A method of processing a sound field representation which can be used to provide a representation which has been modified by at least spatial processing. The method includes: the establishment of an initial representation of the sound field and the establishment of a representation which has been modified by filtering combinations corresponding to the processing operation. The method also includes a step which defines the processing operation and which includes: a step involving the establishment of a set of directivity functions, which is representative of the processing and which takes the form of a set of coefficients corresponding to the spherical harmonic decomposition of the functions; and a step in which the filtering combinations are determined from the decomposition so that another step involving the application of the defined filtering combinations can be performed.