Method and system for generating an instantaneous decoding refresh (IDR) picture slice in an H.264/AVC compliant video data stream
Method for generating an IDR picture slice in an H.264/AVC stream, the method including finding an I-type picture and identifying it as a start picture, modifying and/or adding, in each slice header of the start picture, IDR-specific elements, removing, from each slice header of the start picture, non-IDR-specific elements, removing all reference and non-reference B-slices between the start picture and the next I-picture or P-picture, following the start picture, that is a frame or the first field, in decoding order, of a complementary field pair. Starting from the first picture after the removed slices, for each slice header executing: correcting frame_num so that they are valid with reference to the previous picture, correcting references, in the MMCO, as a function of the removed reference B-pictures, correcting references, found in the “Reference picture list reordering” structure, as a function of the removed reference B-pictures and restoring slice encapsulation for modified slices.
1. Method for generating an Instantaneous Decoding Refresh (IDR) picture slice in an H.264/AVC compliant video data stream, the method comprising the steps of:
Starting reception of the H.264/AVC compliant video data stream ( 201 ) by a data receiving block;
with a conversion unit Finding in the H.264/AVC compliant video data stream an I-type picture and identifying it as a start picture ( 202 );
Modifying and/or adding ( 203 ), in each slice header of the start picture, IDR specific syntax elements;
Removing ( 204 ), from each slice header of the start picture, non-IDR specific syntax elements;
Removing ( 205 ), from the H.264/AVC compliant video data stream, all reference and non-reference B slices between the start picture and the next I picture or P picture, following the start picture, that is either a frame or the first field, in decoding order, of a complementary field pair;
Starting from the first picture after the start picture, for each slice header ( 206 ), received from the H.264/AVC compliant video data stream, executing the following steps:
Correcting frame_num parameters ( 207 ) so that they are valid with respect to the previous picture;
Correcting picture references ( 208 ), found in the “Memory Management Control Operation” (MMCO) structure, as a function of the removed reference B pictures;
Correcting picture references ( 209 ), found in the “Reference picture list reordering” structure, as a function of the removed reference B pictures;
Restoring H.264/AVC slice encapsulation ( 210 ) for each modified slice; and
with an output unit, outputting an H.264/AVC compliant video data stream comprising an Instantaneous Decoding Refresh (IDR) picture slice.
2. The method according to claim 1 characterised in that the step of finding in the H.264/AVC compliant video data stream an I-type picture uses supplemental enhancement information (SEI) data and in particular recovery point SEI message semantics.
3. The method according to claim 1 characterised in that the IDR specific syntax elements are frame_num value set to 0 and picture_order_count value set to 0.
4. The method according to claim 3 characterised in that the IDR specific syntax elements are further no_output_of_prior_pics_flag, long_term_reference_flag.
5. The method according to claim 1 characterised in that the non-IDR specific syntax elements include adaptive_ref_pic_marking.
6. The method according to claim 1 characterised in that the step of correcting frame_num parameters so that they are valid with respect to the previous picture also includes correction of the picture_order_count parameter.
7. The method according to claim 1 characterised in that correcting picture references, found in the “Memory Management Control Operation” structure, as a function of the removed reference B pictures includes verifying whether in the “Memory Management Control Operation” entries, differentially encoded target short-term reference picture indices do not refer to the removed B pictures or pictures preceding the start picture in the decoding order.
8. The method according to claim 1 characterised in that MMCO correction is effected for the number of pictures specified by MaxFrameNum parameter.
9. The method according to claim 1 characterised in that correcting picture references, found in the “Memory Management Control Operation” loop, as a function of the removed reference B pictures includes:
verification ( 701 ) if memory_management_control_operation equals 1 or if the parameter of memory_management_control_operation equals 3;
when the foregoing requirement is true, checking ( 702 ), whether the referenced picture has a frame_num indicating a picture that precedes the next I or P picture following the start picture, in decoding order, that is either a frame or the first field of a complementary field pair;
when the foregoing requirement is true, checking ( 703 ) if the referenced picture is an IDR frame/picture 703 or a field of a complementary pair that contains the processed IDR picture;
when the foregoing requirement is true, correcting ( 705 ), the parameter of difference_of_pic_nums_minus1 such that:
corrected_difference_of_pic_nums_minus1=difference_of_pic_nums_minus1−number_of_dropped_reference_frames<<field_shift, where the field_shift value equals field_pic_flag of the currently processed picture;
else ( 705 ), removing the memory_management_control_operation from the processed picture.
10. The method according to claim 1 characterised in that correcting picture references ( 209 ), found in the “Reference picture list reordering” structure, as a function of the removed reference B pictures is effected for the number of pictures specified by MaxFrameNum parameter.
11. The method according to claim 1 characterised in that correcting picture references ( 209 ), found in the “Reference picture list reordering” structure, as a function of the removed reference B pictures includes correction of reference lists entries or their removal.
12. The method according to claim 1 characterised in that restoring H.264/AVC slice encapsulation ( 210 ) for each modified slice includes encapsulation at RBSP, NAL and Annex-B byte stream levels.
13. A non-transitory computer readable non-volatile memory storing computer-executable instructions performing all the steps of the computer-implemented method according to claim 1 when executed on a computer.
14. A system ( 901 ) for generating an Instantaneous Decoding Refresh (IDR) picture slice in an H.264/AVC compliant video data stream, the system comprising:
a data receiving block ( 910 ) for reception of the H.264/AVC compliant video data stream;
the system being characterized in that it further comprises:
an I-type picture conversion unit ( 921 ) configured to:
find in the H.264/AVC compliant video data stream an I-type picture and identifying it as a start picture;
modify and/or add, in each slice header of the start picture, IDR specific syntax elements;
remove, from each slice header of the start picture, non-IDR specific syntax elements;
a remove, from the H.264/AVC compliant video data stream, all reference and non-reference B slices between the start picture and the next I picture or P picture, following the start picture, that is either a frame or the first field, in decoding order, of a complementary field pair;
a pictures conversion unit ( 922 ) configured to, starting from the first picture after the removed slices, for each slice header, received from the H.264/AVC compliant video data stream, execute the following steps:
correct frame_num parameter so that they are valid with respect to the previous picture;
correct picture references, found in the “Memory Management Control Operation” (MMCO) structure, as a function of the removed reference B pictures;
correct picture references, found in the “Reference picture list reordering” structure, as a function of the removed reference B pictures;
restore H.264/AVC slice encapsulation for each modified slice; and
an output unit ( 923 ) for outputting an H.264/AVC compliant video data stream comprising an Instantaneous Decoding Refresh (IDR) picture slice.