An audio/video reproducing apparatus comprises an activity detector operable to receive video signals representing images and to generate an activity signal indicative of an amount of activity within the sound/images represented by the audio/video signal. An image generator is coupled to the activity detector and is operable to generate sample images at temporal positions within the audio/video signal, which temporal positions are determined from the activity signal. The processing apparatus thereby provides a facility for automatically generating meta data from received audio/video signals. The meta data can be used to select the audio/video material.