High layer syntax for temporal scalability
Disclosed are techniques for temporal scalable video coding. By including a temporal_switching_point_flag and/or a temporal_nesting_flag into a normative part of the scalable bitstream, reference picture management in decoder can be simplified, and media-aware network elements and decoder can easily identify those pictures from which onward a higher temporal layer can be added.
1. A method for decoding with a decoding device media encoded using scalable coding with temporal scalability at a pre-determined maximum temporal level, the method comprising:
decoding a single bit flag temporal_nesting_flag in a sequence parameter set that is indicative of a temporally nested structure of layers and pertains to all frames of a coded video sequence,
decoding a coded first picture of the coded video sequence with a first layer_id;
in response to the decoded temporal_nesting_flag equal to 1, removing at least one second picture of the coded video sequence from a reference picture list maintained by the decoder, the at least one second picture in the reference picture list having a second layer_id, wherein a value of the second layer_id is higher than a value of the first layer_id and wherein the second picture is not used for a prediction of the first picture or any pictures following the first picture in decoding order;
wherein the at least one second picture of the coded video sequence would be included in the reference picture list if the decoded temporal_nesting_flag were equal to 0;
wherein the temporal_nesting_flag set to one indicates for any three frames picA, picB, picC included in the coded video sequence, that picB is not used for reference of picA:
under a first condition that picB is of a temporal level lower or equal than the temporal level of picA, and
under a second condition that the temporal level of picC is lower than the temporal level of picB, and
under a third condition that picC follows picB in decoding order, and
under a fourth condition that picC precedes picA in decoding order; and
wherein picB is not included in a reference picture list of picA.
2. The method of claim 1 , wherein when the temporal_nesting_flag is equal to 0 the second picture is used for prediction of the first picture or a picture following the first picture in decoding order.