stereo rendering problem/bug

[michaelrepucci]

We would like to use frame sequential (active) stereo with shutter goggles to study the perception of depth. (Currently, we are using red/green anaglyphs.) Unfortunately, there seems to be a bug in the Vizard rendering engine that prevents its use in this application. It seems that, with active stereo enabled, the position of a moving object is rendered as equivalent in sequential left and right frames.

For example, say the refresh rate is 100 Hz, an object is moving at 100 meters/second, and at time t=0 (immediately before the first frame) the object is at position x=0. Then what SHOULD happen is that the object is rendered at x=0 on the first (left eye) frame, x=1 on the second (right eye) frame, x=2 on the third (left eye) frame, etc. However, what DOES happen, instead, is that the object is rendered at x=0 for both the first and second frames, x=2 for both the third and fourth frames, etc. This behavior causes moving objects to be perceived at different depths than intended, thus making Vizard unusable for experiments in depth perception with active stereo.

This is clearly a very disappointing discovery. Is there perhaps a patch, fix, or work-around that we could use to get the correct behavior?