This is a description of how to take ordinary 2 dimensional photographic imagery and add a pseudo (but very convincing) 3d effect. One important consideration is beginning with an image that has depth to it. A photograph of a wall is a poor choice but a photograph of a landscape with a foregound object in it is a good choice.

Figure 1. This is an example of a poor choice for this process. The image has no visual clues for depth. It is a flat plane. Figure 2. This is an example of a good choice for the anaglyph process as it has a foreground objct (the fountain and angel) a midground (the balcony area) and a background (the trees).

      It helps to understand how stereo vision works and understand how red blue glasses fool the brain into believing there is depth when there really isn't.
      We have two eyes so that we can see in stereo. If you really look with one eye closed you will notice that forgound background relationships are obscured and depth is difiicult to gauge. There are obvious visual clues that cue depth such as overlapping objects and atmospheric fading. Without some of these obvious cues it is difficult to judge depth with one eye.
      Red blue glasses fool the eye by not allowing red information into the right eye and not allowing blue information into the left eye therefore setting up a stereo perception. Our brain fills in the missing color information from the other eye so that we seem to see everything in color but actually only see different parts of the spectrum out of each eye.
      Luckily the web represents imagery using the rgb color space. We can take advantage of the strong Red and Blue color of the rgb color space.
      Adobe Photoshop represents an image by using 3 grayscale channels representing the range of red, green and blue values in an image.

Figure 3. Red Channel Figure 4. Green Channel
Figure 5. Blue Channel

WHat I was doing was making saved selections to indicate foreground mid and back. looking at each red and blue channel and shifting them. Red to Right (RR). THe more you shift the more pronounced the foreground popping effect. For on-screen I used 4 pixels for the fore and 2 pixels for the mid. I then used "erase to saved" to bring the background back to no shift. If you are going to print them you need to do more than a few pixels.

An animated gif image made by a friend of mine named Dan Ross living in Cali

sites of interest to anaglyph 3d people:
the shockwave 3d web

Jaron Lanier's homepage, the inventor of the term virtual reality.
virtual reality and art The knowledgebase...
I glassesan inexpensive supplier of immersive goggles
3d devices
stereographics corp
SGIs virtual reality page
ChromaTek's unique system of polarized threeDee
MITs spatial imaging group
DIPV three Dee environmental system?
stereo pairs
The New York Public library has a large collection of stereo pairs of cnetral park at the turn of the century online
Macweek did an article a while ago about 3d resources
Finally a competing process for making anaglyph images