Over the years there has been increasing interest in gold recovery from many different kinds of waste. Beneficiation processes, applied to gold recovery from old mine dumps, electronic scrap, dental scrap and jewellery waste have been extensively studied. To obtain good results, these processes need to start from a reliable physical characterisation of the occurrence of gold particles. This can be quite difficult due to the great heterogeneity of the matrix that holds the precious metal. The present work, applied to jewellery waste as a case study, shows that image processing techniques, applied to pictures obtained from optical microscopy or SEM, can offer good results. Nevertheless, information from chemical analyses, related to a size distribution, carried out by traditional wet-sieving, can be misinterpreted due to an overestimation of the influence of the associations between constituent particulate elements into waste.