So, just want to also give you a big ole heads up now, pretty important
So I have a collection of CDs and a collection of bad diaphragms.
When using linear phase crossovers , I was using rectangular windows and optimized to -150db ..... and than using peq to cut the peaks in the acoustical response. So, the fir performance was good using that method, however I kept blowing CDs , one of my 2408h voice coil came unglued from diaphragm. What was happening was the stop band wasnít cutting LF. After a really close look , a 100db measurement showed at -50db some LF peaks at 200hz and 50hz ........ so it was getting a few watts still way down low.
A purely linear phase crossover in an fir will have stop band issues. I talked to pos over at diy and apparently thatís pretty normal for an fir. So, I came up with a better approach and a better sounding filter all together.
1. Use an IIR crossover to find what slope and xo point you want to use, than make it in a
Fir but use this recipe.
If you want a LR4 and you notice the acoustical shape is LR2 with your IIR -LR4 than make a linear phase LR2 and add a BW2 on top of it in IIR. This will do two things , 1. Make it truly linear phase and 2, give you adequate stop band rejection.
I didnít want the ringing from an IIR from its inherent ringing from its loopbacks, but itís such a tiny tiny tiny arguably inaudible ring that it so much outweighs the risks of blown drivers.
On a different subject. Iíve also found the GD caused by an IIR can be desirable to get alignment to mid on the HP side , meaning the horn using a purely minimum phase crossover.
On my setup if I use a LR4 at 1k acoustically it looks like approximately 9 dB oactave roll off using a LR4 overlay shows that also. I am now using a 18db linearphase LR xo and a BW2 IIR on top of it (at 1Khz) Iím liking it quite a bit. Itís acoustically a LR4 and overlay is almost perfect. However electrically its a mutant LR BW with 36db attenuation. The measured phase is flat. And uniform at stop band and sums to mid much better.