ok, my 4 cents (adjusted for inflation):
I believe low frequency signals in a room are normally fairly correlated at the ears - I mean left/right ear cross correlation at low frequencies. Bill Martens, who has done a lot of real research in this area has found it to be true as have I. Just to check, I just now loaded a recent measurement (random noise burst from multiple subwoofers measured at blocked ear canal of a dummy head) in our reference room, low pass filtered it at 80 Hz and plotted the left right ears. Very similar, correlation coefficient has to be near one.
I also did some research to try to demonstrate spatial bass (I call it "Bassiousness"). I went to great lengths to make the test sensitive to the effect I was studying. I used some program material that did have decorrelation at low frequencies. I even went so far as to find material that had simultaneous high signal levels and decorrelation (if you look closely at different programs, often the decorrelation comes at times when there is insignificant signal level). I also included as one of my programs the same contrived warble tone signals that Dave Greisinger uses in his demos. I used trained listeners, with a pretraining period (subwoofer only so they could hear any effects). I used the very sensitive 3 alternative forced choice test and double blind of course.
Several subwoofers configurations were tested, from a single sub at front/center to "stereo" subs at +/- 90 degrees, using either "stereo" or "mono" presentation of the bass, and listeners tried to detect the difference. Also, the responses for the difference configuraitons were pre-equalized to be flat (to remove that as a cue).
After statistical analysis, I found that listeners could only detect the difference between front/center mono and left/right stereo bass presentation (the two most dissimilar configuraitons), and that just barely, and under the most favorable conditions. And keep in mind we did not ask if they preferred one configuraiton over the other, just whether they could hear a difference. In the real world, you really need a contrived or otherwise unusual signal to hear it.
Harman International Industries