at further angles. If If I estimate the F328Be's 75-degree angle in VituixCAD by interpolating the 70 and 80 degree results, and compare it against the Salon2's measured 75 degree angle as actually measured by SSN, the Salon2 appears to maintain its wide directivity a little more (not to mention the 1kHz problem remains in the F328Be):
View attachment 151029
Although the F328Be does appear to be a little flatter than the Salon2 and it has better controlled vertical directivity per the spinorama, the Salon2 is still the slightly wider speaker and appears to have better controlled horizontal directivity overall. Indeed, I'm quite sure the Salon2 is still the widest directivity speaker that also has directivity this well-controlled that I've seen anechoic data for (and I've messed with the data for a LOT of speakers). I think the spinorama graph alone significantly underestimates the Salon2's performance.
Couple that with better bass output (per Harman's spins, the NRC measurements arent reliable below 100Hz), and IMO the Salon2 still reigns as the best of the Harman speakers in terms of my interpretations of the measurements.