idnits 2.17.1 draft-ietf-payload-rtp-h265-13.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- == There are 2 instances of lines with non-RFC6890-compliant IPv4 addresses in the document. If these are example addresses, they should be changed. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (June 3, 2015) is 3249 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) -- Looks like a reference, but probably isn't: '0' on line 1749 == Unused Reference: 'RFC6190' is defined on line 3841, but no explicit reference was found in the text == Unused Reference: 'I-D.ietf-mmusic-sdp-bundle-negotiation' is defined on line 3878, but no explicit reference was found in the text -- Possible downref: Non-RFC (?) normative reference: ref. 'HEVC' ** Obsolete normative reference: RFC 4566 (Obsoleted by RFC 8866) == Outdated reference: A later version (-11) exists of draft-ietf-avtcore-rtp-multi-stream-05 == Outdated reference: A later version (-54) exists of draft-ietf-mmusic-sdp-bundle-negotiation-02 == Outdated reference: A later version (-08) exists of draft-ietf-avtext-rtp-grouping-taxonomy-02 -- Obsolete informational reference (is this intentional?): RFC 2326 (Obsoleted by RFC 7826) -- Obsolete informational reference (is this intentional?): RFC 5117 (Obsoleted by RFC 7667) Summary: 1 error (**), 0 flaws (~~), 7 warnings (==), 5 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 1 Network Working Group Y.-K. Wang 2 Internet Draft Qualcomm 3 Intended status: Standards track Y. Sanchez 4 Expires: December 2015 T. Schierl 5 Fraunhofer HHI 6 S. Wenger 7 Vidyo 8 M. M. Hannuksela 9 Nokia 10 June 3, 2015 12 RTP Payload Format for High Efficiency Video Coding 13 draft-ietf-payload-rtp-h265-13.txt 15 Abstract 17 This memo describes an RTP payload format for the video coding 18 standard ITU-T Recommendation H.265 and ISO/IEC International 19 Standard 23008-2, both also known as High Efficiency Video Coding 20 (HEVC) and developed by the Joint Collaborative Team on Video 21 Coding (JCT-VC). The RTP payload format allows for packetization 22 of one or more Network Abstraction Layer (NAL) units in each RTP 23 packet payload, as well as fragmentation of a NAL unit into 24 multiple RTP packets. Furthermore, it supports transmission of 25 an HEVC bitstream over a single as well as multiple RTP streams. 26 When multiple RTP streams are used, a single or multiple 27 transports may be utilized. The payload format has wide 28 applicability in videoconferencing, Internet video streaming, and 29 high bit-rate entertainment-quality video, among others. 31 Status of this Memo 33 This Internet-Draft is submitted to IETF in full conformance with 34 the provisions of BCP 78 and BCP 79. 36 Internet-Drafts are working documents of the Internet Engineering 37 Task Force (IETF), its areas, and its working groups. Note that 38 other groups may also distribute working documents as Internet- 39 Drafts. 41 Internet-Drafts are draft documents valid for a maximum of six 42 months and may be updated, replaced, or obsoleted by other 43 documents at any time. It is inappropriate to use Internet- 44 Drafts as reference material or to cite them other than as "work 45 in progress." 47 The list of current Internet-Drafts can be accessed at 48 http://www.ietf.org/ietf/1id-abstracts.txt. 50 The list of Internet-Draft Shadow Directories can be accessed at 51 http://www.ietf.org/shadow.html. 53 This Internet-Draft will expire on December 3, 2015. 55 Copyright and License Notice 57 Copyright (c) 2015 IETF Trust and the persons identified as the 58 document authors. All rights reserved. 60 This document is subject to BCP 78 and the IETF Trust's Legal 61 Provisions Relating to IETF Documents 62 (http://trustee.ietf.org/license-info) in effect on the date of 63 publication of this document. Please review these documents 64 carefully, as they describe your rights and restrictions with 65 respect to this document. Code Components extracted from this 66 document must include Simplified BSD License text as described in 67 Section 4.e of the Trust Legal Provisions and are provided 68 without warranty as described in the Simplified BSD License. 70 Table of Contents 72 Abstract.........................................................1 73 Status of this Memo..............................................1 74 Table of Contents................................................3 75 1 Introduction...................................................5 76 1.1 Overview of the HEVC Codec................................5 77 1.1.1 Coding-Tool Features.................................5 78 1.1.2 Systems and Transport Interfaces.....................7 79 1.1.3 Parallel Processing Support.........................14 80 1.1.4 NAL Unit Header.....................................16 81 1.2 Overview of the Payload Format...........................18 82 2 Conventions...................................................18 83 3 Definitions and Abbreviations.................................19 84 3.1 Definitions..............................................19 85 3.1.1 Definitions from the HEVC Specification.............19 86 3.1.2 Definitions Specific to This Memo...................21 87 3.2 Abbreviations............................................23 88 4 RTP Payload Format............................................25 89 4.1 RTP Header Usage.........................................25 90 4.2 Payload Header Usage.....................................27 91 4.3 Transmission Modes.......................................28 92 4.4 Payload Structures.......................................29 93 4.4.1 Single NAL Unit Packets.............................29 94 4.4.2 Aggregation Packets (APs)...........................30 95 4.4.3 Fragmentation Units (FUs)...........................35 96 4.4.4 PACI packets........................................38 97 4.4.4.1 Reasons for the PACI rules (informative).......41 98 4.4.4.2 PACI extensions (Informative)..................42 99 4.5 Temporal Scalability Control Information.................43 100 4.6 Decoding Order Number....................................45 101 5 Packetization Rules...........................................47 102 6 De-packetization Process......................................48 103 7 Payload Format Parameters.....................................50 104 7.1 Media Type Registration..................................51 105 7.2 SDP Parameters...........................................76 106 7.2.1 Mapping of Payload Type Parameters to SDP...........77 107 7.2.2 Usage with SDP Offer/Answer Model...................78 108 7.2.3 Usage in Declarative Session Descriptions...........87 109 7.2.4 Parameter Sets Considerations.......................88 110 7.2.5 Dependency Signaling in Multi-Stream Mode...........89 111 8 Use with Feedback Messages....................................89 112 8.1 Picture Loss Indication (PLI)............................89 113 8.2 Slice Loss Indication (SLI)..............................90 114 8.3 Reference Picture Selection Indication (RPSI)............91 115 8.4 Full Intra Request (FIR).................................92 116 9 Security Considerations.......................................92 117 10 Congestion Control...........................................93 118 11 IANA Consideration...........................................95 119 12 Acknowledgements.............................................95 120 13 References...................................................95 121 13.1 Normative References....................................95 122 13.2 Informative References..................................97 123 14 Authors' Addresses...........................................99 125 1 Introduction 127 1.1 Overview of the HEVC Codec 129 High Efficiency Video Coding [HEVC], formally known as ITU-T 130 Recommendation H.265 and ISO/IEC International Standard 23008-2 131 was ratified by ITU-T in April 2013 and reportedly provides 132 significant coding efficiency gains over H.264 [H.264]. 134 As both H.264 [H.264] and its RTP payload format [RFC6184] are 135 widely deployed and generally known in the relevant implementer 136 communities, frequently only the differences between those two 137 specifications are highlighted in non-normative, explanatory 138 parts of this memo. Basic familiarity with both specifications 139 is assumed for those parts. However, the normative parts of this 140 memo do not require study of H.264 or its RTP payload format. 142 H.264 and HEVC share a similar hybrid video codec design. 143 Conceptually, both technologies include a video coding layer 144 (VCL), which is often used to refer to the coding-tool features, 145 and a network abstraction layer (NAL), which is often used to 146 refer to the systems and transport interface aspects of the 147 codecs. 149 1.1.1 Coding-Tool Features 151 Similarly to earlier hybrid-video-coding-based standards, 152 including H.264, the following basic video coding design is 153 employed by HEVC. A prediction signal is first formed either by 154 intra or motion compensated prediction, and the residual (the 155 difference between the original and the prediction) is then 156 coded. The gains in coding efficiency are achieved by 157 redesigning and improving almost all parts of the codec over 158 earlier designs. In addition, HEVC includes several tools to 159 make the implementation on parallel architectures easier. Below 160 is a summary of HEVC coding-tool features. 162 Quad-tree block and transform structure 164 One of the major tools that contribute significantly to the 165 coding efficiency of HEVC is the usage of flexible coding blocks 166 and transforms, which are defined in a hierarchical quad-tree 167 manner. Unlike H.264, where the basic coding block is a 168 macroblock of fixed size 16x16, HEVC defines a Coding Tree Unit 169 (CTU) of a maximum size of 64x64. Each CTU can be divided into 170 smaller units in a hierarchical quad-tree manner and can 171 represent smaller blocks down to size 4x4. Similarly, the 172 transforms used in HEVC can have different sizes, starting from 173 4x4 and going up to 32x32. Utilizing large blocks and transforms 174 contribute to the major gain of HEVC, especially at high 175 resolutions. 177 Entropy coding 179 HEVC uses a single entropy coding engine, which is based on 180 Context Adaptive Binary Arithmetic Coding (CABAC) [CABAC], 181 whereas H.264 uses two distinct entropy coding engines. CABAC in 182 HEVC shares many similarities with CABAC of H.264, but contains 183 several improvements. Those include improvements in coding 184 efficiency and lowered implementation complexity, especially for 185 parallel architectures. 