Andrew T. Duchowski--Gaze-Contingent Display Programming Examples

GLSL: The most recent GCD incarnation using OpenGL 2.0 and its GLSL shading language. This program has a simple GLUT interface and should compile under Linux and MacOS X.

glsl.tar.gz

Notes:

This program has been tested under Linux and Mac OS X and ports between the two painlessly (just type 'make' following extraction of the tar file). To run, on the Mac type 'open texture.app' from the src/ subdir (or double-click the Qt icon), on Linux type './texture' in the src/ subdir.
The mipmap LOD selection doesn't seem to work on my ATI cards (see above).
Nice features include:
- Same as the Qt version below (except the Tobii eye tracking hooks).

Qt: A more robust GCD example with mouse-controlled foveal region, using Qt for the GUI and the Cg runtime programming model.

gcd-img++.tar.gz

Notes:

This program has been tested under Linux and Mac OS X and ports between the two painlessly (just type 'make' following extraction of the tar file). To run, on the Mac type 'open texture.app' from the src/ subdir (or double-click the Qt icon), on Linux type './texture' in the src/ subdir.
The code has hooks for eye tracker functions such as Connect, Calibrate, and Track but in this version these are just stubs. (The code has successfully been expanded to work with a Tobii eye tracker.)
Nice features include:
- User-selectable spatial and color degradation maps (ppm images): spatial degradation is governed by a pgm (greyscale) image, color degradation is goverend by a ppm (color, 3-channel) image - each of the color image's R,G,B channels can be different allowing different degradation functions per color channel.
- User-selectable stimulus images: "File->Open" a ppm image, there are a couple provided in the src directory including lena and runway.
- User-controllable degradation smoothing: select from none, intra-map, inter-map, or both types of smoothing - this controls minification (filtering) between mipmap levels.
- User-controllable LOD: currently set to 4 as the coarsest level, you can select from 1-4 (the coarsest level depends on the image size and corresponding number of mipmap levels, e.g., log_2(n) where n is image size).

Cg: A simple GCD with mouse-controlled foveal region, using GLUT and the Cg runtime programming model.

Makefile
gcd.cpp
vert.cg
frag.cg
You can get all of the above files in this tar file, glut-gcd.tar.gz.

Notes:

I like Cg runtime, except for the fact that as is, it won't run on my Mac since ATI doesn't seem to support semantic binding in immediate mode. The code compiles on both OS X and Linux but runs only on Linux.
As of the 10.3.4 upgrade, this has now been fixed! Thanks, ATI!

ARB_fragment_program: A similarly simple GCD with (passive) mouse-controlled foveal region, using GLUT and GL extensions.

Makefile
gcd.cpp
vert.cg
frag.cg
You can get all of the above files in this tar file, arb-gcd.tar.gz.

Notes:

I prefer Cg runtime, however, the GL ARB_fragment_program extension seems to be more portable. This program thus uses vertex and fragment programs as compiled by the Cg compiler. (Note the slight difference in the vert.cg program -- the appropriate cgc compile lines are contained within the Makefile.) The code compiles and runs on both OS X and Linux.

SDL: A robust GCD with mouse- or eye-controlled foveal region (requires ISCAN eye tracker).

README
passthrough_vp.cg
degrade_grey_fp.cg
You can get all of the above files in this tar file, gcd.tar.gz.

Notes:

This is the full GCD implementation written by Nathan Cournia. This GCD example will display a gaze-contingent foveal region over still images (PPMs or just about any format if you have DevIL installed) and video (if you have xine-lib installed).
Use scons to configure/compile.
This has not been ported to the Mac hence only runs on Linux. (If you port it to the Mac, please let me know.)
See also Nathan's page for further info.
On the VR Net, the code is in CVS in the gcd module:
```
cvs -d/pub/research/uber/cvsroot co gcd
	
```

Sample usage:

./gcd test/kitten.ppm test/falloff.ppm