That 91 dB in-room figure is derived by factoring in the additive effect of reverberant sound within the room, and no doubt varies from one room to another.
Also, the reverberant field wouldn't add significantly to perceived loudness at near-field distances, but at normal listening distances the reverberant field does raise the volume a bit. So while at 1 meter the speaker would probably measure 88 dB efficient, back at 4 meters the in-room reading would probably be about 79 dB instead of the anechoic-predicted 76 dB (SPL falls off by 6 dB for each doubling of distance from a point source under anechoic conditions). I've taken in-room measurements of various loudspeakers at different distances and gotten similar results.
I have always assumed the anechoic figure to be the convention, but there really are no firmly established standards. We should give Paradigm due credit for their full disclosure in quoting both figures.
For the sake of comparing apples with apples, I'd say consider the Paradigms to be 88 dB efficient speakers.
Duke