187 In-loop filtering 189 H.264 includes an in-loop adaptive deblocking filter, where the 190 blocking artifacts around the transform edges in the 191 reconstructed picture are smoothed to improve the picture quality 192 and compression efficiency. In HEVC, a similar deblocking filter 193 is employed but with somewhat lower complexity. In addition, 194 pictures undergo a subsequent filtering operation called Sample 195 Adaptive Offset (SAO), which is a new design element in HEVC. 196 SAO basically adds a pixel-level offset in an adaptive manner and 197 usually acts as a de-ringing filter. It is observed that SAO 198 improves the picture quality, especially around sharp edges 199 contributing substantially to visual quality improvements of 200 HEVC. 202 Motion prediction and coding 204 There have been a number of improvements in this area that are 205 summarized as follows. The first category is motion merge and 206 advanced motion vector prediction (AMVP) modes. The motion 207 information of a prediction block can be inferred from the 208 spatially or temporally neighboring blocks. This is similar to 209 the DIRECT mode in H.264 but includes new aspects to incorporate 210 the flexible quad-tree structure and methods to improve the 211 parallel implementations. In addition, the motion vector 212 predictor can be signaled for improved efficiency. The second 213 category is high-precision interpolation. The interpolation 214 filter length is increased to 8-tap from 6-tap, which improves 215 the coding efficiency but also comes with increased complexity. 216 In addition, the interpolation filter is defined with higher 217 precision without any intermediate rounding operations to further 218 improve the coding efficiency. 220 Intra prediction and intra coding 222 Compared to 8 intra prediction modes in H.264, HEVC supports 223 angular intra prediction with 33 directions. This increased 224 flexibility improves both objective coding efficiency and visual 225 quality as the edges can be better predicted and ringing 226 artifacts around the edges can be reduced. In addition, the 227 reference samples are adaptively smoothed based on the prediction 228 direction. To avoid contouring artifacts a new interpolative 229 prediction generation is included to improve the visual quality. 230 Furthermore, discrete sine transform (DST) is utilized instead of 231 traditional discrete cosine transform (DCT) for 4x4 intra 232 transform blocks. 234 Other coding-tool features 236 HEVC includes some tools for lossless coding and efficient screen 237 content coding, such as skipping the transform for certain 238 blocks. These tools are particularly useful for example when 239 streaming the user-interface of a mobile device to a large 240 display. 242 1.1.2 Systems and Transport Interfaces 244 HEVC inherited the basic systems and transport interfaces 245 designs, such as the NAL-unit-based syntax structure, the 246 hierarchical syntax and data unit structure from sequence-level 247 parameter sets, multi-picture-level or picture-level parameter 248 sets, slice-level header parameters, lower-level parameters, the 249 supplemental enhancement information (SEI) message mechanism, the 250 hypothetical reference decoder (HRD) based video buffering model, 251 and so on. In the following, a list of differences in these 252 aspects compared to H.264 is summarized. 254 Video parameter set 256 A new type of parameter set, called video parameter set (VPS), 257 was introduced. For the first (2013) version of [HEVC], the 258 video parameter set NAL unit is required to be available prior to 259 its activation, while the information contained in the video 260 parameter set is not necessary for operation of the decoding 261 process. For future HEVC extensions, such as the 3D or scalable 262 extensions, the video parameter set is expected to include 263 information necessary for operation of the decoding process, e.g. 264 decoding dependency or information for reference picture set 265 construction of enhancement layers. The VPS provides a "big 266 picture" of a bitstream, including what types of operation points 267 are provided, the profile, tier, and level of the operation 268 points, and some other high-level properties of the bitstream 269 that can be used as the basis for session negotiation and content 270 selection, etc. (see Section 7.1). 272 Profile, tier and level 274 The profile, tier and level syntax structure that can be included 275 in both VPS and sequence parameter set (SPS) includes 12 bytes of 276 data to describe the entire bitstream (including all temporally 277 scalable layers, which are referred to as sub-layers in the HEVC 278 specification), and can optionally include more profile, tier and 279 level information pertaining to individual temporally scalable 280 layers. The profile indicator indicates the "best viewed as" 281 profile when the bitstream conforms to multiple profiles, similar 282 to the major brand concept in the ISO base media file format 283 (ISOBMFF) [ISOBMFF] and file formats derived based on ISOBMFF, 284 such as the 3GPP file format [3GPPFF]. The profile, tier and 285 level syntax structure also includes indications such as 1) 286 whether the bitstream is free of frame-packed content, 2) whether 287 the bitstream is free of interlaced source content, and 3) 288 whether the bitstream is free of field pictures. When the answer 289 is yes for both 2) and 3), the bitstream contains only frame 290 pictures of progressive source. Based on these indications, 291 clients/players without support of post-processing 292 functionalities for handling of frame-packed, interlaced source 293 content or field pictures can reject those bitstreams that 294 contain such pictures. 296 Bitstream and elementary stream 298 HEVC includes a definition of an elementary stream, which is new 299 compared to H.264. An elementary stream consists of a sequence 300 of one or more bitstreams. An elementary stream that consists of 301 two or more bitstreams has typically been formed by splicing 302 together two or more bitstreams (or parts thereof). When an 303 elementary stream contains more than one bitstream, the last NAL 304 unit of the last access unit of a bitstream (except the last 305 bitstream in the elementary stream) must contain an end of 306 bitstream NAL unit and the first access unit of the subsequent 307 bitstream must be an intra random access point (IRAP) access 308 unit. This IRAP access unit may be a clean random access (CRA), 309 broken link access (BLA), or instantaneous decoding refresh (IDR) 310 access unit. 312 Random access support 314 HEVC includes signaling in the NAL unit header, through NAL unit 315 types, of IRAP pictures beyond IDR pictures. Three types of IRAP 316 pictures, namely IDR, CRA and BLA pictures are supported, wherein 317 IDR pictures are conventionally referred to as closed group-of- 318 pictures (closed-GOP) random access points, and CRA and BLA 319 pictures are those conventionally referred to as open-GOP random 320 access points. BLA pictures usually originate from splicing of 321 two bitstreams or part thereof at a CRA picture, e.g. during 322 stream switching. To enable better systems usage of IRAP 323 pictures, altogether six different NAL units are defined to 324 signal the properties of the IRAP pictures, which can be used to 325 better match the stream access point (SAP) types as defined in 326 the ISOBMFF [ISOBMFF], which are utilized for random access 327 support in both 3GP-DASH [3GPDASH] and MPEG DASH [MPEGDASH]. 328 Pictures following an IRAP picture in decoding order and 329 preceding the IRAP picture in output order are referred to as 330 leading pictures associated with the IRAP picture. There are two 331 types of leading pictures, namely random access decodable leading 332 (RADL) pictures and random access skipped leading (RASL) 333 pictures. RADL pictures are decodable when the decoding started 334 at the associated IRAP picture, and RASL pictures are not 335 decodable when the decoding started at the associated IRAP 336 picture and are usually discarded. HEVC provides mechanisms to 337 enable the specification of conformance of bitstreams with RASL 338 pictures being discarded, thus to provide a standard-compliant 339 way to enable systems components to discard RASL pictures when 340 needed. 342 Temporal scalability support 344 HEVC includes an improved support of temporal scalability, by 345 inclusion of the signaling of TemporalId in the NAL unit header, 346 the restriction that pictures of a particular temporal sub-layer 347 cannot be used for inter prediction reference by pictures of a 348 lower temporal sub-layer, the sub-bitstream extraction process, 349 and the requirement that each sub-bitstream extraction output be 350 a conforming bitstream. Media-aware network elements (MANEs) can 351 utilize the TemporalId in the NAL unit header for stream 352 adaptation purposes based on temporal scalability. 354 Temporal sub-layer switching support 356 HEVC specifies, through NAL unit types present in the NAL unit 357 header, the signaling of temporal sub-layer access (TSA) and 358 stepwise temporal sub-layer access (STSA). A TSA picture and 359 pictures following the TSA picture in decoding order do not use 360 pictures prior to the TSA picture in decoding order with 361 TemporalId greater than or equal to that of the TSA picture for 362 inter prediction reference. A TSA picture enables up-switching, 363 at the TSA picture, to the sub-layer containing the TSA picture 364 or any higher sub-layer, from the immediately lower sub-layer. 365 An STSA picture does not use pictures with the same TemporalId as 366 the STSA picture for inter prediction reference. Pictures 367 following an STSA picture in decoding order with the same 368 TemporalId as the STSA picture do not use pictures prior to the 369 STSA picture in decoding order with the same TemporalId as the 370 STSA picture for inter prediction reference. An STSA picture 371 enables up-switching, at the STSA picture, to the sub-layer 372 containing the STSA picture, from the immediately lower sub- 373 layer. 375 Sub-layer reference or non-reference pictures 377 The concept and signaling of reference/non-reference pictures in 378 HEVC are different from H.264. In H.264, if a picture may be 379 used by any other picture for inter prediction reference, it is a 380 reference picture; otherwise it is a non-reference picture, and 381 this is signaled by two bits in the NAL unit header. In HEVC, a 382 picture is called a reference picture only when it is marked as 383 "used for reference". In addition, the concept of sub-layer 384 reference picture was introduced. If a picture may be used by 385 another other picture with the same TemporalId for inter 386 prediction reference, it is a sub-layer reference picture; 387 otherwise it is a sub-layer non-reference picture. Whether a 388 picture is a sub-layer reference picture or sub-layer non- 389 reference picture is signaled through NAL unit type values. 391 Extensibility 393 Besides the TemporalId in the NAL unit header, HEVC also includes 394 the signaling of a six-bit layer ID in the NAL unit header, which 395 must be equal to 0 for a single-layer bitstream. Extension 396 mechanisms have been included in VPS, SPS, PPS, SEI NAL unit, 397 slice headers, and so on. All these extension mechanisms enable 398 future extensions in a backward compatible manner, such that 399 bitstreams encoded according to potential future HEVC extensions 400 can be fed to then-legacy decoders (e.g. HEVC version 1 decoders) 401 and the then-legacy decoders can decode and output the base layer 402 bitstream. 