Dome arrays with significant distance between the tweeters and horizontal lobing,
The Grundig 190s were wonderful to listen to but then that was 50 years ago.
The Trick is to get the drivers close enough together and stack the bottom and top of the wave guides as close together as you can. It mitigates most of the lobing. One long continuous ribbon (by truncating) and by aligning the drivers further back from the front baffle vs the mids. The same needs to be done with horizontal and vertical truncation of the mids or sub/bass drivers. Staggering and fitting a second line of cones BETWEEN the inside or outside line of mid or bass drivers. I like cones to the outside, preferably in a separate cabinet. I've never had much luck with an LS of bass drivers in the same cabinet. The math just doesn't work.
They used to lay one ribbon on top of the other. It tipped the drivers either forward or backward (negative or positive caster) to a center point. It acted as a point source and where negative met positive was the alignment for ear height.
Sub/bass drivers that were in two or 3 vertical staggered arrays had very little lobing. If you add convex to the array and floor to ceiling, there is literally zero lobbing.
The same goes for neg/pos lapping of ribbon tweeter drivers that are tipped towards the center at the top and bottom. Infinity QL1 and IIs had an adjustment that few people took advantage of. It reduced tweeter lobing considerably if you followed the instructions. It helped a lot by lowering the XO point between domes and ribbons, too.
Domes vs cones work the same way. The problem is stacking. Domes don't stack very well in my experience and truncation is darn near impossible, BUT staggering 2 to 3 line arrays works very well if you can afford the drivers (for one). Most people don't understand that the closer you get to the top and bottom (ceiling and floor plane) the fewer problems you have overall. In other words, short LS seldom works, whereas hybrid LS/PS can work very well. A single tweeter in the front or back or both that is time-aligned with the mids is one of my favorite types of speakers to listen to.
Some speaker designers get all bent out of shape behind the theory vs practicality/cost. Simply put, it is NOT rocket science. Joe Blow couldn't tell the difference. Look at Infinity IRS V. They lob worse than any speaker I've ever listened to, and they were still a lot of fun. But to listen to the critics, they were the greatest. The VMPS RM50 was far superior with a hybrid LS/PS. The RM50 also closed off the rear of the front and rear drivers, where the IRS V didn't.
The depth of the RM50 cabinet 24" (?), mitigated arrival times from the front and rear baffles. The front wall distance to the rear of the cabinet was crucial. There is a formula that is fairly easy to work out if you tape a grid on the floor and start your measurements at the seated position. About 24 to 36", depending on the rear wall distance. (RM50). IRS V the middle of the room is as good a place to start as any. LOL
The mid pics in 7311 are an example of what not to do. Staggering and lapping might not have looked better, but they sure would have measured better. There would be a lot of cancellations. The one saving grace is being close to the ceiling and floor plane.
The IRS V and Betas had ZERO imaging but great immersion sound. I owned IRS Betas for 17 years. Tons of fun and a great servo bass system, considering the price difference between the V and the Betas. 10-12K vs 80+K delivered. I repaired quite a few sets, 1980-2000.
Regards
teufel.de
Finally enough drivers in one speaker 
.jpg)
www.on-mag.fr
