A system and method for signaling and indicating track relationship information in media files provides a mechanism for signaling the information of layered track groups and the information of multiple description coding (MDC) track groups, as well as a mechanism for signaling track relationship information in an efficient way. In local playback or unicast streaming applications, to select an independently decodable track for a certain media type, the information of alternate track groups is first found via the track relation box, and one track is selected from an alternate track group for the media type. If stream switching is desired, the switching track group information is found via the track relation box. In multicast applications with scalable or MDC streams, the tracks in a layered or MDC group are found via the track relation box and selected among all the layered or MDC groups.