In this study a maximised cross-correlation technique is applied to two very different experiments. The first is a laminar, slowly developing experiment where the main scales of the motion are comparable to the image size. Here extensive use of multiple passes with the technique and various filters are applied to replace erroneous measurements. The second experiment has both laminar and turbulent regions of flow, and in addition, many of the scales of motion are smaller than the correlation window size. Mean velocities are resolved in this application, as well as a qualitative impression of the turbulent fluctuations.