A method for compressing multiple channels using an MPEG-compatible technique (Fig.: block 96). The method encodes differences between frames within and among each channel (Fig.: blocks C1, C2, and CN). Some frames are predicted and some are skipped. Differences between the current and predicted frames for each channel are compared to a threshold E0 for each successive group of frames, after each group of each channel is encoded. If for any of the groups of frames, the difference between the current and predicted frames of any of the channels reaches the threshold, the method inserts an I-frame into the sequence of encoded frames, to restart encoding. Computing the differences and encoding I-frames is implemented for both intra- and inter-channel differences. The encoded frames are transmitted to a decoder for display. The method uses artificial intelligence or fuzzy logic.