How do anaglyphs work? Essentially, they let you see right and left views in the same image using coloured glasses. To help you understand the process, the diagram at left shows how an anaglyph was created from a computer model of a flower. Two cameras representing the left and right eyes were trained on the flower model yielding two slightly different views of the model. These renderings appear in the viewfinders of the virtual cameras. The left eye view is then converted to red (blue and green channels set to zero) and the right eye view is converted to cyan (red channel set to zero). These two images are then combined. They can be seen overlapping in the middle of the diagram. When they are fully overlapped the result is the anaglyph shown at the bottom of the diagram. This anaglyph therefore contains the left eye view and the right eye view, but as different colours. When the image is viewed through red green spectacles then the two images can be separated. This is because the left eye only receives the red image and the right eye only the green.
3D computer models can be used to create anaglyphs but so can photos (which is generally a bit easier). In both cases the principle is the same. The anaglyph encodes both right and left viewpoints which the coloured glasses allow the viewer to separate. The brain receives these two views and so perceives depth (stereopsis).
Instead of red green glasses we can use polarising lenses. In this case the right and left views are differently polarised. This has the advantage of preserving colour and so is the technique used for 3D movies in the cinema.