You still want us to go easy on ya? There's gobs and gobs of
theories on what constitutes good EQ practices - most of em seem to be
incomplete or make some bad assumptions, which often leads to a lot of
confusion. With that in mind, I'll just share what I've come to
understand, knowing that there is still more to learn. You might want
to grab a cup of coffee 
There are a few things happening and
I think it's best to look at each one individually. In the anechoic
chamber, we see exactly what the speaker does by itself. In the
listening room, you have the speaker-to-room interaction, which can be
divided into the direct sound and the reflected sound. The important
thing here is the direct sound is actually anechoic until the first
reflection arrives at the listening position. If the reflections takes
long enough to arrive, then we percieve it as a distinct and separate
reflection (echo). If it arrives early, our ears don't pick up on the
fact that it's a reflection so we percieve it as part of the direct
sound. It of course doesn't sound like the direct sound though....it's
a smeared version of the direct sound.
An RTA doesn't make any distinction between early and late
arriving reflections. In fact, you could think of an RTA as lumping all
of the time-smear into a single measurement as if everything arrived at
the same time. Normally this would look very chaotic, but that's why
they add a 1/3 octave smoothing to them (to help correlate the visual
representation to correlate to about what we would hear).
The
problem with EQ is that it doesn't change the fact that the sound is
still bouncing off the walls in your room. In other words, it doesn't
fix the time-smear, which is where the increase in clarity needs to
happen. You will, however, see dips and peaks in the frequency response
that are caused by the reflections.
The other problem with EQ
is that you need to match the Q of the problem, which is rather
ambiguous when using an RTA. Just to illustrate that fact, here's an
example of a raw measurement with no smoothing:
Here's the same measurement with 1/3 octave smoothing:
And there here's the measurement with 1/3 octave smoothing and EQ applied:
(doesn't that look a lot flatter there in the middle?) 
But here's the measurement without smoothing and EQ applied:
Notice how the EQ at marker 2 looks better, but 1, 3, and 4
aren't? Ultimately this can come down to the fact that 1,3,4 are not
minimum phase, but 2 is. The reason the EQ at 2 works is because that
is part of the raw response of the driver. The craziness at 1,3,4 is
due to a reflection with the floor. Normally you see a lot more
craziness in the high frequencies, but I was using some foam during
this measurement to reduce the effects of reflections in the room. In
other words, I tried to minimize the effects of reflections to provide
a simpler example! It only gets worse than this. If we looked at the
response in an anechoic chamber (I don't have one so can't do an actual
measurement), we wouldn't seen any peaks or dips at 1,3,4, but we would
still see the peak at 2.
Since 1,3,4 are the result of
reflections in the room, moving the mic as little as an inch is going
to dramatically change their location and magnitude. I did not perform
any measurements to demonstrate this so you'll have to take my word on
it. When I get back to campus I can show some unsmoothed plots where
the microphone moves if anyone is interested. Anyways, my point in
showing these pictures is to illustrate that the frequency response is
crazy insane....when measured in a room.
In the anechoic chamber
things look a lot more like the smoothed plot, except that no smoothing
is applied. The most obvious EQ being used on the Jubilee is the shelf
filter bring up the falling response of the tweeter. Because the K402
is a constant-directivity horn, the HF output is going to drop
6dB/octave...this is actually expected behavior and in a way a good
thing, but got a bad wrap during its early inception because early CD
horns were using diffraction slots (yuck) and they weren't very good at
providing a constant directivity either (so no real gains). Without the
"CD EQ" (compensation for the falling response of
constant-directivity), you probably had the tweeter turned up louder so
that you could hear the highs, which in turn would make the mids too
loud (thus sounding "forward").
If you go a step further, one
could think of using EQ to correct the power response of the
speaker...then, if you put the speaker in the room, you can guarantee
that the direct sound and reflected sound share the same tonal balance.
Without a flat power response, the reflected sound is going to have
different frequency content than the direct sound. If you put a flat
power response speaker in your room, and then treated all the early
reflections so that every reflection at the listening position was
outside the Haas window, then you don't end up with any time-smearing
and all of the reflections sound as if they are supposed to be there.
In fact, listening in an anechoic chamber doesn't sound natural at all
(too dead)....we need the reverberation of the room, but we need to
make sure the reverberation doesn't reduce intelligibility - which is
achieved by getting rid of early reflections.
The other consideration though is that homes are too small to
achieve true reverberation. In other words, there aren't enough
surfaces and the room isn't large enough to create a true reverberant
field. Because of this, reflections should only be absorbed when
absolutely necessary. In other words, they should be diffracted so as
to breakup specular reflections into multiple smaller reflections
spread out over time. Here's an article that provides some pictures of
an ETC with good labels:
http://www.rpginc.com/cgi-bin/byteserver.pl/news/library/HT_AcD.pdf
And just in case you're interested, here's the ETC for my measurements above (notice the floor bounce at 6.88ms):
The lack of reverberant field in my measurement is mostly the result
of my putting acoustic absorbtion behind the microphone because this
was part of a series of measurements of dialing in my Chorus II's with
an active crossover and I wanted to get as anechoic as possible.
Ok, I dunno if you wanted to know any of this crap from above, but
there it is for you to spend as much time on as you want, lol. To
answer some of your questions. I always use EQ for the anechoic
response of the speaker. Any frequency response abberations resulting
from the speaker-to-room interaction should be addressed with
acoustical treatment because EQ doesn't fix the time-smear that results
from early reflections. I do not EQ my source material either (or
change EQ based on source material). The nice thing about the Jubs is
they have extremely controlled polars, in fact, so controlled that you
wouldn't need to address early reflections for the right speaker on the
right wall. In other words, the first early reflection from the right
speaker is going to be on the left wall...(vice versa for the left
speaker). I believe Roy has mentioned that the Jubs keep the sound off
the floor (which means its gonna be kept off the ceiling too). The
easiest way to tell would be to just measure the room. If you're
interested in measuring, I've started putting together a website where
I hope to detail everything I've been learning (and correcting my own
misconceptions as I learn about them too):
http://measuresound.googlepages.com/
I mention it because I have only spent $100 on my measurement rig and it can do some rather powerful things.
I don't mean to drop a bomb of a post like this, but you did
ask...and it's winter break so I've got plenty of time on my hands