Method for rendering shadows with blended transparency without producing visual artifacts in real time applications

摘要

A method, system and computer-readable medium for generating a shadow for a three dimensional model without visual artifacts. A shadow of a model is projected onto a shadow plane using a shadow projection matrix and disabling alpha blending, creating a projected shadow model. A three-dimensional (3D) bounding volume of the model is projected onto the shadow plane and clipped against the view frustum. Based on the bounding volume projected onto the shadow plane, a screen extent that encompasses the entire projected bounding volume is determined. A texture map sized in relation to the screen extent is created in memory for storing a texture map. This texture map is cleared to a non-shadow color and a chromakey is set to the non-shadow color. The projected shadow model without alpha blending has its colors and texture set to the desired shadow color. The projected shadow model is then rendered onto the texture map. Thereafter, the projected bounding volume vertices are set to the shadow's alpha value and the texture coordinates of the projected bounding volume vertices are set to evenly map texture into screen space. With prospective texture correction disabled and blended transparency enabled, the projected bounding volume is rendered into the scene with the shadow texture applied.