.JPG

Introduction

JPEG is the most common format for photographs. It is named after the Joint Photographic Experts Group, whose 1992 standard (ITU-T T.81, ISO/IEC 10918-1) defines the compression. The compression is lossy: it divides the image into 8 by 8 blocks, keeps the detail the eye notices most and discards the rest, so files are small and the quality setting trades size against artefacts.

The standard defines the compressed data but not a complete file. The file around it comes from JFIF, the JPEG File Interchange Format, or from Exif, which cameras use to add the date, exposure, orientation and location. Both use the same extensions: .jpg, .jpeg, .jpe and .jfif.

Structure

A JPEG file is a sequence of segments, each starting with a marker: the byte FF and a code. Most markers are followed by a 2-byte big-endian length that counts itself but not the marker. The file starts with SOI (FF D8) and ends with EOI (FF D9).

Between them, application segments carry metadata: APP0 for JFIF and APP1 for Exif and XMP. Quantisation tables (DQT), the frame header (SOF0 for baseline, SOF2 for progressive) and Huffman tables (DHT) follow. The start of scan (SOS) is followed by the compressed image data, in which any FF byte is written as FF 00 so it cannot be mistaken for a marker.

Links