Looking at your un-eq'd curves, I'd say it's a toss-up between your settings with the sub LP at 120Hz and speaker HP at 100Hz, and the one without any HP filtering for the speakers (but you get phase cancellation at 105Hz or so). You might be able to mitigate that somewhat by using a sharp notch filter for the sub at the frequency, maybe, perhaps -- wouldn't hurt to try.
Given your huge room-related cancellation issue around 80Hz, it really is a compromise, either way.
Anyhows, since you are using REW...
If you look at the phase traces (using an acoustic timing reference), you will be able to see where the sub and speakers’ phases converge and diverge -- given your fixed set of delays and filters.
I need help from those with experience from two+ subs though, to figure out if I can keep my DAC in the chain or not and if it is realistic to hope for a major improvement with two subs.
I'm not using multiple subs... but from what I understand, one may want to adjust the relationship of their phases to be somewhat different so that they even-out in their overall summed response. There will be some energy loss, but that trade off is well worth-it.
It's not possible to do this finely with your current interface, though... the Genelec subs do have simple phase dip switches. Maybe with the GLM software, there is also the ability to adjust timing automatically/manually as well? I don't know.
...
EQ is necessary with subs. It doesn't matter how ruler flat your speakers are since speaker-room acoustics interaction will absolutely dominate what you get down there. EQ and other filters can be beneficial since they may be able to "smoothen" or "flatten" not only the frequency response, but also the phase response (to a lesser degree) of the sub as well -- perhaps even filter out non-coherent frequencies causing cancellations. But when it comes to wide and deep nulls, EQ can't do much.
If pairing non-matched subs and speakers, there definitely is a huge benefit in using an interface that allows more fine control over bass management e.g. hardware like miniDSP or even software like JRiver.
Been doing some experimentation with
REW's new
alignment tool myself:
White trace is the prediction. Yellow and green traces are actual measurements taken with different amount of delays applied. You see phases diverge increasingly above 140Hz or so... and so at the expense of better phase alignment up between 150-250Hz -- esp. if no sub LP is applied (very small bass boost benefit overall), I opted for 16ms delay instead.
The Neumann KH120s, while not high-passed by me, already have a natural 4th order acoustic roll-off on their own being bass reflex speakers. The sealed sub's phase with/without a low pass is good except at around 140Hz and above the room just messes it up significantly more than what's normal, otherwise -- filtering above that point is quite beneficial.
My ceiling is quite short as well -- maybe 7 and a half feet -- but I was able to navigate away from my own 70-80Hz cancellation issue by using a standing desk.
REW's alignment tools are very useful, but it doesn't always work without a hitch. For more complex bass management crossovers in problematic rooms and setups, its predictions may not be so on-point. As in the case of my sofa-couch listening/movie watching area where I use overlapping filters... and these are for very, very good reasons.
Nevertheless, different setups (midfield vs nearfield) with different settings and curves result in the end with similar evenness in their alignment and steady-state responses.
Spectrogram wavelets
*can't really get rid of the inherent, problematic room nodes so what one can try to do is navigate around them.
