1、 Recommendation ITU-R BT.2073-0 (02/2015) Use of the high efficiency video coding (HEVC) standard for UHDTV and HDTV broadcasting BT Series Broadcasting service (television) ii Rec. ITU-R BT.2073-0 Foreword The role of the Radiocommunication Sector is to ensure the rational, equitable, efficient and
2、 economical use of the radio-frequency spectrum by all radiocommunication services, including satellite services, and carry out studies without limit of frequency range on the basis of which Recommendations are adopted. The regulatory and policy functions of the Radiocommunication Sector are perform
3、ed by World and Regional Radiocommunication Conferences and Radiocommunication Assemblies supported by Study Groups. Policy on Intellectual Property Right (IPR) ITU-R policy on IPR is described in the Common Patent Policy for ITU-T/ITU-R/ISO/IEC referenced in Annex 1 of Resolution ITU-R 1. Forms to
4、be used for the submission of patent statements and licensing declarations by patent holders are available from http:/www.itu.int/ITU-R/go/patents/en where the Guidelines for Implementation of the Common Patent Policy for ITU-T/ITU-R/ISO/IEC and the ITU-R patent information database can also be foun
5、d. Series of ITU-R Recommendations (Also available online at http:/www.itu.int/publ/R-REC/en) Series Title BO Satellite delivery BR Recording for production, archival and play-out; film for television BS Broadcasting service (sound) BT Broadcasting service (television) F Fixed service M Mobile, radi
6、odetermination, amateur and related satellite services P Radiowave propagation RA Radio astronomy RS Remote sensing systems S Fixed-satellite service SA Space applications and meteorology SF Frequency sharing and coordination between fixed-satellite and fixed service systems SM Spectrum management S
7、NG Satellite news gathering TF Time signals and frequency standards emissions V Vocabulary and related subjects Note: This ITU-R Recommendation was approved in English under the procedure detailed in Resolution ITU-R 1. Electronic Publication Geneva, 2015 ITU 2015 All rights reserved. No part of thi
8、s publication may be reproduced, by any means whatsoever, without written permission of ITU. Rec. ITU-R BT.2073-0 1 RECOMMENDATION ITU-R BT.2073-0 Use of the high efficiency video coding (HEVC) standard for UHDTV and HDTV broadcasting (Question ITU-R 12-3/6) (2015) Scope This Recommendation specifie
9、s the use of the High Efficiency Video Coding (HEVC) standard as per Recommendation ITU-T H.265 | ISO/IEC 23008-2 for ultra-high definition television (UHDTV) and high definition television (HDTV) broadcasting. Keywords UHDTV, HDTV, Sub layered coding, parallel coding sub bit stream The ITU Radiocom
10、munication Assembly, considering a) that there are applications where UHDTV and HDTV programme material is desired to be transported at a highly reduced bit rates with minimal visible degradation in quality; b) that Recommendation ITU-R BT.2020 specifies the parameters for a family of UHDTV video sy
11、stems; c) that Recommendation ITU-R BT.709 specifies the parameters for a family of HDTV video systems; d) that Recommendation ITU-T H.265 | ISO/IEC 23008-2 specifies the High Efficiency Video Coding (HEVC) standard that enables significantly improved compression performance relative to former stand
12、ards; e) that HEVC is increasingly adopted for various applications including broadcasting, recommends that when it is necessary to transport UHDTV and HDTV programme material at a highly reduced bit rate for broadcasting, the High Efficiency Video Coding (HEVC) standard specified in Recommendation
13、ITU-T H.265 | ISO/IEC 23008-2 should be used. Note 1 Annex 1 shows basic parameters for UHDTV and HDTV broadcasting using the HEVC standard. Note 2 Annex 2 shows a preferred coding scheme for temporal sub-layer coding of UHDTV at 120 or 100 Hz frame frequency using the HEVC standard. Note 3 Annex 3
