Lossy image, audio and video compression: Science
Principles
The science it rests on, with why each principle matters to it
Relies on- Fourier analysis The discrete cosine transform splits a block of pixels into frequencies, so the coder can keep the low ones and approximate the high ones [1].
- Information theory Ahmed's work put into practice the lossy compression that Shannon's information theory had foreseen in the 1940s [1].
- Sound waves Audio coders such as mp3 use findings about human hearing to decide which sounds can be dropped [2].
as of 2026-11-11
Discoveries
The discoveries that made it possible: when, by whom and what was found
| Year | Who | What |
|---|---|---|
| 1948 | Claude Shannon | Founded information theory and showed that every channel has a maximum rate for sending information reliably [5] |
| 1974 | Nasir Ahmed and two collaborators | Published the discrete cosine transform [1] |
as of 2026-11-11
The science explained
The science behind it explained for a curious reader
Think of a photo as music played on a piano. A smooth sky is a low, slow note; the edge of a leaf or a strand of hair is a high, quick one. The discrete cosine transform works out which notes, at what loudness, make up each small block of the picture [1]. Most of the loudness sits in the low notes. The high notes are quiet, and the eye barely misses them if they are played roughly or not at all. So the coder writes down the low notes exactly and the high ones crudely, and the score shrinks to a fraction of the size [1]. Sound coders use the same idea with real notes, guided by what is known about hearing [2]: quiet sounds that louder ones hide from the ear are dropped.
as of 2026-11-11
Sources
- Nasir Ahmed: An Unsung Hero of Digital Media IEEE Spectrum (Willie D. Jones), 2024
- The mp3 History: Timeline Fraunhofer Institute for Integrated Circuits IIS
- Claude Shannon: Tinkerer, Prankster, and Father of Information Theory IEEE Spectrum (John Horgan), 2016