404 Bitstream extraction 406 HEVC includes a bitstream extraction process as an integral part 407 of the overall decoding process, as well as specification of the 408 use of the bitstream extraction process in description of 409 bitstream conformance tests as part of the hypothetical reference 410 decoder (HRD) specification. 412 Reference picture management 414 The reference picture management of HEVC, including reference 415 picture marking and removal from the decoded picture buffer (DPB) 416 as well as reference picture list construction (RPLC), differs 417 from that of H.264. Instead of the sliding window plus adaptive 418 memory management control operation (MMCO) based reference 419 picture marking mechanism in H.264, HEVC specifies a reference 420 picture set (RPS) based reference picture management and marking 421 mechanism, and the RPLC is consequently based on the RPS 422 mechanism. A reference picture set consists of a set of 423 reference pictures associated with a picture, consisting of all 424 reference pictures that are prior to the associated picture in 425 decoding order, that may be used for inter prediction of the 426 associated picture or any picture following the associated 427 picture in decoding order. The reference picture set consists of 428 five lists of reference pictures; RefPicSetStCurrBefore, 429 RefPicSetStCurrAfter, RefPicSetStFoll, RefPicSetLtCurr and 430 RefPicSetLtFoll. RefPicSetStCurrBefore, RefPicSetStCurrAfter and 431 RefPicSetLtCurr contain all reference pictures that may be used 432 in inter prediction of the current picture and that may be used 433 in inter prediction of one or more of the pictures following the 434 current picture in decoding order. RefPicSetStFoll and 435 RefPicSetLtFoll consist of all reference pictures that are not 436 used in inter prediction of the current picture but may be used 437 in inter prediction of one or more of the pictures following the 438 current picture in decoding order. RPS provides an "intra-coded" 439 signaling of the DPB status, instead of an "inter-coded" 440 signaling, mainly for improved error resilience. The RPLC 441 process in HEVC is based on the RPS, by signaling an index to an 442 RPS subset for each reference index; this process is simpler than 443 the RPLC process in H.264. 445 Ultra low delay support 447 HEVC specifies a sub-picture-level HRD operation, for support of 448 the so-called ultra-low delay. The mechanism specifies a 449 standard-compliant way to enable delay reduction below one 450 picture interval. Sub-picture-level coded picture buffer (CPB) 451 and DPB parameters may be signaled, and utilization of these 452 information for the derivation of CPB timing (wherein the CPB 453 removal time corresponds to decoding time) and DPB output timing 454 (display time) is specified. Decoders are allowed to operate the 455 HRD at the conventional access-unit-level, even when the sub- 456 picture-level HRD parameters are present. 458 New SEI messages 460 HEVC inherits many H.264 SEI messages with changes in syntax 461 and/or semantics making them applicable to HEVC. Additionally, 462 there are a few new SEI messages reviewed briefly in the 463 following paragraphs. 465 The display orientation SEI message informs the decoder of a 466 transformation that is recommended to be applied to the cropped 467 decoded picture prior to display, such that the pictures can be 468 properly displayed, e.g. in an upside-up manner. 470 The structure of pictures SEI message provides information on the 471 NAL unit types, picture order count values, and prediction 472 dependencies of a sequence of pictures. The SEI message can be 473 used for example for concluding what impact a lost picture has on 474 other pictures. 476 The decoded picture hash SEI message provides a checksum derived 477 from the sample values of a decoded picture. It can be used for 478 detecting whether a picture was correctly received and decoded. 480 The active parameter sets SEI message includes the IDs of the 481 active video parameter set and the active sequence parameter set 482 and can be used to activate VPSs and SPSs. In addition, the SEI 483 message includes the following indications: 1) An indication of 484 whether "full random accessibility" is supported (when supported, 485 all parameter sets needed for decoding of the remaining of the 486 bitstream when random accessing from the beginning of the current 487 CVS by completely discarding all access units earlier in decoding 488 order are present in the remaining bitstream and all coded 489 pictures in the remaining bitstream can be correctly decoded); 2) 490 An indication of whether there is no parameter set within the 491 current CVS that updates another parameter set of the same type 492 preceding in decoding order. An update of a parameter set refers 493 to the use of the same parameter set ID but with some other 494 parameters changed. If this property is true for all CVSs in the 495 bitstream, then all parameter sets can be sent out-of-band before 496 session start. 498 The decoding unit information SEI message provides coded picture 499 buffer removal delay information for a decoding unit. The 500 message can be used in very-low-delay buffering operations. 502 The region refresh information SEI message can be used together 503 with the recovery point SEI message (present in both H.264 and 504 HEVC) for improved support of gradual decoding refresh. This 505 supports random access from inter-coded pictures, wherein 506 complete pictures can be correctly decoded or recovered after an 507 indicated number of pictures in output/display order. 509 1.1.3 Parallel Processing Support 511 The reportedly significantly higher encoding computational demand 512 of HEVC over H.264, in conjunction with the ever increasing video 513 resolution (both spatially and temporally) required by the 514 market, led to the adoption of VCL coding tools specifically 515 targeted to allow for parallelization on the sub-picture level. 516 That is, parallelization occurs, at the minimum, at the 517 granularity of an integer number of CTUs. The targets for this 518 type of high-level parallelization are multicore CPUs and DSPs as 519 well as multiprocessor systems. In a system design, to be 520 useful, these tools require signaling support, which is provided 521 in Section 7 of this memo. This section provides a brief 522 overview of the tools available in [HEVC]. 524 Many of the tools incorporated in HEVC were designed keeping in 525 mind the potential parallel implementations in multi-core/multi- 526 processor architectures. Specifically, for parallelization, four 527 picture partition strategies are available. 529 Slices are segments of the bitstream that can be reconstructed 530 independently from other slices within the same picture (though 531 there may still be interdependencies through loop filtering 532 operations). Slices are the only tool that can be used for 533 parallelization that is also available, in virtually identical 534 form, in H.264. Slices based parallelization does not require 535 much inter-processor or inter-core communication (except for 536 inter-processor or inter-core data sharing for motion 537 compensation when decoding a predictively coded picture, which is 538 typically much heavier than inter-processor or inter-core data 539 sharing due to in-picture prediction), as slices are designed to 540 be independently decodable. However, for the same reason, slices 541 can require some coding overhead. Further, slices (in contrast 542 to some of the other tools mentioned below) also serve as the key 543 mechanism for bitstream partitioning to match Maximum Transfer 544 Unit (MTU) size requirements, due to the in-picture independence 545 of slices and the fact that each regular slice is encapsulated in 546 its own NAL unit. In many cases, the goal of parallelization and 547 the goal of MTU size matching can place contradicting demands to 548 the slice layout in a picture. The realization of this situation 549 led to the development of the more advanced tools mentioned 550 below. 552 Dependent slice segments allow for fragmentation of a coded slice 553 into fragments at CTU boundaries without breaking any in-picture 554 prediction mechanism. They are complementary to the 555 fragmentation mechanism described in this memo in that they need 556 the cooperation of the encoder. As a dependent slice segment 557 necessarily contains an integer number of CTUs, a decoder using 558 multiple cores operating on CTUs can process a dependent slice 559 segment without communicating parts of the slice segment's 560 bitstream to other cores. Fragmentation, as specified in this 561 memo, in contrast, does not guarantee that a fragment contains an 562 integer number of CTUs. 564 In wavefront parallel processing (WPP), the picture is 565 partitioned into rows of CTUs. Entropy decoding and prediction 566 are allowed to use data from CTUs in other partitions. Parallel 567 processing is possible through parallel decoding of CTU rows, 568 where the start of the decoding of a row is delayed by two CTUs, 569 so to ensure that data related to a CTU above and to the right of 570 the subject CTU is available before the subject CTU is being 571 decoded. Using this staggered start (which appears like a 572 wavefront when represented graphically), parallelization is 573 possible with up to as many processors/cores as the picture 574 contains CTU rows. 576 Because in-picture prediction between neighboring CTU rows within 577 a picture is allowed, the required inter-processor/inter-core 578 communication to enable in-picture prediction can be substantial. 579 The WPP partitioning does not result in the creation of more NAL 580 units compared to when it is not applied, thus WPP cannot be used 581 for MTU size matching, though slices can be used in combination 582 for that purpose. 584 Tiles define horizontal and vertical boundaries that partition a 585 picture into tile columns and rows. The scan order of CTUs is 586 changed to be local within a tile (in the order of a CTU raster 587 scan of a tile), before decoding the top-left CTU of the next 588 tile in the order of tile raster scan of a picture. Similar to 589 slices, tiles break in-picture prediction dependencies (including 590 entropy decoding dependencies). However, they do not need to be 591 included into individual NAL units (same as WPP in this regard), 592 hence tiles cannot be used for MTU size matching, though slices 593 can be used in combination for that purpose. Each tile can be 594 processed by one processor/core, and the inter-processor/inter- 595 core communication required for in-picture prediction between 596 processing units decoding neighboring tiles is limited to 597 conveying the shared slice header in cases a slice is spanning 598 more than one tile, and loop filtering related sharing of 599 reconstructed samples and metadata. Insofar, tiles are less 600 demanding in terms of inter-processor communication bandwidth 601 compared to WPP due to the in-picture independence between two 602 neighboring partitions. 604 1.1.4 NAL Unit Header 606 HEVC maintains the NAL unit concept of H.264 with modifications. 