For Roger and then Nyall - comments appreciated
First, thank you Roger for prodding me to experiment with the sub. Here is a series of tests dealing with some of the vagaries of the Theta Casablanca III.
A few things have changed since I did my last tests. First, the Modex Plates are installed (side walls, very front corners), the 6' x 4' panel of Skyline diffusers is installed and the larger/thicker carpet is in place. Also, after a debate on the REW forum, I changed the orientation of my mic to horizontal (using the 0 degree calibration file) and pointed in the general direction of the main speaker being tested.
Perversely, the way to get pure stereo, front L/R only, from a digital input into the Casablanca is to use Matrix mode and tell the Casablanca that there are no other speakers (as opposed to selecting "stereo"). Don't ask - it's a hangover from days gone by and, well, is the way it is. Obviously, a sub can be introduced by setting the number of subs to greater than or equal to one. Perversely, even if you select Full in the CB's crossover settings, as I was reminded while doing these tests, if Matrix mode is selected and subs >=1 then a signal is still (counterintuitively) sent to the sub (quite what, I do not know). If you select Full/LP then full range is sent to the main speakers and low pass is sent to the sub according to your crossover settings. For these experiments I had my crossover settings set to: Butterworth, 80dB, slope 24dB, normal phase for both the low and high.
The results of these tests were rather intriguing - especially the phase reversal at the sub (Rel Stentor II).
First, "pure stereo", Full / Sub=0:
Now "Full" in the crossover settings but with the sub enabled. Some sort of Lord-knows-what mix.
Then "Full/LP" and the sub enabled (crossover settings as noted above). Normal phase set on the Rel. So here I have LF going to 3 speakers but it just keeps getting worse.
I then tried reversing the phase on the Rel and things start cooking (well, relatively speaking). But why? I'd be really interested in people's thoughts on this.
Here are the first and last sets overlaid on each other:
Lastly, here are some of the charts which Nyal and Jeff's very good paper focus on. All are run with the settings of my last run (i.e. sub with reversed phase etc). I have tried to use the same ranges and settings in Nyal's paper.
ETC. Nyal, I've not looked at this before and I must admit I find it difficult to interpret. I would appreciate your thoughts.
Left and right spectrograms. Very low end is still awful and I suspect a challenge to fix:
Left and right mid-range RT60/20/30. I know I need to deal with left-side first reflections (in particular).
I would really appreciate any thoughts on these. The room setup is depicted in my first post. More work to do...
Thanks, Steve (and apologies for all the charts!!)