14、shows a preferred coding scheme for interlaced video using the HEVC standard. Note 4 Annex 4 shows a preferred parallel coding scheme for the 7 680 4 320 format of UHDTV using the HEVC standard. 2 Rec. ITU-R BT.2073-0 Abbreviations CVS Coded Video Sequence DTS Decoding Time Stamp GOP Group of Pictur
15、es IRAP Intra Random Access Point PTS Presentation Time Stamp SEI Supplemental Enhancement Information Annex 1 (informative) Basic parameters for UHDTV and HDTV broadcasting using the HEVC standard This Annex shows basic parameters for UHDTV and HDTV broadcasting using the HEVC standard. TABLE 1 Bas
16、ic parameters for UHDTV and HDTV broadcasting using the HEVC standard Video format Level Profile Tier Suggested bit rate for broadcasting emission (Mbit/s) Spatial resolution Frame frequency (Hz) 7 680 4 320 120*, 100(1) 6.2 Main 10 Main 90-120 60*, 50 6.1 Main 10 Main 80-100 3 840 2 160 120*, 100(1
17、) 5.2 Main 10 Main 35-50 60*, 50 5.1 Main 10 Main 30-40 1 920 1 080 60*, 50 4.1 Main 10 or Main Main 10-15 30*, 25 (interlaced) 4.1(2) Main 10 or Main Main 10-15 * Those divided by 1.001 are also included. (1) The use of temporal sub-layer coding is detailed in Annex 2. (2) To allow coding at a suff
18、icient bit rate as needed, level 4.1 (the maximum bit rate is 20 Mbit/s) is preferred to level 4 (the maximum bit rate is 12 Mbit/s). Rec. ITU-R BT.2073-0 3 Annex 2 (informative) Preferred coding scheme for temporal sub-layer coding for UHDTV at 1201 or 100 Hz frame frequency using the HEVC standard
19、 This Annex shows a preferred coding scheme to achieve temporal sub-layer coding for UHDTV at 120 or 100 Hz frame frequency using the HEVC standard. Introduction The purpose of this preferred coding scheme is to enable a decoder with the decoding capability of Level 6.1 (or 5.1) bitstream for 60 or
20、50 Hz video to correctly decode the 60 or 50 Hz portion of a Level 6.2 (or 5.2) bitstream for 120 or 100 Hz video. Such decoding capability is realized by temporal sub-layer coding specified in the HEVC standard. To maximize the adaptability of a Level 6.1 (or 5.1) decoder to a Level 6.2 (or 5.2) te
21、mporal sub-layer coding bitstream, a further constraint on decoding order is introduced in such a way that the DTS/PTS value of an access unit in Level 6.1 (or 5.1) sub-bitstream can be applied both to Level 6.2 (or 5.2) bitstream decoding and Level 6.1 (or 5.1) sub-bitstream decoding. Temporal sub-
22、layer coding Every second frame of a 120 or 100 Hz video is encoded into an access unit of a sub-bitstream. All the other frames of the 120 or 100 Hz video are encoded into access units in a subset. A Level 6.1 (or 5.1) decoder decodes the sub-bitstream and outputs decoded frames with the frame freq
23、uency of 60 or 50 Hz. A Level 6.2 (or 5.2) decoder decodes both the sub-bitstream and the subset and outputs decoded frames with the frame frequency of 120 Hz. Constraint on decoding order It is mandated that the decoding order of each access unit in the sub-bitstream and each access unit in the sub
24、set shall be interleaved. That is, an access unit in the sub-bitstream is decoded immediately after an access unit in the subset, and vice versa. Figure 1 shows an example of the decoding order of an access unit in a Level 6.2 (or 5.2) temporal sub-layer coding bitstream. It is noted that there is n
25、o need to overwrite the values of au_cpb_removal_delay_minus1 and pic_dpb_output_delay of an access unit in the sub-bitstream for decoding the sub-bitstream in a Level 6.1 (or 5.1) decoder. That is, a nested picture timing SEI message is not needed. 1 Also includes 120/1.001 4 Rec. ITU-R BT.2073-0 F
