Baffle-step. A fact of physics.
What is Baffle-step diffraction -- article on t-linespeakers.org:
<
https://www.t-linespeakers.org/tech/bafflestep/index.html>
There are a number of ways of dealing with this:
1/ a huge baffle so that the BS is so low it doesn't matter. An approximation can be had by pushing the speakers closer to the wall... the smaller the actual baffle the closer you need to be to the wall. The ultimate variation on this theme is to build the speakers into the wall.
2/ passive BS compensation between amp & speaker -- cored inductors outperform air-core inductors in most cases because you can dramatically decrease the series R in the inductor. Your problem is most likely the fancy inductor and not the Rs. Still this inserts some ugly reactive components where they have a chance to do their worst.
3/ passive line-level BS compensation -- recalc the LR for your amps input impedance -- here you could use really small single layer inductors -- anyone tell me how you would do this with an RC?
Yes:
Connect (R2 and C in series)in parallel with the input and R1 between the hot connectors of input and output, grounds together.
This would give a low corner of fl = 1/(2*pi*(R1+R2)*C) and a high corner of fh = 1/(2*pi*R2*C). For a 6 dB "step", R1 should equal R2, and C = 1/(2*pi*fh*R2), where fh is the high corner frequency.
I would love some graphics here...
4/ active line-level BS compensation -- <
https://www.t-linespeakers.org/tech/bafflestep/bstepcompo.html> The same thing could be done with tubes, but it is an extra stage or two.
5/ The addition of a 0.5 bass speaker -- Set up a low pass filter for the lower woofer to coincide with the spreading loss of the upper woofer, net result is constant power output through the bass to midrange frequencies. Simple to do, all you have to do is run the upper woofer all the way down, while the lower woofer gets a 1st order rolloff. But it's best to do this with measurements, as for any of these techniques.
These are real tricky to get right, since as the bottom woofer rolls off its phase goes thru 90 degrees wrt the midbass that goes all the way up, so you get some phase disturbance in this region. Now if you put the 0.5 woofer on the back (see 6/) the phase rotation is in the shadow of the cabinet so is much less a problem... the cabinet also guarantees that the BS fill is perfect even if the XO is a little high.
6/ and my favorite approach -- what i call brute-force baffle-step compensation -- the bipole -- identical speakers on the front & the back. If you think about what happens when you hit baffle-step -- the sound is progressively wraping around the cabinet more & more as the frequency goes down. The same is happening with the back driver but facing the other way -- as the output of the front driver starts to fall off, it is exactly filled in by the bafflestep on the back driver. You can also roll off the back speaker above bafflestep giving you a 1.5 way without the phase problems inherent in the 0.5 part being on the front of the baffle
Fig 5.2 <
https://www.t-linespeakers.org/projects/tlB/response.html> 1 driver vrs 2.
The compromise is that the speaker has to be out into the room more (not necessarily a bad thing) -- my room is excellent with a bipole. And of course the expense of that extra set of drivers -- the last actually cheap when you consider the 6dB power increase required for the same midrange levels if you use a filter.
You can simulate the baffle step, including the transition region with this program:
https://www.tolvan.com/edge
