An adaptive compression/decompression method for color video data with an anti-aliasing mode. 4.times.4 blocks of pixel data are examined to determine which one of four compression techniques should be used on each block. User settable thresholds can be used to shift the types of compression used. Highest compression is obtained when more data is stored in run length blocks of single colors and lowest compression when more data is stored as two colors with a 32-bit bitmap for each 4.times.4 block. One type of compression used provides anti-aliasing.