The separation of figures from ground is achieved by the visual system based on differences in features such as luminance, colour, depth, orientation, texture or motion. Temporal information, namely phase differences between groups of spatially homogeneous points, can also lead to a clear discrimination of an object shape. The time difference needed to separate figure and ground is around 5 ms over a range of temporal frequencies between 1.3 Hz and 30 Hz, both for sharply focused and for blurred points. These short delays are clearly below the temporal integration time of the visual system. The results have implications for theories on temporal binding and object recognition.