The long and short of it is that bracing the cabinet is stiffening, obviously. What comes from that is the natural frequency goes up and the displacement under a given magnitude of loading should be reduced. The lower amplitude movement translates to lower movement of air and thereby lower SPL, in general. The only way I can see this NOT happening is if the braces aren’t cross coupled and simply adding mass without stiffness. If you brace a cabinet panel to the point where its resonance spectrum starts somewhere above & beyond the HF limit of a drivers bandwidth, then the panel will not resonate. It won't resonate because given the above stated conditions, the mechanical system isn't being fed any energy within the frequency range needed to trigger mechanical resonance. Basic physics.
What the Axiom study does not make clear is the geometry of the cabinet or the braces. How stiff are the braces in an absolute sense and relative to the side wall, how well are they attached to the side wall, are they uniform in size and/or spacing? In other words, it is not clear if the braces stiff enough and well enough coupled to the side walls to make a significant difference to the actual modal behavior. It might be possible with a uniform or nonuniform arrangement of bracing that makes one or more of the subpanels match resonant frequency with another panel, say with the top/bottom panel, and that might increase output.
It is also not clear, but likely, that a single point or a small number of points are measured. The mode shapes of the panel are likely to be fairly complex and increasingly complex with the addition of the bracing. Limited information such as this would not describe the overall effect of the entire panel with and without bracing where complex response would produce canceling effects. Although it would be a poor representation of the response, the best place to measure to approximate the response would be the centerpoint of the panel/subpanel, the likeliest location to capture maxima/minima. If the same point is measured in both cases the change in relative proximity to the boundaries and braces would give results that are not representative of the behavior.
Also note, the results are reported as accelerations, and although differentially related, SPL is a function displacement, so the plot could be viewed as misleading. Regardless, all of the measurements are reported as below their established threshold of hearing which we also have to wonder how they actually derived that in itself.
Some more questions that come to mind:
• First, did they figure masking into the curve they applied in the graphs as the upper boundary? Second, in comparing accelerometer data to an overlaid dB spl plot, isn't that something akin to comparing apples & oranges? (That's why you need to separate accelerometer data and the dB spl data)
• Last, why no mention of the contribution to panel vibration made by the driver bolted into said panel? Owing to the far closer mechanical impedance match between the driver frame/panel, it’s a far more efficient mechanism for transferring energy than the air/panel combination.
These days, measuring the acoustical behavior of a cabinet panel is comparatively easy, using something like the B&K 3599 microphone rig. From the sound intensity data generated, the dB spl plots of the panel can be derived.
