Interferometry is a very powerful tool for making accurate measurements. However, conventional interferometric methods have some serious limitations. With two-beam interference, the accuracy with which the centre of a fringe can be located is limited by the sinusoidal intensity distribution in the fringe pattern. In addition, quantitative information on the shape of a surface is available only along the intensity minima and maxima. Data at other points can be obtained only by interpolation or by introducing a number of fringes across the field. Finally, it is often difficult to identify the sense of a slope and ambiguities can arise. This is particularly true when the fringes are irregular and unequally spaced. Various methods have been explored to solve these problems.' One of the earliest was the use of a television camera linked to a digital computer to store and process the intensity distribution in the fringes to locate the maxima and minima. A commonly used method is heterodyne interferometry, in which a frequency difference is introduced between the beams. However, the technique most widely used currently is digital phase-shifting interferometry.