That's almost exactly what I did. I got rid of the 5.5mh and 2.7 ohm baffle step circuit. I went 150uf and 4mh for high pass into 1.8mh and 10uf low pass into a 3.3 ohm resistor. This made it better and the frequency response is better in that region.
That is a near-field mid range measurement. Earthworks M30 through a Scarlet interface. Microphone about 1 cm away from the exact center of the cone.
Here are my final impedance and on-axis frequency response. The frequency response is near +/-1.5db from a 1 meter on-axis measurement (200Hz-20Khz). The impedance still dips to 2.5 ohms but is above 3 ohms for most of it so my Class D amps should be fine. I've also attached a graph of the on-axis response plus an averaged horizontal off-axis response to show the downward tilt. It's probably the best result I've ever gotten with that. All of these measurements use gating of about 3.5ms and no smoothing. I listened to them on my listening room last night and they were revealing, smooth and all the good stuff. I will take measurements in there when I have some time.
