Here's a post I made highlighting some of the problems that exist:

http://www.zerogain.com/forum/showpost.php?p=177591&postcount=78

I don't manufacturer speakers at all just an enthusiast who builds his own stuff for pleasure:

You'll see that they use ATC drivers for mid and bass. They use FIR filters on the crossover along with phase/time alignment. I cross the mid at 3.2Khz and 400hz using 96dB/oct slopes. ATC use 24dB/oct at 380hz and 3.5Khz which is wholey inadequate at those frequencies since there's a breakup a 4.5Khz and drivers resonance is 320hz(lots of distortion there and big impedance phase problems). I've done a lot of experimenting and the ATC mid sounds harsh when following ATC's footstep and considering the measurements its easy to see why.

I've also got a +/- 0.5dB in-room response from ~180hz to 20Khz+, in that same range the ATC implementation is more like +/-3dB in-room if your very lucky.

Why didn't ATC do a better job? The ATC SB75-234SC doesn't run up very high and they couldn't push the bass/mid crossover point to 600hz which is what you really should be doing with 24dB/oct slopes on the ATC mid. Kinda defeats the object of running a pure midrange when you raise the XO point slap bang in the middle of an important frequency range.

The FR is largely uncorrected because they're using analogue active electronics. Its completely impractical to offer the flexibility to perfect in-room response. The same goes for the phase/time alignment.

God only knows why they didn't drop the XO point of the mid/treble. Maybe someone likes the lower treble a little confusion - interestingly these very findings are backed up in the SCM70 review:

http://ultimateavmag.com/speakersystems/5