What type of compression algorithm does GIF use?

GIF uses LZW (Lempel-Ziv-Welch) compression. It is a lossless algorithm, so decoding restores the exact pixel data without the quality loss associated with formats such as JPEG.

CompuServe selected LZW when it introduced GIF in 1987 because the method worked well for images with repeated pixel patterns, such as logos, diagrams, icons, and areas of flat color. LZW builds a dictionary of recurring sequences and replaces repeated sequences with shorter codes, reducing file size.

A common mix-up is confusing LZW with run-length encoding. GIF's predecessor at CompuServe used run-length encoding, but GIF itself uses LZW for its image data. Huffman coding is important in other compression systems, while JPEG uses a very different, generally lossy approach based on transforming image data.

GIF's compression is especially useful for simple graphics but less effective for photographs, gradients, and noisy images. GIF also limits each frame to a palette of at most 256 colors, which is why PNG is usually better for high-color static images.