A video encoder/decoder, or codec, is a device or software program that converts visual information between analog and digital formats. The primary goal for most compression methods is to generate video that most closely maintains the original high-fidelity source while delivering the smallest file-size possible for storage and transmission. A file is compressed for transmission and storage and decompressed for viewing or transcoding.

MPEG-4 and H.264 video compression algorithms reduce video data between sets of frames by using inter-frame prediction. Difference coding, which refers to a situation in which one frame is compared with a reference frame and only changed pixels (with respect to the reference frame) are coded, is one technique for accomplishing this, thus reducing the number of pixel values being coded and sent. If such a sequence is displayed, images appear in such a way as they did in the first, unaltered video sequence.

With inter-frame prediction, there are three types of frames, ‘I’, ‘P’, and ‘B’ frames. The first image in a video is always an I-frame, which stands for intra-frame. This is a self-contained frame that can be viewed independently. A P-frame, or predictive inter-frame, makes references to parts of earlier I or P frames. These require fewer bits than I-frames. A B-frame, or bi-predictive inter-frame, is a frame that makes references to both an earlier reference frame and a future frame.

When a video is restored via video decoder decoding the bit stream shot by shot, decoding should start with an I-frame. Should you use P-frames and B-frames, they should be decoded together and with the reference frames.

Common forms of video encoder/decoders are H.264, MPEG-4, MJPEG (motion JPEG), DV and HDV.