607 HEVC uses a two-byte NAL unit header, as shown in Figure 1. The 608 payload of a NAL unit refers to the NAL unit excluding the NAL 609 unit header. 611 +---------------+---------------+ 612 |0|1|2|3|4|5|6|7|0|1|2|3|4|5|6|7| 613 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 614 |F| Type | LayerId | TID | 615 +-------------+-----------------+ 617 Figure 1 The structure of HEVC NAL unit header 619 The semantics of the fields in the NAL unit header are as 620 specified in [HEVC] and described briefly below for convenience. 621 In addition to the name and size of each field, the corresponding 622 syntax element name in [HEVC] is also provided. 624 F: 1 bit 625 forbidden_zero_bit. Required to be zero in [HEVC]. Note that 626 the inclusion of this bit in the NAL unit header was to enable 627 transport of HEVC video over MPEG-2 transport systems 628 (avoidance of start code emulations) [MPEG2S]. In the context 629 of this memo, the value 1 may be used to indicate a syntax 630 violation, e.g. for a NAL unit resulted from aggregating a 631 number of fragmented units of a NAL unit but missing the last 632 fragment, as described in Section 4.4.3. 634 Type: 6 bits 635 nal_unit_type. This field specifies the NAL unit type as 636 defined in Table 7-1 of [HEVC]. If the most significant bit 637 of this field of a NAL unit is equal to 0 (i.e. the value of 638 this field is less than 32), the NAL unit is a VCL NAL unit. 639 Otherwise, the NAL unit is a non-VCL NAL unit. For a 640 reference of all currently defined NAL unit types and their 641 semantics, please refer to Section 7.4.1 in [HEVC]. 643 LayerId: 6 bits 644 nuh_layer_id. Required to be equal to zero in [HEVC]. It is 645 anticipated that in future scalable or 3D video coding 646 extensions of this specification, this syntax element will be 647 used to identify additional layers that may be present in the 648 CVS, wherein a layer may be, e.g. a spatial scalable layer, a 649 quality scalable layer, a texture view, or a depth view. 651 TID: 3 bits 652 nuh_temporal_id_plus1. This field specifies the temporal 653 identifier of the NAL unit plus 1. The value of TemporalId is 654 equal to TID minus 1. A TID value of 0 is illegal to ensure 655 that there is at least one bit in the NAL unit header equal to 656 1, so to enable independent considerations of start code 657 emulations in the NAL unit header and in the NAL unit payload 658 data. 660 1.2 Overview of the Payload Format 662 This payload format defines the following processes required for 663 transport of HEVC coded data over RTP [RFC3550]: 665 o Usage of RTP header with this payload format 667 o Packetization of HEVC coded NAL units into RTP packets using 668 three types of payload structures, namely single NAL unit 669 packet, aggregation packet, and fragment unit 671 o Transmission of HEVC NAL units of the same bitstream within a 672 single RTP stream or multiple RTP streams (within one or more 673 RTP sessions), where within an RTP stream transmission of NAL 674 units may be either non-interleaved (i.e. the transmission 675 order of NAL units is the same as their decoding order) or 676 interleaved (i.e. the transmission order of NAL units is 677 different from their decoding order) 679 o Media type parameters to be used with the Session Description 680 Protocol (SDP) [RFC4566] 682 o A payload header extension mechanism and data structures for 683 enhanced support of temporal scalability based on that 684 extension mechanism. 686 2 Conventions 688 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL 689 NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and 690 "OPTIONAL" in this document are to be interpreted as described in 691 BCP 14, RFC 2119 [RFC2119]. 693 In this document, these key words will appear with that 694 interpretation only when in ALL CAPS. Lower case uses of these 695 words are not to be interpreted as carrying the RFC 2119 696 significance. 698 This specification uses the notion of setting and clearing a bit 699 when bit fields are handled. Setting a bit is the same as 700 assigning that bit the value of 1 (On). Clearing a bit is the 701 same as assigning that bit the value of 0 (Off). 703 3 Definitions and Abbreviations 705 3.1 Definitions 707 This document uses the terms and definitions of [HEVC]. Section 708 3.1.1 lists relevant definitions copied from [HEVC] for 709 convenience. Section 3.1.2 provides definitions specific to this 710 memo. 712 3.1.1 Definitions from the HEVC Specification 714 access unit: A set of NAL units that are associated with each 715 other according to a specified classification rule, are 716 consecutive in decoding order, and contain exactly one coded 717 picture. 719 BLA access unit: An access unit in which the coded picture is a 720 BLA picture. 722 BLA picture: An IRAP picture for which each VCL NAL unit has 723 nal_unit_type equal to BLA_W_LP, BLA_W_RADL, or BLA_N_LP. 725 coded video sequence (CVS): A sequence of access units that 726 consists, in decoding order, of an IRAP access unit with 727 NoRaslOutputFlag equal to 1, followed by zero or more access 728 units that are not IRAP access units with NoRaslOutputFlag equal 729 to 1, including all subsequent access units up to but not 730 including any subsequent access unit that is an IRAP access unit 731 with NoRaslOutputFlag equal to 1. 733 Informative note: An IRAP access unit may be an IDR access 734 unit, a BLA access unit, or a CRA access unit. The value of 735 NoRaslOutputFlag is equal to 1 for each IDR access unit, each 736 BLA access unit, and each CRA access unit that is the first 737 access unit in the bitstream in decoding order, is the first 738 access unit that follows an end of sequence NAL unit in 739 decoding order, or has HandleCraAsBlaFlag equal to 1. 741 CRA access unit: An access unit in which the coded picture is a 742 CRA picture. 744 CRA picture: A RAP picture for which each VCL NAL unit has 745 nal_unit_type equal to CRA_NUT. 747 IDR access unit: An access unit in which the coded picture is an 748 IDR picture. 750 IDR picture: A RAP picture for which each VCL NAL unit has 751 nal_unit_type equal to IDR_W_RADL or IDR_N_LP. 753 IRAP access unit: An access unit in which the coded picture is an 754 IRAP picture. 756 IRAP picture: A coded picture for which each VCL NAL unit has 757 nal_unit_type in the range of BLA_W_LP (16) to RSV_IRAP_VCL23 758 (23), inclusive. 760 layer: A set of VCL NAL units that all have a particular value of 761 nuh_layer_id and the associated non-VCL NAL units, or one of a 762 set of syntactical structures having a hierarchical relationship. 764 operation point: bitstream created from another bitstream by 765 operation of the sub-bitstream extraction process with the 766 another bitstream, a target highest TemporalId, and a target 767 layer identifier list as inputs. 769 random access: The act of starting the decoding process for a 770 bitstream at a point other than the beginning of the bitstream. 772 sub-layer: A temporal scalable layer of a temporal scalable 773 bitstream consisting of VCL NAL units with a particular value of 774 the TemporalId variable, and the associated non-VCL NAL units. 776 sub-layer representation: A subset of the bitstream consisting of 777 NAL units of a particular sub-layer and the lower sub-layers. 779 tile: A rectangular region of coding tree blocks within a 780 particular tile column and a particular tile row in a picture. 782 tile column: A rectangular region of coding tree blocks having a 783 height equal to the height of the picture and a width specified 784 by syntax elements in the picture parameter set. 786 tile row: A rectangular region of coding tree blocks having a 787 height specified by syntax elements in the picture parameter set 788 and a width equal to the width of the picture. 790 3.1.2 Definitions Specific to This Memo 792 dependee RTP stream: An RTP stream on which another RTP stream 793 depends. All RTP streams in an MRST or MRMT except for the 794 highest RTP stream are dependee RTP streams. 796 highest RTP stream: The RTP stream on which no other RTP stream 797 depends. The RTP stream in an SRST is the highest RTP stream. 799 media aware network element (MANE): A network element, such as a 800 middlebox, selective forwarding unit, or application layer 801 gateway that is capable of parsing certain aspects of the RTP 802 payload headers or the RTP payload and reacting to their 803 contents. 805 Informative note: The concept of a MANE goes beyond normal 806 routers or gateways in that a MANE has to be aware of the 807 signaling (e.g. to learn about the payload type mappings of 808 the media streams), and in that it has to be trusted when 809 working with SRTP. The advantage of using MANEs is that they 810 allow packets to be dropped according to the needs of the 811 media coding. For example, if a MANE has to drop packets due 812 to congestion on a certain link, it can identify and remove 813 those packets whose elimination produces the least adverse 814 effect on the user experience. After dropping packets, MANEs 815 must rewrite RTCP packets to match the changes to the RTP 816 stream as specified in Section 7 of [RFC3550]. 818 Media Transport: As used in the MRST, MRMT, and SRST definitions 819 below, Media Transport denotes the transport of packets over a 820 transport association identified by a 5-tuple (source address, 821 source port, destination address, destination port, transport 822 protocol). See also Section 2.1.13 of [I-D.ietf-avtext-rtp- 823 grouping-taxonomy]. 825 Informative note: The term "bitstream" in this document is 826 equivalent to the term "encoded stream" in [I-D.ietf-avtext- 827 rtp-grouping-taxonomy]. 829 Multiple RTP streams on a Single Transport (MRST): Multiple RTP 830 streams carrying a single HEVC bitstream on a Single Transport. 831 See also Section 3.5 of [I-D.ietf-avtext-rtp-grouping-taxonomy]. 833 Multiple RTP streams on Multiple Transports (MRMT): Multiple RTP 834 streams carrying a single HEVC bitstream on Multiple Transports. 835 See also Section 3.5 of [I-D.ietf-avtext-rtp-grouping-taxonomy]. 837 NAL unit decoding order: A NAL unit order that conforms to the 838 constraints on NAL unit order given in Section 7.4.2.4 in [HEVC]. 840 NAL unit output order: A NAL unit order in which NAL units of 841 different access units are in the output order of the decoded 842 pictures corresponding to the access units, as specified in 843 [HEVC], and in which NAL units within an access unit are in their 844 decoding order. 846 NAL-unit-like structure: A data structure that is similar to NAL 847 units in the sense that it also has a NAL unit header and a 848 payload, with a difference that the payload does not follow the 849 start code emulation prevention mechanism required for the NAL 850 unit syntax as specified in Section 7.3.1.1 of [HEVC]. Examples 851 NAL-unit-like structures defined in this memo are packet payloads 852 of AP, PACI, and FU packets. 854 NALU-time: The value that the RTP timestamp would have if the NAL 855 unit would be transported in its own RTP packet. 