26、IGURE 1 Constraint on decoding order for temporal sub-layer coding bitstream Annex 3 (informative) Preferred coding scheme for interlaced video using the HEVC standard This Annex shows a preferred coding scheme for interlaced video using the HEVC standard. Introduction The preferred coding scheme in
27、 this Annex utilizes the coding capability for interlaced video of the HEVC standard. That is, either frame-based coding or field-based coding is used in each CVS. If a CVS is encoded by field-based coding (i.e. field_seq_flag equal to 1), constraints described below are imposed. Otherwise, if a CVS
28、 is encoded by frame-based coding (i.e. field_seq_flag is equal to 0), no further constraints are imposed. BT .2073 -01012nT em po ra l IDD ec od in g o rd er o fac c es s un itA cc es s un it in12 0 or 1 00 H z su bs et O utp ut fr am e ra te: 12 0 or 1 00 H zG OP s tru ct ur eA cc es s un it in60
29、o r 5 0 H z su b- bi tst re amAc cess u nitN: Di spl ay o rderM: D ecodin g o rderIn L 6. 2 or 5 .2 de co de rIn L 6. 1 or 5 .1 de co de rD ec od in g o rd er o fac c es s un itO utp ut o rd er o fac c es s un itO utp ut fr am e ra te: 60 o r 5 0 H z(a u_cpb_ re mo val_ del ay_m inu s 1 + 1)of ac ce
30、s un itpic_ dp b_ ou tpu t_d el ayof ac ces un it(a u_cpb_ re mo val_ del ay_m inu s 1 + 1)of ac ces un itpic_ dp b_ ou tpu t_d el ayof ac ces un itDi spla yorde r0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16- 2 3 4 5 6 7 8 9 10 11 12 13 14 154 10 2 416 172 0 2 2 2 10 2 4 2 0 2 2 2NM1 3 5 7 9 11 13 15008
31、1124334255667781699101211111210131314141515160 842 140 1234 6 8 10 12 14 1661612105791113 150 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16- 2 4 6 8 10 12 144 10 4160 2 10 4 0 20 2 4 6 8 10 12 14 160 2 5 4 3 6 7 8 1 10 13 12 11 14 15 16 900814 256716912111013141530 5 3 7 1 13 11 15 9O utp ut o rd er o fac
32、c es s un itRec. ITU-R BT.2073-0 5 Constraint on GOP structure It is mandated that the first and second field pictures shall be encoded successively if the two fields are contained in the same frame. Figure 2 shows an example of the GOP structure according to the constraint of this Annex. It is note
33、d that any field in a frame can refer to any previously decoded fields in other frames. FIGURE 2 Constraint on GOP structure in field-based coding B T . 2 0 7 3 - 0 1012nT emp o ral IDD eco d i n g o rd er o facce s s u n i tA cces s u n i t i n1 2 0 o r 1 0 0 H z s u b s et O u t p u t frame rat e:
34、 1 2 0 o r 1 0 0 H zG O P s t ru ct u reA cces s u n i t i n6 0 o r 5 0 H z s u b -b i t s t reamA c c e s s u n i tN : D i s p l a y o r d e rM : D e c o d i n g o r d e rIn L 6 . 2 o r 5 . 2 d eco d erIn L 6 . 1 o r 5 . 1 d eco d erD eco d i n g o rd er o facce s s u n i tO u t p u t o rd er o fac
35、ce s s u n i tO u t p u t frame rat e: 6 0 o r 5 0 H z( a u _ c p b _ r e m o v a l _ d e l a y _ m i n u s 1 + 1 )o f a c c e s u n i tp i c _ d p b _ o u t p u t _ d e l a yo f a c c e s u n i t( a u _ c p b _ r e m o v a l _ d e l a y _ m i n u s 1 + 1 )o f a c c e s u n i tp i c _ d p b _ o u t
36、p u t _ d e l a yo f a c c e s u n i tD i s pl a yor de r0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16- 2 3 4 5 6 7 8 9 10 11 12 13 14 154 10 2 416 172 0 2 2 2 10 2 4 2 0 2 2 2NM1 3 5 7 9 11 13 150081124334255667781699101211111210131314141515160 842 140 1234 6 8 10 12 14 1661612105 791113 150 1 2 3 4 5 6
37、 7 8 9 10 11 12 13 14 15 16- 2 4 6 8 10 12 144 10 4160 2 10 4 0 20 2 4 6 8 10 12 14 160 2 5 4 3 6 7 8 1 10 13 12 11 14 15 16 900814 256716912111013141530 5 3 7 1 13 11 15 9Constraint on IRAP access unit Since the HEVC standard does not allow encoding a trailing picture access unit before any leading
