You have drawn a figure in front of a doorway and want it to rise off the page when someone puts on red/cyan glasses. A red/cyan stereoscopic picture needs two views of that scene, one for each eye. A colored shadow alone gives a different result.
Anaglyph 3D images begin with two views
In anaglyph 3D images, the left and right views are combined into one picture. The glasses separate them: the red lens sits over the left eye and passes the red channel; the cyan lens over the right eye passes green and blue. Each eye receives a slightly different view, and the visual system can interpret their differences as depth. NASA explains the two-view principle with images made from separate camera angles.
For simple grayscale 3D anaglyphs, put the left view into the red channel and the right view into both green and blue. Colored fringes visible without glasses mark places where the views differ; they are not an extra outline.
Anaglyph images can also be made from two photographs taken from slightly different positions. For illustration, you can construct two views from separately drawn elements. The important part is that their positions differ according to the depth you intend, rather than every piece being moved by the same amount.
Separate the scene into planes
Start with a small test: a figure, a doorway behind it, and a distant wall. Put each on its own transparent layer, with enough artwork behind the figure to fill any gap that appears when the views separate. Keep a copy of the original file; the individual layers make later corrections much easier.
Make two groups named Left view and Right view, each containing the three scene layers. Their canvases must have the same dimensions. In pictures made with anaglyphs, choose the doorway as the reference plane and leave it in the same position in both groups. Corresponding points there have zero horizontal disparity: through the glasses, that plane should appear at the screen surface.
Now move the figure in the left view a few pixels to the right relative to its position in the right view. With the usual red-left, cyan-right glasses, that crossed displacement places the figure in front of the reference plane. Move the distant wall’s left-view layer a few pixels left relative to its right-view counterpart to place it behind. Keep corresponding features at the same vertical height. A vertical mismatch makes the two views harder to combine.
For 3D anaglyphs, begin with a small separation and test at the intended viewing size. If the figure feels detached or difficult to fuse, reduce its relative shift. Restrained offsets often read more clearly than an attempt to put the figure in the viewer’s lap.
Combine the channels
To make pictures with anaglyphs, use an RGB editor that supports channel editing. Preserve the two working groups so you can change their positions.
- Export or duplicate each view as a flattened grayscale image on the same canvas. Label them left and right.
- Create an RGB composite. Copy the left image into its red channel.
- Copy the right image into both the green and blue channels. The right view now supplies cyan light where those channels match.
- Display all three channels together. Save a separate composite, keeping the two source views editable.
- Put on glasses with red over the left eye and cyan over the right. Check the doorway first, then compare the figure and wall.
The University of Southern California’s stereoscopy guide describes the red-left, cyan-right viewing convention. If your editor uses layer blend modes rather than channel pasting, the goal remains the same: left-view luminance in red and right-view luminance in green and blue. Verify the channel assignment rather than relying on layer names.
Know what a duplicated offset can do
Duplicate one flat illustration, shift the entire copy, and you can make attractive red/cyan edges. That is a useful graphic treatment for a poster, especially when the viewer has no glasses. But a uniform shift gives every point the same disparity. It can place the whole flat picture at a different apparent plane; it does not independently separate the figure, doorway, and wall.
The anaglyphic approach uses different shifts for different planes. Translating a flat figure does not create the changing perspective or newly revealed surfaces of a fully modeled object. Draw the hidden background and adjust overlaps in both views so an edge does not reveal a hole.
For more dimensional illustrations, draw or render the left and right views from slightly different viewpoints. A hand in one view should correspond to the same hand in the other, even if its edge falls in a slightly different place. Anaglyphs work when those differences describe a coherent scene.
If you see a double outline
The most common problem is ghosting: through the glasses, a faint copy or colored edge remains visible beside the intended shape. First make sure the red lens is over the left eye and the views were assigned to the correct channels. Then align the doorway vertically and reduce large horizontal separations. If the echo remains around a saturated color or strong contrast, try the grayscale composite and inspect the screen and glasses; imperfect color filtering can let part of one view leak into the other.
