A video indexing system uses pictures representative of a recorded video program to assist a user in determining the contents of a recorded medium without having to view the program itself. The pictures preferably represent segments of the program which are spaced apart in time, enabling a relatively small number of such pictures to characterize a lengthy program, and are presented in separate windows on the screen of the same device used to display the video program. The pictures may include still or moving imagery. A viewer optionally may select a particular picture with a pointing device, to commence replay of the recorded program from that period in the program, or to recall stored audio information so as to assist in identifying the selection. The picture information may be stored on the same medium as that used to record the video program, or a different medium may be used.