I asked Wayne Parham about putting my 4pi speakers in a baffle wall and here is what he said on his forum
"All of my speakers respond well to this kind of placement. They are all designed to be used in quarter-space or eighth-space, and I don't use baffle step compensation filters. My thinking has always been that the baffles are rather large, so the transition happens in the lower midrange, close to where room modes take over. That's not where you really want to bump up the power, but then again, you don't want it to be lean down low either. So instead of including baffle step filters, I've suggested placements that reduce the problem acoustically. And as for room modes, a different but related matter, they can be effectively mitigated using a multisub approach.
Placing the speaker close to the boundary reduces baffle step, among other things. The closer the speaker baffle is to the back wall, the less the baffle step. If you can recess the speaker in the wall flush, the baffle step goes to zero. It also makes the reflection off the back wall go to zero too. Those are both important goals.
Some people try to reduce the negative influence of the back wall reflection by moving their speakers out away from the back wall. That's reasonable, a good idea, I suppose. It reduces the amplitude of the rear reflection because of distance. It increases the delay of the reflection too. There's a window of time where reflections are most troublesome, and by moving the speakers away from the back wall, they hope to decrease the early reflections, and to delay them out of that window.
But I go the other way, for several reasons. It is hard to space the speakers far enough away from the back wall to do much good in most living spaces. Most rooms in peoples' homes aren't big enough. So I use another approach. I prefer to use directional speakers that limit the rear wave in the first place, which decreases the amplitude of the rear wave by virtue of directivity. Instead of trying to make the rear reflection late, I try to make it so early it is indistinguishable from the direct wave. As you decrease the distance from baffle to boundary, when the distance becomes zero, so does the reflection delay. So that's the direction I go.
That's why I like using directional speakers designed to mount right up against the rear wall or in corners. The rear wave reflection becomes less of a problem and baffle step is nearly eliminated, because there is so little transition region. If the speakers are mounted with their baffles completely flush with the walls, or their drivers mounted at the apex of the corners, baffle step and rear reflection are completely eliminated."
On another question I shared with him some test results that Ruben posted on his SMX screens and the measured frequency response and he said:
"In truth, there are a lot of things in the environment that cause 10dB swings in what would be an otherwise perfectly flat response curve in an anechoic environment. Heck, just the reflections off the walls do that in the region below 200Hz or so.
Then again, I'm not sure I would be willing to accept response swings so large in that frequency range, right where our hearing is most sensitive. Here are a couple things I'd try:
First, angling the speakers might help because the much of the reflected energy would be directed away from the speaker instead of back towards the source. The self-interference ripple is caused by a reflection that combines with the source, and is therefore strongest when it is straight on.
When a boundary is a quarter-wave away from the source, the reflection is a half-wavelength late, which forms destructive interference. When the boundary is a half-wavelength away, the reflection is a full-wavelength late, which forms constructive interference. The relationship between source and reflection changes with wavelength i.e. frequency, so the resulting response curve develops ripples, as some frequencies combine constructively and others combine destructively.
If the reflection isn't directed back towards the source, the ripple would be reduced. So angling them inward should help. On-axis sound is at an incident angle that reflects away from the source. Far side off-axis is even further angled away from the source. Only near side off-axis sound is reflected back towards the source. Of course, that's the problem with this approach. These speakers generate sufficient output off-axis that there will be a fairly significant reflection back to the source even with a lot of toe-in. But it's worth trying. Try the recommended 45° toe-in to see if it helps reduce screen ripple.
Second, putting some open cell foam or other semi-absorbant material between the speaker and the screen may smooth the ripple. It would probably only be required in front of the tweeter if the screen is placed close enough. Increase tweeter output slightly to compensate. It would make the reflective "boundary" formed by the screen be somewhat "fuzzy". Instead of having a partially reflective boundary that is at a single distance from the source, you'll have a region of semi-transparent, semi-reflective material for the sound to pass through. The self-interference reflection could not line up with well-defined quarter-wave nodes."
I asked about the open cell foam from Tropicals
"I think this is a good application for this open cell foam. I'd place it between the horn mouth and the screen, perhaps attaching it to the screen in front of the horn so sound has to travel through the span, not just the screen. Don't leave a gap because we're trying to eliminate transitions, having only one, where the sound leaves the screen towards the listeners. You can fill the horn with it as Geddes prescribes too, which will cause the sound to travel through the same sort of space all the way from the phase plug to the screen."
One last thought, with acoustically transparent woven screens you need to be aware that the level of transparency decreases with an increase in the angle.of incidence.