Multi-layer partitioned timed media data comprising timed samples is encapsulated. Each timed sample is encoded into a first layer and at least one second layer, at least one timed sample comprising at least one subsample, each subsample being encoded into the first layer or the at least one second layer. The method includes obtaining at least one subsample, belonging to said first layer, from at least one of the timed samples; creating a first track comprising the at least one obtained subsample; obtaining at least another subsample, belonging to said second layer, from the same one of the timed samples; creating a second track comprising said at least another obtained subsample; and generating descriptive metadata associated with the second track.