A method of digital signal processing

A method of digitally processing images and/or audio in the form of a plurality of data samples, each data sample having a value representative of a visual or audio property, the method comprising: identifying a first set of data samples having values outside a first predefined range of data values (B-C, fig 3); identifying a second set of data samples having values outside a second predefined range of data values (A-D, fig 3), the second predefined range being within the first predefined range; and applying a non-linear scaling algorithm to the data values of the first and second sets of data samples to generate a set of modified data samples having data values within the first predefined range (A-D, fig 4). The method may be applied to spatially or temporally spaced samples (10, fig 5). When applied to audio the technique is akin to compressing/limiting the sample.