Video compression uses the sum of absolute differences (SAD) algorithm to reduce video file sizes by removing redundancy between consecutive video frames. The SAD algorithm calculates the pixel-by-pixel differences between two frames and discards pixels that have not changed. This results in data loss but high compression ratios needed for transmitting video over the internet. The document presents the SAD algorithm and its Verilog implementation to compress a video by calculating the absolute differences between corresponding pixels in two frames and summing those differences.