A personalized audio and/or video (AV) stream is created for a user. A first stream of AV data is selected by reference to predetermined user preferences. One or more further data streams having a predetermined relation to the first stream are selected. Portions of the first stream are identified and captured by reference to identifying predetermined indicia in a group of streams consisting of the first stream and the one or more further data streams, where the capturing comprises capturing one or more time stamps from the group of streams. The personalized AV stream, comprising a captured portion of the first stream and data selected by reference to the predetermined indicia from at least one of the further data streams is generated, where the data selected is based on each captured timestamp.