I actually use a Sunfire TG4 processor that has Bob Carver's Holography imaging, works great on both stereo and multichannel.
I have a Carver C-9 Sonic Holography Generator. It basically works by feeding some of the L signal into the R speaker and vice versa to reduce crosstalk. If done exactly right, the crosstalk cancels out at your ear. There was also a system called M.A.R.S. that did much the same thing in the speaker itself. Polk had something similar too.
In all cases, the method used for the solutions is less complex than the original problem, and winds up not completely solving it.
I've had my C-9 for about 30 years, IIRC. I take it out once in a while and play with it. It will make a couple of recordings at one listening position sound great (just don't move your head at all!), but then it makes other recordings sound worse. I always wind up putting it back in the closet.
You can do almost the same thing if you put some acoustic foam like Sonex as a barrier between the speakers. I mounted some on a hat. It looked really weird, but it worked as well as the Carver method. Unfortunately, it blocks a basic part of your hearing mechanism, the filtering you get from your head, the head-related transfer function (HRTF). So it still only works on some recordings and makes others sound worse. (Headphones have the same problem.)
I've also come to wonder whether crosstalk is really a major problem. Logically it is a fundamental flaw of stereo speakers. An event in the concert hall produces two events, one in each channel or the stereo recording. No problem so far, the same thing happens with your ears. Then you play it back and the one initial event becomes four events.
The question I would ask is whether it really makes a difference? Is it really a major problem, or only a minor one?
I'm coming to think it's only a minor problem, and the available cures are worse then the disease.
