In theory; a horn, vent, etc., has an end correction, i.e. its air mass plug hangs out a bit, which in freespace is ~0.613*radius to find the terminus [mouth] acoustic radius and space up the horn however amount this new radius 'touches' the nearest reflective boundary. This assumes a round horn of course, but until the terminus becomes a high aspect ratio this 'rule-of-thumb' is close enough and ensures an adequate space for horizontally opposed layouts, but if used vertically to increase horizontal dispersion, then you have to convert the radius to the terminus shape to find the spacing, though folks normally just use the regular spacing.
When XO'd to another horn though, ideally both horn's acoustic radius needs to be accounted for, so you want the spacing such that their output overlaps at the XO point and since the bigger horn has a larger radius it's going to jack up the HF horn a bit more than just its radius.
Normally, this is all done from measurements, but doing a scale drawing of frequency circles or rectangles is plenty good enough. All that's left then is to empirically find the best HF tilt angle if a HIFI/HT app, otherwise measurements made at various seating positions to find an acceptable compromise.
If there's 'wings', then the end correction is 0.732.
GM
Bookmarks