As a "kid" I did some work in NYC for Dynaco. The best way to use this concept, in my opinion, is to merely take the passive approach. Wire a single or pair of rear speakers to the front speakers with wire only. The positive of the left to the positive of the left rear, the positive of the right to the positive of the right rear, and the two negative posts on the rear speakers wired together.
<snip>
The little passive boxes Dynaco sold had a variable resister on the ground side of the hookup to control rear level. I tried this with some success in several of my vehicles but found that the changes in stereo seperation with reception quality was a fly in the ointment, so to speak. I think it worked better that some of the complex decoding systems we have experienced over the years, and it introduces no noise or distortion of its own into the mix.
Richard
Wow I completely agree with you. The interesting thing about two rear speakers hooked up that way is that the left rear and right rear are in phase with the front speakers on their sides, but out of phase with each other. So as a sound moves off-center it wraps to the side, and every bit of leakage to the opposite rear channel is completely out of phase, so there is acoustic cancellation. Center front has TOTAL separation from center rear at all times WITHOUT artificial logic steering, and dead-center rear stays in the rear because the psychoacoustic effect of 180 degree out of phase signals in front is to feel vaguely to the rear, and can be cancelled from the front further by minimal front blending. It will be out of phase and blurred in back too, but the point is it will stay in back where it belongs. If recording engineers thought in terms of left front-center front-right front and a diffused rear, then it would be ideal. Ironically the actual 'diamond' layout is worse for reproducing that diamond pattern, because a single rear speaker will be in phase with one front channel but out of phase with the other.
So let's say you design a new decoder, one that treats stereo as an actual quad matrix. The artist or engineer still thinks in terms of a left-to-right stereo spread, except it will bend around the listener that has a passive adapter, and a seperate rear sound position is a possible extra bonus. And although there is really only a single dead-center rear channel, it can generate it's left-or-right directionality by choosing which front speaker it's in phase with, by choosing if it's encoded as (+L-R) or (+R-L).
You have alot more separation to work with if you only derive 3 channels out of stereo mathematically like this instead of 4, and harness psychoacoustics for the minimal rear separation that's needed for 4 apparent channels.
My last voyage into passive quad 10 years ago I tried both the four-corners and the Wendy Carlos speaker placements, and each had songs they were better suited to. And I don't mean just slightly better, I mean profoundly and absolutely, making the difference of whether a 3D soundstage was percieved or not.
So I dunno, either you plan for 6 speakers, or you use four corners and have a setting that wraps dead center left (and right) equally between the side speakers instead of the "horseshoe" wrap, in other words dead-center left appears midway between front left and back left instead of wrapping all the way to left rear. I look at the distribution patterns of SQ and QS and I see that *neither one* can place a sound dead-center-left (or right) in-phase, only pure amplitude matrixes like EV and Dynaco can do that. I'm not sure how QS got away with calling itself "regular matrix" and calling SQ "phase matrix" since both are equally phase matrixes and neither regular in any way.
Way back in high school the complaint I heard about quad was that it sounded "too spacey". I even knew a guy that got a cheap-o quad system for his birthday, and sold his two rear speakers because he didn't like it. I believe this is because of all the weird random phase angles, and wanting to listen AT something with a clear location instead of being immersed in it.
This is just my opinion but I think quad should have concentrated on making the front soundstage 3D instead of trying to place you in the center of the band. I think my current interest is to distribute a normal stereo signal among four speakers arranged in a 180 degree arc, and then two speakers in back for the out-of-phase sounds. It would probably work great for any quad or Dolby you run through it.
The only thing about passive matrixes is the L-R sounds have a hissy, spitty quality in the highs. I suspect this would not be the case if sounds were deliberately placed there by recording engineers instead of random ambient noise being extracted and reproduced. But that's what first generation Dolby did, deliberate L-R for the rear, so I'm eager to try.