We're not communicating too well at the moment. Let me try to clarify.

1. Monochrome anaglyphs are easy. You use the left image in the R channel, and the right image in both the blue and green channels. It's the blue+green that produces the characteristic cyan.

I recently posted about not knowing how to produce conventional anaglyphs, but that was a brain fart. I'd done it years ago with Viking Lander images (and about popped my eyeballs out of their sockets with that 1-meter separation), but had forgotten how I did it. It's easy once I was reminded of the technique:

2. Color anaglyphs are just an elaboration on the technique. You use an actual red image from the left eye in the R channel, and use green and blue images from the right eye in their respective channels. It's actually not necessary to present full RGB images to each eye, which is, in any case, an impossibility in the anaglyph format. You have only three channels to work with, not six. If I had to make an anaglyph from two RGB images, I'd extract the channels from each and use them as outlined.

That color anaglyphs are possible is demonstrated by the example I posted. The relief is inverted, but it's distinctly 3D. As JMS pointed out (the obvious--doh!), turning the glasses upside down does somewhat mitigate the problem.

That said, I'm faced with the puzzle of why, all else being equal, it wasn't the right eye on the rover cameras that was equipped with the full spectrum of filters. It's not an earthshaking mystery, since I can reverse the glasses, but it's a puzzler.