Hi Mark -
Interesting approach, using low pass and group delay. I wrestled with a 16 wide of d&b B2s back in March, modeled with Arrayshow. You are indeed correct about strange things happening at higher frequencies as you add delay incrementally to the outside. I managed to get it so that the weirdness was far enough outside the low pass for the B2s, and was down in level enough it didn't matter in the end.
This was 16 standing on end, next to each other. I ended with 21 msec. of delay on the outer pair. It seems a wider block of course requires more delay on the outside, which lowers the frequency at which the weird things start to happen. In the end I was good up to around 60 Hz, which was fine for the application (B2s in infra mode). Walking the amphitheater the level was pretty much equal out to maybe 140 degrees, with no hot spot and no nulls created by interference by separated sources.
One thing that comes with this is the loss of apparent punch, or whatever you choose to call it. I attributed it to spreading an acoustic source out to 140 degrees or so, vs. 20 degrees. 16 B2s in a row obviously sets up a killer power alley
What got all this started was a client request for even subbass throughout the venue for a rave (Ultra Music Festival 2004 in Miami) without the power alley. The previous year, there was no type of delay scheme for a similar setup. This thing really worked well. However, I did spend quite a few hours modeling it.
Do you have examples of the implementation by low pass, rather than "regular" delay?
Thanks, this is interesting stuff IMO......
The attached photo is a view of part of the amphitheater. You can just barely see some of the subs in the lower left hand corner, against the curved rail. Interestingly, the radius of the railing factored into the equation and had to be taken into account when modeling.
-doug