One more thing, here is an equally priced alternative to the Paradigm speakers:
JBL LSR6332, $1,560 each. If you can stand their, well, 'plain' appearance, you get one of the most accurate loudspeakers on the planet.
Here is what I am talking about: Frequency response of the 95F, from
this review:
This is a bit rocky, to put it mildly. You would expect something less erratic for that price. Now here is the Frequency response for the LSR6332:
That is a thing of beauty. One of the most linear speakers out there, and it has tremendous dynamic range too. The only caveat is that it is not a terrible stylish speaker, and it is intended for studios and sound stages, not really for home use. It is a true high-fidelity speaker. It is THX PM3 certified meaning it is certified to
create THX soundtracks on, not just playback! These have a good enough response to be used in a THX mixing stage.
Harman actually used LSR6332s in their reference listening room. By the way, Harman was widely regarded to have the best research and development group for loudspeakers in the world. Here is a pic of the LSR6332s in their reference room:
Anyway, here is some good news about the 95F speakers:
You can see that the horizontal dispersion is very uniform, which is good. However, and what can be easy to forget in John Atkonson's measurements, is that the direct axis is zeroed out in this graph, and what that means is all of that rockiness in the frequency response graph carries over at all angles of the speaker's response. Sine the dispersion is very uniform, this is the sort of behavior is what auto-EQ systems like Audyssey can actually help with. However, they are just not needed with a speaker like the LSR6332. There is very little that EQing can do to assist a speaker like that; it is that good.
The bottom line is that the 95Fs look nice, but their frequency response, not so much. The LSR6332 do not look so hot, but their performance measurements are drop dead gorgeous.