857 RTP stream: See [I-D.ietf-avtext-rtp-grouping-taxonomy]. Within 858 the scope of this memo, one RTP stream is utilized to transport 859 one or more temporal sub-layers. 861 Single RTP stream on a Single Transport (SRST): Single RTP 862 stream carrying a single HEVC bitstream on a Single (Media) 863 Transport. See also Section 3.5 of [I-D.ietf-avtext-rtp- 864 grouping-taxonomy]. 866 transmission order: The order of packets in ascending RTP 867 sequence number order (in modulo arithmetic). Within an 868 aggregation packet, the NAL unit transmission order is the same 869 as the order of appearance of NAL units in the packet. 871 3.2 Abbreviations 873 AP Aggregation Packet 875 BLA Broken Link Access 877 CRA Clean Random Access 879 CTB Coding Tree Block 881 CTU Coding Tree Unit 883 CVS Coded Video Sequence 885 DPH Decoded Picture Hash 887 FU Fragmentation Unit 889 HRD Hypothetical Reference Decoder 891 IDR Instantaneous Decoding Refresh 893 IRAP Intra Random Access Point 895 MANE Media Aware Network Element 896 MRMT Multiple RTP streams on Multiple Transports 898 MRST Multiple RTP streams on a Single Transport 900 MTU Maximum Transfer Unit 902 NAL Network Abstraction Layer 904 NALU Network Abstraction Layer Unit 906 PACI PAyload Content Information 908 PHES Payload Header Extension Structure 910 PPS Picture Parameter Set 912 RADL Random Access Decodable Leading (Picture) 914 RASL Random Access Skipped Leading (Picture) 916 RPS Reference Picture Set 918 SEI Supplemental Enhancement Information 920 SPS Sequence Parameter Set 922 SRST Single RTP stream on a Single Transport 924 STSA Step-wise Temporal Sub-layer Access 926 TSA Temporal Sub-layer Access 928 TSCI Temporal Scalability Control Information 930 VCL Video Coding Layer 932 VPS Video Parameter Set 934 4 RTP Payload Format 936 4.1 RTP Header Usage 938 The format of the RTP header is specified in [RFC3550] and 939 reprinted in Figure 2 for convenience. This payload format uses 940 the fields of the header in a manner consistent with that 941 specification. 943 The RTP payload (and the settings for some RTP header bits) for 944 aggregation packets and fragmentation units are specified in 945 Sections 4.4.2 and 4.4.3, respectively. 947 0 1 2 3 948 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 949 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 950 |V=2|P|X| CC |M| PT | sequence number | 951 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 952 | timestamp | 953 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 954 | synchronization source (SSRC) identifier | 955 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 956 | contributing source (CSRC) identifiers | 957 | .... | 958 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 960 Figure 2 RTP header according to [RFC3550] 962 The RTP header information to be set according to this RTP 963 payload format is set as follows: 965 Marker bit (M): 1 bit 967 Set for the last packet of the access unit, carried in the 968 current RTP stream. This is in line with the normal use of 969 the M bit in video formats to allow an efficient playout 970 buffer handling. When MRST or MRMT is in use, if an access 971 unit appears in multiple RTP streams, the marker bit is set on 972 each RTP stream's last packet of the access unit. 974 Informative note: The content of a NAL unit does not tell 975 whether or not the NAL unit is the last NAL unit, in 976 decoding order, of an access unit. An RTP sender 977 implementation may obtain these information from the video 978 encoder. If, however, the implementation cannot obtain 979 these information directly from the encoder, e.g. when the 980 bitstream was pre-encoded, and also there is no timestamp 981 allocated for each NAL unit, then the sender implementation 982 can inspect subsequent NAL units in decoding order to 983 determine whether or not the NAL unit is the last NAL unit 984 of an access unit as follows. A NAL unit is determined to 985 be the last NAL unit of an access unit if it is the last 986 NAL unit of the bitstream. A NAL unit naluX is also 987 determined to be the last NAL unit of an access unit if 988 both the following conditions are true: 1) the next VCL NAL 989 unit naluY in decoding order has the high-order bit of the 990 first byte after its NAL unit header equal to 1, and 2) all 991 NAL units between naluX and naluY, when present, have 992 nal_unit_type in the range of 32 to 35, inclusive, equal to 993 39, or in the ranges of 41 to 44, inclusive, or 48 to 55, 994 inclusive. 996 Payload type (PT): 7 bits 998 The assignment of an RTP payload type for this new packet 999 format is outside the scope of this document and will not be 1000 specified here. The assignment of a payload type has to be 1001 performed either through the profile used or in a dynamic way. 1003 Informative note: It is not required to use different 1004 payload type values for different RTP streams in MRST or 1005 MRMT. 1007 Sequence number (SN): 16 bits 1009 Set and used in accordance with RFC 3550 [RFC3550]. 1011 Timestamp: 32 bits 1013 The RTP timestamp is set to the sampling timestamp of the 1014 content. A 90 kHz clock rate MUST be used. 1016 If the NAL unit has no timing properties of its own (e.g. 1017 parameter set and SEI NAL units), the RTP timestamp MUST be 1018 set to the RTP timestamp of the coded picture of the access 1019 unit in which the NAL unit (according to Section 7.4.2.4.4 of 1020 [HEVC]) is included. 1022 Receivers MUST use the RTP timestamp for the display process, 1023 even when the bitstream contains picture timing SEI messages 1024 or decoding unit information SEI messages as specified in 1025 [HEVC]. However, this does not mean that picture timing SEI 1026 messages in the bitstream should be discarded, as picture 1027 timing SEI messages may contain frame-field information that 1028 is important in appropriately rendering interlaced video. 1030 Synchronization source (SSRC): 32-bits 1032 Used to identify the source of the RTP packets. When using 1033 SRST, by definition a single SSRC is used for all parts of a 1034 single bitstream. In MRST or MRMT, different SSRCs are used 1035 for each RTP stream containing a subset of the sub-layers of 1036 the single (temporally scalable) bitstream. A receiver is 1037 required to correctly associate the set of SSRCs that are 1038 included parts of the same bitstream. 1040 4.2 Payload Header Usage 1042 The first two bytes of the payload of an RTP packet are referred 1043 to as the payload header. The payload header consists of the 1044 same fields (F, Type, LayerId, and TID) as the NAL unit header as 1045 shown in Section 1.1.4, irrespective of the type of the payload 1046 structure. 1048 The TID value indicates (among other things) the relative 1049 importance of an RTP packet, for example because NAL units 1050 belonging to higher temporal sub-layers are not used for the 1051 decoding of lower temporal sub-layers. A lower value of TID 1052 indicates a higher importance. More important NAL units MAY be 1053 better protected against transmission losses than less important 1054 NAL units. 1056 4.3 Transmission Modes 1058 This memo enables transmission of an HEVC bitstream over 1060 . a single RTP stream on a single Media Transport (SRST), 1061 . multiple RTP streams over a single Media Transport (MRST), 1062 or 1063 . multiple RTP streams over multiple Media Transports (MRMT). 1065 Informative Note: While this specification enables the use of 1066 MRST within the H.265 RTP payload, the signaling of MRST within 1067 SDP Offer/Answer is not fully specified at the time of this 1068 writing. See [RFC5576] and [RFC5583] for what is supported 1069 today as well as [I-D.ietf-avtcore-rtp-multi-stream] and [I- 1070 D.ietf-mmusic-sdp-bundle-negotiation] for future directions. 1072 When in MRMT, the dependency of one RTP stream on another RTP 1073 stream is typically indicated as specified in [RFC5583]. 1074 [RFC5583] can also be utilized to specify dependencies within 1075 MRST, but only if the RTP streams utilize distinct payload types. 1076 When an RTP stream A depends on another RTP stream B, the RTP 1077 stream B is referred to as a dependee RTP stream of the RTP 1078 stream A. 1080 SRST or MRST SHOULD be used for point-to-point unicast scenarios, 1081 while MRMT SHOULD be used for point-to-multipoint multicast 1082 scenarios where different receivers require different operation 1083 points of the same HEVC bitstream, to improve bandwidth utilizing 1084 efficiency. 1086 Informative note: A multicast may degrade to a unicast after 1087 all but one receivers have left (this is a justification of 1088 the first "SHOULD" instead of "MUST"), and there might be 1089 scenarios where MRMT is desirable but not possible e.g. when 1090 IP multicast is not deployed in certain network (this is a 1091 justification of the second "SHOULD" instead of "MUST"). 1093 The transmission mode is indicated by the tx-mode media parameter 1094 (see Section 7.1). If tx-mode is equal to "SRST", SRST MUST be 1095 used. Otherwise, if tx-mode is equal to "MRST", MRST MUST be 1096 used. Otherwise (tx-mode is equal to "MRMT"), MRMT MUST be used. 1098 Informative note: When an RTP stream does not depend on other 1099 RTP streams, any of SRST, MRST and MRMT may be in use for the 1100 RTP stream. 1102 Receivers MUST support all of SRST, MRST, and MRMT. 1104 Informative note: The required support of MRMT by receivers 1105 does not imply that multicast must be supported by receivers. 1107 4.4 Payload Structures 1109 Four different types of RTP packet payload structures are 1110 specified. A receiver can identify the type of an RTP packet 1111 payload through the Type field in the payload header. 1113 The four different payload structures are as follows: 1115 o Single NAL unit packet: Contains a single NAL unit in the 1116 payload, and the NAL unit header of the NAL unit also serves 1117 as the payload header. This payload structure is specified in 1118 Section 4.4.1. 1120 o Aggregation packet (AP): Contains more than one NAL unit 1121 within one access unit. This payload structure is specified 1122 in Section 4.4.2. 1124 o Fragmentation unit (FU): Contains a subset of a single NAL 1125 unit. This payload structure is specified in Section 4.4.3. 1127 o PACI carrying RTP packet: Contains a payload header (that 1128 differs from other payload headers for efficiency), a Payload 1129 Header Extension Structure (PHES), and a PACI payload. This 1130 payload structure is specified in Section 4.4.4. 1132 4.4.1 Single NAL Unit Packets 1134 A single NAL unit packet contains exactly one NAL unit, and 1135 consists of a payload header (denoted as PayloadHdr), a 1136 conditional 16-bit DONL field (in network byte order), and the 1137 NAL unit payload data (the NAL unit excluding its NAL unit 1138 header) of the contained NAL unit, as shown in Figure 3. 