I've decided to go through the data Scott Sehlin posted on the 3.1's years ago. I want to create an EQ alternative to his crossover changes that creates a flat on-axis response and is not room dependent. To do this, I merged his nearfield data with a box simulation accurate enough for our purposes.
I would argue that the crossover is well enough designed, it's just
wrong. The resistances are silly and the tuning is extremely bright. We all know this, we've all seen it, it's been discussed to death. Here's my simulation, done similarly to Scott's, but in VCAD.
It's bright.
And if I throw in the estimated driver separation, our phase looks pretty great, I must say.
There are all sorts of variations, highlighted in green (but not run through the optimizer) that Scott made to the original crossover with incremental improvement.
However, I'm sure the stock crossover can support equalization for tone. Starting with the stock crossover, I have a 6-band PEQ that can flatten the tone, and a 5-band PEQ that will tilt it for an extremely high preference score without a sub, for a total of 11 bands. It would make sense to ignore the 128hz EQ and simply correct for your room below 500hz, especially since the data here is simulated. I have previously shared my REW data and in-room response. This bookshelf can play 40hz in-room very well and very loudly. I am constantly impressed by this woofer.
A very worthwhile investment for the speaker quality. My old REW distortion measurements were at -60dB even as high as 110dB SPL in-room. I haven't seen a better deal in audio. JBL LSR-308's sound better at low volumes, but can't touch the SPL levels these reach. Dynamics are fabulous. Mine pull double duty as home theater mains and my main listening stereo. I keep thinking of bass boxes as stands and I'm nearly settled on the SB29NRX75-6 in series. With a bit of boost they hit 20hz no problem.