Data Compression by Applying First and Second Compression Methods to the Values of First and Second Variates (groups of values)

Compressing a multivariate dataset involves selecting a dataset comprising a plurality of variates (i.e. a set or group of values), applying a first compression method to the values of a first variate and applying a second compression method to the values of a second variate of the dataset. The second compression method compresses the second variate values relative to the variation of the corresponding first variate values. The compression of the second variate values may be inversely related to the variation of the corresponding first variate values, and the second variate values may be compressed relative to the uncompressed or compressed first variate values. The second variate values may be compressed to a range. The compressed second variate values may be interpolated to a set of value ranges, in which the interpolation may be linear. The dataset may represent image data, e.g. high dynamic range or hyperspectral data, and the first and second variates may comprise pixel amplitude (which may represent luminance) and frequency. The first compression method may comprise a tone mapping method which may be arranged to perform spatially uniform or varying compression.