The curves in the T10 document you referenced are for the Model T active; not the T10.
I'll say it again. The Model T active LW curve is horizontal. The LFR1100 active is downward sloping at 3dB/decade. Why is that?
Page 3 has the Model T curves, but page 11 has a curve that I'm assuming is one of the T10's. It would be nice if they labeled it. Looking closer, it seems like it could be +/- 0.75 dB from 60 Hz to 15,000 Hz. LW and SP are a single composite curves from dozens to hundreds of measurements, so that's some serious engineering or software wizardry.
The LFR 1100 Active listening window slopes because it has to, in order for an omnidirectional speaker to sound good.
https://youtu.be/sL-W3qn4-WU?list=LL&t=1111(time cued)
A bipolar or line array design might have different listening window slopes from a point source design. I haven't seen anyone talk about it. In the end, the listening tests determine which measurements matter.
