Streamlined tetrahedral interpolation

Abstract

A streamlined technique for obtaining interpolated color output values from corresponding color input values based on color output values stored in a color look-up table, with tetrahedral interpolation being applied so as to dissect each cubic grid of the color look-up table into plural tetrahedrons. According to the invention, processing efficiency is increased by avoiding the use of conditional branches when making a determination as to which of the plural dissected tetrahedrons a particular color input value falls. Preferably, rather than making such a determination, weights for all possible tetrahedrons for one cube are stored in a temporary look-up table, and an index into the temporary look-up table is calculated based on the magnitude of component values for each color input value. The index is calculated in a block of code that does not include conditional branches that are conditioned on the relative magnitude of the component values of the color input value. Most preferably, the block of code calculates the index through use of arithmetic operations, such as subtractions, which throw off a condition code that can be used to build the index into the temporary look-up table of weights. Because conditional branches are avoided, processing of such tetrahedral interpolation is speeded, particularly for processors such as a Pentium® processor that employ dual processor pipelines together with branch prediction.