Re: Does this apply to our BF cabs ?
Posted: Wed Aug 01, 2012 7:09 pm
Well, it's hard to tell if he knows what he's talking about, or just lucky. He made some comments that show he's not ignorant of real-world issues, but glosses over some details. That may just be in the name of brevity, though.
In regards to the main point of controversy, when he says that changing the *slope* of the crossover can change the phase, he's right. Different alignments and slopes will affect phase at the crossover point. That's why (if I've kept all this straight) with a 2-way, 2nd-order crossover, you might need to invert the polarity of one of the outputs. 1st-order would only be 90 degrees shifted between the outputs, so you just have to deal with it. At multiples of 24dB (4th-order), like in his video, it should be back in phase again.
I don't know for sure whether aligning phase via delay will always mesh with actual time alignment or not -- assuming even poles and inversion as necessary. It seems the driver/enclosure response would have some impact, but it may be negligible. At any rate, phase will only be really important around the crossover frequency, whereas time is more critical for the difference between entire passbands. If I had to pick one, I'd pick time and get the phase as close as possible with polarity (180 degree steps). So, for a fixed install, you would ideally want to get an impulse response graph. I think that's probably beyond this guy's intended audience though.
In regards to the main point of controversy, when he says that changing the *slope* of the crossover can change the phase, he's right. Different alignments and slopes will affect phase at the crossover point. That's why (if I've kept all this straight) with a 2-way, 2nd-order crossover, you might need to invert the polarity of one of the outputs. 1st-order would only be 90 degrees shifted between the outputs, so you just have to deal with it. At multiples of 24dB (4th-order), like in his video, it should be back in phase again.
I don't know for sure whether aligning phase via delay will always mesh with actual time alignment or not -- assuming even poles and inversion as necessary. It seems the driver/enclosure response would have some impact, but it may be negligible. At any rate, phase will only be really important around the crossover frequency, whereas time is more critical for the difference between entire passbands. If I had to pick one, I'd pick time and get the phase as close as possible with polarity (180 degree steps). So, for a fixed install, you would ideally want to get an impulse response graph. I think that's probably beyond this guy's intended audience though.