Image compression method and apparatus

Controlling compression of image data by determining a desired quantisation level 41 based on a desired target number of bits 40, and selecting one a set of discrete quantisation levels based on the desired quantisation level 42. If a predicted number of bits arising from the selected quantisation level is below the desired target number of bits 44, providing information to an encoder to allow the encoder to encode image data with the selected quantisation level 46. Another compression control embodiment involves using a current weight for each area of an image to determine a desired target bitrate of the image area. Weight determination comprises normalising a number of bits following compression of a corresponding image area in a previous image by a total number of bits following compression of the previous image. Normalised weights derived from the normalised number of bits are combined with a previously used weight to create a current weight. Another compression control embodiment involves selecting one of a set of discrete quantisation levels, determining an average value of quantisation levels used for all image areas of a previous frame, and using the selected level and average level to determine an updated discrete quantisation level.