Abstract: A direct interaction stereoscopic display system that produces an augmented or virtual reality environment. The system comprises one or more displays and one or more beam combiners to virtually project high-resolution flicker-free stereoscopic 3D imagery into a graphics volume in an open region. Viewpoint tracking is provided enabling motion parallax cues. A user interaction volume co-inhabits the graphics volume and a precise low-latency sensor allows users to directly interact with 3D virtual objects or interfaces without occluding the graphics. An adjustable support frame permits the 3D imagery to be readily positioned in situ with real environments for augmented reality applications. Individual display components may be adjusted to precisely align the 3D imagery with components of real environments for high-precision applications and also to match accommodation-vergence distances to prevent eye strain.