The stereoimg command generates stereo pairs using the specified rendering program (ribbonjr, conic, or neon). By default, the image is stored in a TIFF file called stereoimg.tiff. One shortcoming is that stereoimg assumes that the size of the screen is 1280 by 1024.
The command-line flags specific to stereoimg are:
-c
Generate a crosseye stereo pair instead of walleye.
-t
Normally stereoimg fits each half of the stereo pair into a half-screen by expanding one dimension of the original graphics window until the aspect ratio is correct. If the -t option is given, the window will be brought into the correct aspect ratio by trimming one dimension. This may result in part of the image being clipped, but will also result in less blank screen area. For optimal screen filling, reshape the graphics window to about the same aspect ratio as half the screen and get the image you want before running stereoimg.
-s
Show the image immediately with the SGI utility imgview. Without this flag, the image is deposited in the output file and not displayed.
-o outfile
Deposit the image in outfile instead of the default, stereoimg.tiff.
-p renderer
Use renderer to generate the pair image. Legal values for renderer are conic, ribbonjr and neon. This flag is mandatory.
-v
Make the pair in a fashion appropriate for use with a mechanical stereo viewer, typically used with journal publications. Each eye image will have a square aspect ratio. If you convert the image to PostScript with itops, giving a scale factor of 0.375 (with -s) will produce a final image where each eye image is 2.5 inches square, a typical size used in journals. Other scale factors can be used for larger or smaller images.
-C
This flag indicates the end of arguments for stereoimg. Any arguments which follow the -C flag will be passed on to the rendering program. See the conic, neon, and ribbonjr manual pages for full descriptions of their command-line arguments and parameter files.