38、 picture access unit that has an earlier display order than the trailing picture access unit, the following constraint is imposed to satisfy the constraint on the GOP structure described above. When a leading picture access unit appears in a bitstream, an IRAP picture access unit shall appear only a
39、t the beginning of a CVS. To have frequent random access points, multiple access units associated with a recovery point SEI message may be present in a CVS. In this case, it is encouraged to encode a CVS in such a way that recovery_poc_cnt and exact_match_flag of the recovery point SEI message can b
40、e set equal to 0 and 1, respectively. 6 Rec. ITU-R BT.2073-0 Annex 4 (informative) Preferred parallel coding scheme for the 7 680 4 320 format of UHDTV using the HEVC standard This Annex shows a preferred parallel coding scheme for the 7 680 4 320 format of UHDTV using the HEVC standard. Introductio
41、n Considering the latest technology trends in the broadcasting industry, it is strongly hoped that real-time HEVC decoder of 4K video will be implemented on a single-chip LSI in a short period. On the other hand, it is envisaged that it will take another 5-10 years to realize a single-chip LSI that
42、can decode 8K video in real-time. Therefore the structure of an HEVC bitstream of 8K video shall be defined so that it can be decoded by using multiple 4K HEVC LSIs. This preferred coding scheme for 8K video adopts a parallel coding scheme. An 8K picture is equally partitioned into four sub-pictures
43、. To minimize the loss in coding efficiency caused by the partitioning, sharing reference pictures among sub-pictures and enabling in-loop filters at the boundary of sub-pictures are mandated. Sub-picture partitioning An 8K picture is partitioned into four sub-pictures. Each sub-picture is encoded b
44、y each processing core as an independent slice segment with the parameters shown in Fig. 3. Each slice segment may be further portioned into multiple slices. FIGURE 3 Partitioning scheme of 8K picture into four sub-pictures B T .20 73- 0 3Inde pen dent sli ce se gm ent 0fi rst_sl ic e_se gm ent _i n
45、_pi c_fla g = 17 680 sam pl esInde pen dent sli ce se gm ent 1fi rst_sl ic e_se gm ent _i n_pi c_fla g = 0sl ice _seg m ent _ad dres s = 2 0 40 (64 64 C TB c ase)Inde pen dent sli ce se gm ent 3fi rst_sl ic e_se gm ent _i n_pi c_fla g = 0sl ice _seg m ent _ad dres s = 6 1 20 (64 64 C TB c ase)Inde p
46、en dent sli ce se gm ent 2fi rst_sl ic e_se gm ent _i n_pi c_fla g = 0sl ice _seg m ent _ad dres s = 4 0 80 (64 64 C TB c ase)1088samples1 056samples1 088samples1088samplesxy4320 samplesRec. ITU-R BT.2073-0 7 Constraints on parameters The constraints on parameters in Table 2 are applied. TABLE 2 Con
47、strains on parameters for slice-based sub-picture partitioning Parameter Constraint pic_width_in_luma_samples 7 680 pic_height_in_luma_samples 4 320 first_slice_segment_in_pic_flag slice_segment_address Values shown in Fig. 3 pps_loop_filter_across_slices_enabled_flag slice_loop_filter_across_slices
48、_enabled_flag 1 tiles_enabled_flag 0 NOTE Tile-based partition is not recommended since the vertical partitioning of tile causes significant loss in coding efficiency at typical scenes in programmes that have a large horizontal motion when each processing core shares a limited amount of reference sa
49、mples for motion compensation. The range of the vertical component of a motion vector that crosses a slice boundary It shall be constrained in such a way that any prediction block in an independent slice segment does not refer to samples in a different independent slice segment whose vertical position relative to the boundary of two independent slices is outside the range of (128, 128) for a luma sample and (64, 64) for a chroma sample (in the case of 4:2:0 chroma subsampling). See Fig. 4 for detailed explanation NOTE This constraint is introduced to reduc