1140 0 1 2 3 1141 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1142 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1143 | PayloadHdr | DONL (conditional) | 1144 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1145 | | 1146 | NAL unit payload data | 1147 | | 1148 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1149 | :...OPTIONAL RTP padding | 1150 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1152 Figure 3 The structure a single NAL unit packet 1154 The payload header SHOULD be an exact copy of the NAL unit header 1155 of the contained NAL unit. However, the Type (i.e. 1156 nal_unit_type) field MAY be changed, e.g. when it is desirable to 1157 handle a CRA picture to be a BLA picture [JCTVC-J0107]. 1159 The DONL field, when present, specifies the value of the 16 least 1160 significant bits of the decoding order number of the contained 1161 NAL unit. If sprop-max-don-diff is greater than 0 for any of the 1162 RTP streams, the DONL field MUST be present, and the variable DON 1163 for the contained NAL unit is derived as equal to the value of 1164 the DONL field. Otherwise (sprop-max-don-diff is equal to 0 for 1165 all the RTP streams), the DONL field MUST NOT be present. 1167 4.4.2 Aggregation Packets (APs) 1169 Aggregation packets (APs) are introduced to enable the reduction 1170 of packetization overhead for small NAL units, such as most of 1171 the non-VCL NAL units, which are often only a few octets in size. 1173 An AP aggregates NAL units within one access unit. Each NAL unit 1174 to be carried in an AP is encapsulated in an aggregation unit. 1175 NAL units aggregated in one AP are in NAL unit decoding order. 1177 An AP consists of a payload header (denoted as PayloadHdr) 1178 followed by two or more aggregation units, as shown in Figure 4. 1180 0 1 2 3 1181 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1182 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1183 | PayloadHdr (Type=48) | | 1184 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 1185 | | 1186 | two or more aggregation units | 1187 | | 1188 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1189 | :...OPTIONAL RTP padding | 1190 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1192 Figure 4 The structure of an aggregation packet 1194 The fields in the payload header are set as follows. The F bit 1195 MUST be equal to 0 if the F bit of each aggregated NAL unit is 1196 equal to zero; otherwise, it MUST be equal to 1. The Type field 1197 MUST be equal to 48. The value of LayerId MUST be equal to the 1198 lowest value of LayerId of all the aggregated NAL units. The 1199 value of TID MUST be the lowest value of TID of all the 1200 aggregated NAL units. 1202 Informative Note: All VCL NAL units in an AP have the same TID 1203 value since they belong to the same access unit. However, an 1204 AP may contain non-VCL NAL units for which the TID value in 1205 the NAL unit header may be different than the TID value of the 1206 VCL NAL units in the same AP. 1208 An AP MUST carry at least two aggregation units and can carry as 1209 many aggregation units as necessary; however, the total amount of 1210 data in an AP obviously MUST fit into an IP packet, and the size 1211 SHOULD be chosen so that the resulting IP packet is smaller than 1212 the MTU size so to avoid IP layer fragmentation. An AP MUST NOT 1213 contain Fragmentation Units (FUs) specified in Section 4.4.3. 1214 APs MUST NOT be nested; i.e. an AP must not contain another AP. 1216 The first aggregation unit in an AP consists of a conditional 16- 1217 bit DONL field (in network byte order) followed by a 16-bit 1218 unsigned size information (in network byte order) that indicates 1219 the size of the NAL unit in bytes (excluding these two octets, 1220 but including the NAL unit header), followed by the NAL unit 1221 itself, including its NAL unit header, as shown in Figure 5. 1223 0 1 2 3 1224 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1225 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1226 : DONL (conditional) | NALU size | 1227 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1228 | NALU size | | 1229 +-+-+-+-+-+-+-+-+ NAL unit | 1230 | | 1231 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1232 | : 1233 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1235 Figure 5 The structure of the first aggregation unit in an AP 1237 The DONL field, when present, specifies the value of the 16 least 1238 significant bits of the decoding order number of the aggregated 1239 NAL unit. 1241 If sprop-max-don-diff is greater than 0 for any of the RTP 1242 streams, the DONL field MUST be present in an aggregation unit 1243 that is the first aggregation unit in an AP, and the variable DON 1244 for the aggregated NAL unit is derived as equal to the value of 1245 the DONL field. Otherwise (sprop-max-don-diff is equal to 0 for 1246 all the RTP streams), the DONL field MUST NOT be present in an 1247 aggregation unit that is the first aggregation unit in an AP. 1249 An aggregation unit that is not the first aggregation unit in an 1250 AP consists of a conditional 8-bit DOND field followed by a 16- 1251 bit unsigned size information (in network byte order) that 1252 indicates the size of the NAL unit in bytes (excluding these two 1253 octets, but including the NAL unit header), followed by the NAL 1254 unit itself, including its NAL unit header, as shown in Figure 6. 1256 0 1 2 3 1257 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1258 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1259 : DOND (cond) | NALU size | 1260 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1261 | | 1262 | NAL unit | 1263 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1264 | : 1265 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1267 Figure 6 The structure of an aggregation unit that is not the 1268 first aggregation unit in an AP 1270 When present, the DOND field plus 1 specifies the difference 1271 between the decoding order number values of the current 1272 aggregated NAL unit and the preceding aggregated NAL unit in the 1273 same AP. 1275 If sprop-max-don-diff is greater than 0 for any of the RTP 1276 streams, the DOND field MUST be present in an aggregation unit 1277 that is not the first aggregation unit in an AP, and the variable 1278 DON for the aggregated NAL unit is derived as equal to the DON of 1279 the preceding aggregated NAL unit in the same AP plus the value 1280 of the DOND field plus 1 modulo 65536. Otherwise (sprop-max-don- 1281 diff is equal to 0 for all the RTP streams), the DOND field MUST 1282 NOT be present in an aggregation unit that is not the first 1283 aggregation unit in an AP, and in this case the transmission 1284 order and decoding order of NAL units carried in the AP are the 1285 same as the order the NAL units appear in the AP. 1287 Figure 7 presents an example of an AP that contains two 1288 aggregation units, labeled as 1 and 2 in the figure, without the 1289 DONL and DOND fields being present. 1291 0 1 2 3 1292 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1293 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1294 | RTP Header | 1295 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1296 | PayloadHdr (Type=48) | NALU 1 Size | 1297 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1298 | NALU 1 HDR | | 1299 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ NALU 1 Data | 1300 | . . . | 1301 | | 1302 + +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1303 | . . . | NALU 2 Size | NALU 2 HDR | 1304 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1305 | NALU 2 HDR | | 1306 +-+-+-+-+-+-+-+-+ NALU 2 Data | 1307 | . . . | 1308 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1309 | :...OPTIONAL RTP padding | 1310 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1312 Figure 7 An example of an AP packet containing two aggregation 1313 units without the DONL and DOND fields 1315 Figure 8 presents an example of an AP that contains two 1316 aggregation units, labeled as 1 and 2 in the figure, with the 1317 DONL and DOND fields being present. 1319 0 1 2 3 1320 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1321 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1322 | RTP Header | 1323 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1324 | PayloadHdr (Type=48) | NALU 1 DONL | 1325 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1326 | NALU 1 Size | NALU 1 HDR | 1327 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1328 | | 1329 | NALU 1 Data . . . | 1330 | | 1331 + . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1332 | | NALU 2 DOND | NALU 2 Size | 1333 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1334 | NALU 2 HDR | | 1335 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ NALU 2 Data | 1336 | | 1337 | . . . +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1338 | :...OPTIONAL RTP padding | 1339 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1341 Figure 8 An example of an AP containing two aggregation units 1342 with the DONL and DOND fields 1344 4.4.3 Fragmentation Units (FUs) 1346 Fragmentation units (FUs) are introduced to enable fragmenting a 1347 single NAL unit into multiple RTP packets, possibly without 1348 cooperation or knowledge of the HEVC encoder. A fragment of a NAL 1349 unit consists of an integer number of consecutive octets of that 1350 NAL unit. Fragments of the same NAL unit MUST be sent in consecutive 1351 order with ascending RTP sequence numbers (with no other RTP packets 1352 within the same RTP stream being sent between the first and last 1353 fragment). 1355 When a NAL unit is fragmented and conveyed within FUs, it is 1356 referred to as a fragmented NAL unit. APs MUST NOT be 1357 fragmented. FUs MUST NOT be nested; i.e. an FU must not contain 1358 a subset of another FU. 1360 The RTP timestamp of an RTP packet carrying an FU is set to the 1361 NALU-time of the fragmented NAL unit. 1363 An FU consists of a payload header (denoted as PayloadHdr), an FU 1364 header of one octet, a conditional 16-bit DONL field (in network 1365 byte order), and an FU payload, as shown in Figure 9. 1367 0 1 2 3 1368 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1369 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1370 | PayloadHdr (Type=49) | FU header | DONL (cond) | 1371 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-| 1372 | DONL (cond) | | 1373 |-+-+-+-+-+-+-+-+ | 1374 | FU payload | 1375 | | 1376 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1377 | :...OPTIONAL RTP padding | 1378 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1380 Figure 9 The structure of an FU 1382 The fields in the payload header are set as follows. The Type 1383 field MUST be equal to 49. The fields F, LayerId, and TID MUST 1384 be equal to the fields F, LayerId, and TID, respectively, of the 1385 fragmented NAL unit. 1387 The FU header consists of an S bit, an E bit, and a 6-bit FuType 1388 field, as shown in Figure 10. 