mozjpeg is a fork of libjpeg-turbo that aims to speed up load times of web
pages by reducing the size (and, by extension, the transmission time) of JPEG
files. It accomplishes this by enabling optimized Huffman trees and
progressive entropy coding by default in the JPEG compressor, as well as
splitting the spectrum of DCT coefficients into separate scans and using
Trellis quantisation.

Although it is based on libjpeg-turbo, mozjpeg is not intended to be a
general-purpose or high-performance JPEG library. Its performance is highly
"asymmetric". That is, the JPEG files it generates require much more time to
compress than to decompress. When the default settings are used, mozjpeg is
considerably slower than libjpeg-turbo or even libjpeg at compressing images.
Thus, it is not generally suitable for real-time compression. It is best used
as part of a web encoding workflow.