The encyclopedia · Engineering & Operations · Technical decision · 1992
JPEG compressed photos by deleting what eyes cannot see.
The 1992 standard split images into frequency blocks, threw away imperceptible detail, and made digital photos practical.
Joint Photographic Experts Group
the move
By the late 1980s, storing and transmitting photographic images was brutally expensive: an uncompressed color photo consumed hundreds of thousands of bytes. The ISO/IEC Joint Photographic Experts Group began work in 1986 on a common standard for still-image compression.
The result, JPEG (ISO/IEC 10918), was completed in 1992; the ITU adopted the identical text as Recommendation T.81 in September 1992. The JPEG committee describes the core coding system as the well-known Huffman-coded, DCT-based lossy image format.
The trick is perceptual: the transform concentrates the image's energy into a few coefficients, and quantization discards small differences that human vision cannot resolve. Because encoder and decoder use the same rules, any compliant device can read any compliant image.
why it works
- The DCT concentrates image energy so most coefficients can be discarded.
- Quantization matches the coding to human vision instead of mathematical exactness.
- A single international standard meant one file worked on every device.
- Lossy compression gave a large size cut that made digital photos usable.
what transfers
Optimize for the human perceiving the output, not the data itself: when the consumer cannot detect the difference, the wasted bits are cost with no value.
what came after
JPEG became the most widely used image format on the web and in cameras for two decades, and its success spawned the JPEG 2000 and JPEG XS families. The perceptual-coding trick it proved also underpins video formats such as MPEG.
references
- JPEG — Overview of JPEG 1
- ITU-T Recommendation T.81 (09/92) — Digital compression and coding of continuous-tone still images
spotted an error? The archive wants to know.