1390 +---------------+ 1391 |0|1|2|3|4|5|6|7| 1392 +-+-+-+-+-+-+-+-+ 1393 |S|E| FuType | 1394 +---------------+ 1396 Figure 10 The structure of FU header 1398 The semantics of the FU header fields are as follows: 1399 S: 1 bit 1400 When set to one, the S bit indicates the start of a fragmented 1401 NAL unit i.e. the first byte of the FU payload is also the 1402 first byte of the payload of the fragmented NAL unit. When 1403 the FU payload is not the start of the fragmented NAL unit 1404 payload, the S bit MUST be set to zero. 1406 E: 1 bit 1407 When set to one, the E bit indicates the end of a fragmented 1408 NAL unit, i.e. the last byte of the payload is also the last 1409 byte of the fragmented NAL unit. When the FU payload is not 1410 the last fragment of a fragmented NAL unit, the E bit MUST be 1411 set to zero. 1413 FuType: 6 bits 1414 The field FuType MUST be equal to the field Type of the 1415 fragmented NAL unit. 1417 The DONL field, when present, specifies the value of the 16 least 1418 significant bits of the decoding order number of the fragmented 1419 NAL unit. 1421 If sprop-max-don-diff is greater than 0 for any of the RTP 1422 streams, and the S bit is equal to 1, the DONL field MUST be 1423 present in the FU, and the variable DON for the fragmented NAL 1424 unit is derived as equal to the value of the DONL field. 1425 Otherwise (sprop-max-don-diff is equal to 0 for all the RTP 1426 streams, or the S bit is equal to 0), the DONL field MUST NOT be 1427 present in the FU. 1429 A non-fragmented NAL unit MUST NOT be transmitted in one FU; i.e. 1430 the Start bit and End bit must not both be set to one in the same 1431 FU header. 1433 The FU payload consists of fragments of the payload of the 1434 fragmented NAL unit so that if the FU payloads of consecutive 1435 FUs, starting with an FU with the S bit equal to 1 and ending 1436 with an FU with the E bit equal to 1, are sequentially 1437 concatenated, the payload of the fragmented NAL unit can be 1438 reconstructed. The NAL unit header of the fragmented NAL unit is 1439 not included as such in the FU payload, but rather the 1440 information of the NAL unit header of the fragmented NAL unit is 1441 conveyed in F, LayerId, and TID fields of the FU payload headers 1442 of the FUs and the FuType field of the FU header of the FUs. An 1443 FU payload MUST NOT be empty. 1445 If an FU is lost, the receiver SHOULD discard all following 1446 fragmentation units in transmission order corresponding to the 1447 same fragmented NAL unit, unless the decoder in the receiver is 1448 known to be prepared to gracefully handle incomplete NAL units. 1450 A receiver in an endpoint or in a MANE MAY aggregate the first n- 1451 1 fragments of a NAL unit to an (incomplete) NAL unit, even if 1452 fragment n of that NAL unit is not received. In this case, the 1453 forbidden_zero_bit of the NAL unit MUST be set to one to indicate 1454 a syntax violation. 1456 4.4.4 PACI packets 1458 This section specifies the PACI packet structure. The basic 1459 payload header specified in this memo is intentionally limited to 1460 the 16 bits of the NAL unit header so to keep the packetization 1461 overhead to a minimum. However, cases have been identified where 1462 it is advisable to include control information in an easily 1463 accessible position in the packet header, despite the additional 1464 overhead. One such control information is the Temporal 1465 Scalability Control Information as specified in Section 4.5 1466 below. PACI packets carry this and future, similar structures. 1468 The PACI packet structure is based on a payload header extension 1469 mechanism that is generic and extensible to carry payload header 1470 extensions. In this section, the focus lies on the use within 1471 this specification. Section 4.4.4.2 below provides guidance for 1472 the specification designers in how to employ the extension 1473 mechanism in future specifications. 1475 A PACI packet consists of a payload header (denoted as 1476 PayloadHdr), for which the structure follows what is described in 1477 Section 4.2 above. The payload header is followed by the fields 1478 A, cType, PHSsize, F[0..2] and Y. 1480 Figure 11 shows a PACI packet in compliance with this memo; that 1481 is, without any extensions. 1483 0 1 2 3 1484 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1485 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1486 | PayloadHdr (Type=50) |A| cType | PHSsize |F0..2|Y| 1487 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1488 | Payload Header Extension Structure (PHES) | 1489 |=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=| 1490 | | 1491 | PACI payload: NAL unit | 1492 | . . . | 1493 | | 1494 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1495 | :...OPTIONAL RTP padding | 1496 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1498 Figure 11 The structure of a PACI 1500 The fields in the payload header are set as follows. The F bit 1501 MUST be equal to 0. The Type field MUST be equal to 50. The 1502 value of LayerId MUST be a copy of the LayerId field of the PACI 1503 payload NAL unit or NAL-unit-like structure. The value of TID 1504 MUST be a copy of the TID field of the PACI payload NAL unit or 1505 NAL-unit-like structure. 1507 The semantics of other fields are as follows: 1509 A: 1 bit 1510 Copy of the F bit of the PACI payload NAL unit or NAL-unit- 1511 like structure. 1513 cType: 6 bits 1514 Copy of the Type field of the PACI payload NAL unit or NAL- 1515 unit-like structure. 1517 PHSsize: 5 bits 1518 Indicates the length of the PHES field. The value is limited 1519 to be less than or equal to 32 octets, to simplify encoder 1520 design for MTU size matching. 1522 F0 1523 This field equal to 1 specifies the presence of a temporal 1524 scalability support extension in the PHES. 1526 F1, F2 1527 MUST be 0, available for future extensions, see Section 1528 4.4.4.2. Receivers compliant with this version of the HEVC 1529 payload format MUST ignore F1=1 and/or F2=1, and also ignore 1530 any information in the PHES indicated as present by F1=1 1531 and/or F2=1. 1533 Informative note: The receiver can do that by first 1534 decoding information associated with F0=1, and then 1535 skipping over any remaining bytes of the PHES based on the 1536 value of PHSsize. 1538 Y: 1 bit 1539 MUST be 0, available for future extensions, see Section 1540 4.4.4.2. Receivers compliant with this version of the HEVC 1541 payload format MUST ignore Y=1, and also ignore any 1542 information in the PHES indicated as present by Y. 1544 PHES: variable number of octets 1545 A variable number of octets as indicated by the value of 1546 PHSsize. 1548 PACI Payload 1549 The single NAL unit packet or NAL-unit-like structure (such 1550 as: FU or AP) to be carried, not including the first two 1551 octets. 1553 Informative note: The first two octets of the NAL unit or 1554 NAL-unit-like structure carried in the PACI payload are not 1555 included in the PACI payload. Rather, the respective values 1556 are copied in locations of the PayloadHdr of the RTP 1557 packet. This design offers two advantages: first, the 1558 overall structure of the payload header is preserved, i.e. 1559 there is no special case of payload header structure that 1560 needs to be implemented for PACI. Second, no additional 1561 overhead is introduced. 1563 A PACI payload MAY be a single NAL unit, an FU, or an AP. 1564 PACIs MUST NOT be fragmented or aggregated. The following 1565 subsection documents the reasons for these design choices. 1567 4.4.4.1 Reasons for the PACI rules (informative) 1569 A PACI cannot be fragmented. If a PACI could be fragmented, and 1570 a fragment other than the first fragment would get lost, access 1571 to the information in the PACI would not be possible. Therefore, 1572 a PACI must not be fragmented. In other words, an FU must not 1573 carry (fragments of) a PACI. 1575 A PACI cannot be aggregated. Aggregation of PACIs is inadvisable 1576 from a compression viewpoint, as, in many cases, several to be 1577 aggregated NAL units would share identical PACI fields and values 1578 which would be carried redundantly for no reason. Most, if not 1579 all the practical effects of PACI aggregation can be achieved by 1580 aggregating NAL units and bundling them with a PACI (see below). 1581 Therefore, a PACI must not be aggregated. In other words, an AP 1582 must not contain a PACI. 1584 The payload of a PACI can be a fragment. Both middleboxes and 1585 sending systems with inflexible (often hardware-based) encoders 1586 occasionally find themselves in situations where a PACI and its 1587 headers, combined, are larger than the MTU size. In such a 1588 scenario, the middlebox or sender can fragment the NAL unit and 1589 encapsulate the fragment in a PACI. Doing so preserves the 1590 payload header extension information for all fragments, allowing 1591 downstream middleboxes and the receiver to take advantage of that 1592 information. Therefore, a sender may place a fragment into a 1593 PACI, and a receiver must be able to handle such a PACI. 1595 The payload of a PACI can be an aggregation NAL unit. HEVC 1596 bitstreams can contain unevenly sized and/or small (when compared 1597 to the MTU size) NAL units. In order to efficiently packetize 1598 such small NAL units, AP were introduced. The benefits of APs 1599 are independent from the need for a payload header extension. 1600 Therefore, a sender may place an AP into a PACI, and a receiver 1601 must be able to handle such a PACI. 1603 4.4.4.2 PACI extensions (Informative) 1605 This section includes recommendations for future specification 1606 designers on how to extent the PACI syntax to accommodate future 1607 extensions. Obviously, designers are free to specify whatever 1608 appears to be appropriate to them at the time of their design. 1609 However, a lot of thought has been invested into the extension 1610 mechanism described below, and we suggest that deviations from it 1611 warrant a good explanation. 1613 This memo defines only a single payload header extension (Temporal 1614 Scalability Control Information, described below in Section 4.5), 1615 and, therefore, only the F0 bit carries semantics. F1 and F2 are 1616 already named (and not just marked as reserved, as a typical video 1617 spec designer would do). They are intended to signal two additional 1618 extensions. The Y bit allows to, recursively, add further F and Y 1619 bits to extend the mechanism beyond 3 possible payload header 1620 extensions. It is suggested to define a new packet type (using a 1621 different value for Type) when assigning the F1, F2, or Y bits 1622 different semantics than what is suggested below. 1624 When a Y bit is set, an 8 bit flag-extension is inserted after 1625 the Y bit. A flag-extension consists of 7 flags F[n..n+6], and 1626 another Y bit. 1628 The basic PACI header already includes F0, F1, and F2. 1629 Therefore, the Fx bits in the first flag-extensions are numbered 1630 F3, F4, ..., F9, the F bits in the second flag-extension are 1631 numbered F10, F11, ..., F16, and so forth. As a result, at least 1632 3 Fx bits are always in the PACI, but the number of Fx bits (and 1633 associated types of extensions), can be increased by setting the 1634 next Y bit and adding an octet of flag-extensions, carrying 7 1635 flags and another Y bit. The size of this list of flags is 1636 subject to the limits specified in Section 4.4.4 (32 octets for 1637 all flag-extensions and the PHES information combined). 