As I have previously explained, diffraction from the horizontal baffle edges, i.e., the top and bottom edges, in theory creates on-axis response peaks and dips at specific frequencies; these predicted peaks and dips and readily apparent in Amir's measurement but are not apparent in Rick's measurement.
to me to point out that the major discrepancy between the two measurements are explained thusly, that one (Amir's) shows the effect of diffraction associated with the top and bottom edges of the baffle, while the other one (Rick's) does not. Of course this raises the follow-up question of why Rick's measurement does not show this diffraction effect.
Rick can't eliminate the influence of the speaker cabinet edges (edge diffraction) in his gated measurement - how should this be possible? And I would also rule out the possibility that Rick sets too large a gate and thus allows room reflections to creep into the measurement.
But conjecture is not circumstantial evidence, so there is more circumstantial evidence in the following sections.
When calculating the theoretical frequency ranges in which the effects of edge diffraction are to be expected with the simplest means, one must be careful to estimate their influence correctly.
Therefore I have shown in
Post#694 the BEM simulation for the woofer in the loudspeaker cabinet and the large deviations between BEM simulation and Amir's NFS measurement that occur when normalizing to the axial frequency response. Maybe this was too abstract and complicated to understand.
Therefore here is the simple baffle simulation of the Purifi woofer on-axis in the Directiva enclosure that VCAD provides (the influence of the cabinet depth is not taken into account here):
VCAD now offers the possibility to include a half space measurement (infinite baffle measurement) and to simulate it based on the baffle simulation in full space (4pi).
For the infinite baffle measurement from HifiCompass of the Purifi woofer we get this:
In this way, I took the on-axis half space measurements of the woofer from HifiCompass and the manufacturer and simulated them in the Directiva enclosure for full space, exported them, and compared them to Amir's and Rick's measurements:
Also the VCAD simulation, like the BEM simulation, ignores the protruding surround of the woofer.
Shadowing theory
Now you have to realize that the theoretical maximum around 860Hz for the vertical edge diffraction corresponds to a wavelength of 40cm. The surround protrudes about 1cm. So everybody can imagine how big the influence of the shadowing should be. Nevertheless, let's assume that the shadowing has a significant effect on the vertical edge diffraction.
1) If in Amir's NFS measurement the protruding surround was not in the way vertically, then the NFS measurement should agree well with the simulation.
2) If in Amir's NFS measurement the protruding surround was vertically in the way, then the maximum at 860Hz is attenuated (because of shadowing) and frequency curve should be below the simulation in this frequency range.
No matter if 1 or 2 in both cases the NFS measurement follows the course of the simulation less well than Rick's measurement does.
Which again is a confirmation that Rick's measurement seems to be okay and that the NFS measurement does show unexplainable deviations compared to the simulation (apart from far-fetched theories like protruding measurement platform edges).