1639 Each of the F bits can indicate either the presence of 1640 information in the Payload Header Extension Structure (PHES), 1641 described below, or a given F bit can indicate a certain 1642 condition, without including additional information in the PHES. 1644 When a spec developer devises a new syntax that takes advantage 1645 of the PACI extension mechanism, he/she must follow the 1646 constraints listed below; otherwise the extension mechanism may 1647 break. 1649 1) The fields added for a particular Fx bit MUST be fixed in 1650 length and not depend on what other Fx bits are set (no 1651 parsing dependency). 1652 2) The Fx bits must be assigned in order. 1653 3) An implementation that supports the n-th Fn bit for any 1654 value of n must understand the syntax (though not 1655 necessarily the semantics) of the fields Fk (with k < n), so 1656 to be able to either use those bits when present, or at 1657 least be able to skip over them. 1659 4.5 Temporal Scalability Control Information 1661 This section describes the single payload header extension 1662 defined in this specification, known as Temporal Scalability 1663 Control Information (TSCI). If, in the future, additional 1664 payload header extensions become necessary, they could be 1665 specified in this section of an updated version of this document, 1666 or in their own documents. 1668 When F0 is set to 1 in a PACI, this specifies that the PHES field 1669 includes the TSCI fields TL0PICIDX, IrapPicID, S, and E as 1670 follows: 1672 0 1 2 3 1673 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 1674 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1675 | PayloadHdr (Type=50) |A| cType | PHSsize |F0..2|Y| 1676 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1677 | TL0PICIDX | IrapPicID |S|E| RES | | 1678 |-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 1679 | .... | 1680 | PACI payload: NAL unit | 1681 | | 1682 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1683 | :...OPTIONAL RTP padding | 1684 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 1686 Figure 12 The structure of a PACI with a PHES containing a TSCI 1688 TL0PICIDX (8 bits) 1689 When present, the TL0PICIDX field MUST be set to equal to 1690 temporal_sub_layer_zero_idx as specified in Section D.3.22 of 1691 [H.265] for the access unit containing the NAL unit in the 1692 PACI. 1694 IrapPicID (8 bits) 1695 When present, the IrapPicID field MUST be set to equal to 1696 irap_pic_id as specified in Section D.3.22 of [H.265] for the 1697 access unit containing the NAL unit in the PACI. 1699 S (1 bit) 1700 The S bit MUST be set to 1 if any of the following conditions 1701 is true and MUST be set to 0 otherwise: 1702 o The NAL unit in the payload of the PACI is the first VCL NAL 1703 unit, in decoding order, of a picture. 1704 o The NAL unit in the payload of the PACI is an AP and the NAL 1705 unit in the first contained aggregation unit is the first 1706 VCL NAL unit, in decoding order, of a picture. 1707 o The NAL unit in the payload of the PACI is an FU with its S 1708 bit equal to 1 and the FU payload containing a fragment of 1709 the first VCL NAL unit, in decoding order of a picture. 1711 E (1 bit) 1712 The E bit MUST be set to 1 if any of the following conditions 1713 is true and MUST be set to 0 otherwise: 1714 o The NAL unit in the payload of the PACI is the last VCL NAL 1715 unit, in decoding order, of a picture. 1716 o The NAL unit in the payload of the PACI is an AP and the NAL 1717 unit in the last contained aggregation unit is the last VCL 1718 NAL unit, in decoding order, of a picture. 1719 o The NAL unit in the payload of the PACI is an FU with its E 1720 bit equal to 1 and the FU payload containing a fragment of 1721 the last VCL NAL unit, in decoding order of a picture. 1723 RES (6 bits) 1724 MUST be equal to 0. Reserved for future extensions. 1726 The value of PHSsize MUST be set to 3. Receivers MUST allow 1727 other values of the fields F0, F1, F2, Y, and PHSsize, and MUST 1728 ignore any additional fields, when present, than specified above 1729 in the PHES. 1731 4.6 Decoding Order Number 1733 For each NAL unit, the variable AbsDon is derived, representing 1734 the decoding order number that is indicative of the NAL unit 1735 decoding order. 1737 Let NAL unit n be the n-th NAL unit in transmission order within 1738 an RTP stream. 1740 If sprop-max-don-diff is equal to 0 for all the RTP streams 1741 carrying the HEVC bitstream, AbsDon[n], the value of AbsDon for 1742 NAL unit n, is derived as equal to n. 1744 Otherwise (sprop-max-don-diff is greater than 0 for any of the 1745 RTP streams), AbsDon[n] is derived as follows, where DON[n] is 1746 the value of the variable DON for NAL unit n: 1748 o If n is equal to 0 (i.e. NAL unit n is the very first NAL unit 1749 in transmission order), AbsDon[0] is set equal to DON[0]. 1751 o Otherwise (n is greater than 0), the following applies for 1752 derivation of AbsDon[n]: 1754 If DON[n] == DON[n-1], 1755 AbsDon[n] = AbsDon[n-1] 1757 If (DON[n] > DON[n-1] and DON[n] - DON[n-1] < 32768), 1758 AbsDon[n] = AbsDon[n-1] + DON[n] - DON[n-1] 1760 If (DON[n] < DON[n-1] and DON[n-1] - DON[n] >= 32768), 1761 AbsDon[n] = AbsDon[n-1] + 65536 - DON[n-1] + DON[n] 1763 If (DON[n] > DON[n-1] and DON[n] - DON[n-1] >= 32768), 1764 AbsDon[n] = AbsDon[n-1] - (DON[n-1] + 65536 - 1765 DON[n]) 1767 If (DON[n] < DON[n-1] and DON[n-1] - DON[n] < 32768), 1768 AbsDon[n] = AbsDon[n-1] - (DON[n-1] - DON[n]) 1770 For any two NAL units m and n, the following applies: 1772 o AbsDon[n] greater than AbsDon[m] indicates that NAL unit n 1773 follows NAL unit m in NAL unit decoding order. 1775 o When AbsDon[n] is equal to AbsDon[m], the NAL unit decoding 1776 order of the two NAL units can be in either order. 1778 o AbsDon[n] less than AbsDon[m] indicates that NAL unit n 1779 precedes NAL unit m in decoding order. 1781 Informative note: When two consecutive NAL units in the NAL 1782 unit decoding order have different values of AbsDon, the 1783 absolute difference between the two AbsDon values may be 1784 greater than or equal to 1. 1786 Informative note: There are multiple reasons to allow for the 1787 absolute difference of the values of AbsDon for two 1788 consecutive NAL units in the NAL unit decoding order to be 1789 greater than one. An increment by one is not required, as at 1790 the time of associating values of AbsDon to NAL units, it may 1791 not be known whether all NAL units are to be delivered to the 1792 receiver. For example, a gateway may not forward VCL NAL 1793 units of higher sub-layers or some SEI NAL units when there is 1794 congestion in the network. In another example, the first 1795 intra-coded picture of a pre-encoded clip is transmitted in 1796 advance to ensure that it is readily available in the 1797 receiver, and when transmitting the first intra-coded picture, 1798 the originator does not exactly know how many NAL units will 1799 be encoded before the first intra-coded picture of the pre- 1800 encoded clip follows in decoding order. Thus, the values of 1801 AbsDon for the NAL units of the first intra-coded picture of 1802 the pre-encoded clip have to be estimated when they are 1803 transmitted, and gaps in values of AbsDon may occur. Another 1804 example is MRST or MRMT with sprop-max-don-diff greater than 1805 0, where the AbsDon values must indicate cross-layer decoding 1806 order for NAL units conveyed in all the RTP streams. 1808 5 Packetization Rules 1810 The following packetization rules apply: 1812 o If sprop-max-don-diff is greater than 0 for any of the RTP 1813 streams, the transmission order of NAL units carried in the RTP 1814 stream MAY be different than the NAL unit decoding order and the 1815 NAL unit output order. Otherwise (sprop-max-don-diff is equal 1816 to 0 for all the RTP streams), the transmission order of NAL 1817 units carried in the RTP stream MUST be the same as the NAL unit 1818 decoding order, and, when tx-mode is equal to "MRST" or "MRMT", 1819 MUST also be the same as the NAL unit output order. 1821 o A NAL unit of a small size SHOULD be encapsulated in an 1822 aggregation packet together with one or more other NAL units 1823 in order to avoid the unnecessary packetization overhead for 1824 small NAL units. For example, non-VCL NAL units such as 1825 access unit delimiters, parameter sets, or SEI NAL units are 1826 typically small and can often be aggregated with VCL NAL units 1827 without violating MTU size constraints. 1829 o Each non-VCL NAL unit SHOULD, when possible from an MTU size 1830 match viewpoint, be encapsulated in an aggregation packet 1831 together with its associated VCL NAL unit, as typically a non- 1832 VCL NAL unit would be meaningless without the associated VCL 1833 NAL unit being available. 1835 o For carrying exactly one NAL unit in an RTP packet, a single 1836 NAL unit packet MUST be used. 1838 6 De-packetization Process 1840 The general concept behind de-packetization is to get the NAL 1841 units out of the RTP packets in an RTP stream and all RTP streams 1842 the RTP stream depends on, if any, and pass them to the decoder 1843 in the NAL unit decoding order. 1845 The de-packetization process is implementation dependent. 1846 Therefore, the following description should be seen as an example 1847 of a suitable implementation. Other schemes may be used as well 1848 as long as the output for the same input is the same as the 1849 process described below. The output is the same when the set of 1850 output NAL units and their order are both identical. 1851 Optimizations relative to the described algorithms are possible. 1853 All normal RTP mechanisms related to buffer management apply. In 1854 particular, duplicated or outdated RTP packets (as indicated by 1855 the RTP sequences number and the RTP timestamp) are removed. To 1856 determine the exact time for decoding, factors such as a possible 1857 intentional delay to allow for proper inter-stream 1858 synchronization must be factored in. 1860 NAL units with NAL unit type values in the range of 0 to 47, 1861 inclusive may be passed to the decoder. NAL-unit-like structures 1862 with NAL unit type values in the range of 48 to 63, inclusive, 1863 MUST NOT be passed to the decoder. 1865 The receiver includes a receiver buffer, which is used to 1866 compensate for transmission delay jitter within individual RTP 1867 streams and across RTP streams, to reorder NAL units from 1868 transmission order to the NAL unit decoding order, and to recover 1869 the NAL unit decoding order in MRST or MRMT, when applicable. In 1870 this section, the receiver operation is described under the 1871 assumption that there is no transmission delay jitter within an 1872 RTP stream and across RTP streams. To make a difference from a 1873 practical receiver buffer that is also used for compensation of 1874 transmission delay jitter, the receiver buffer is here after 1875 called the de-packetization buffer in this section. Receivers 1876 should also prepare for transmission delay jitter; i.e. either 1877 reserve separate buffers for transmission delay jitter buffering 1878 and de-packetization buffering or use a receiver buffer for both 1879 transmission delay jitter and de-packetization. Moreover, 1880 receivers should take transmission delay jitter into account in 1881 the buffering operation; e.g. by additional initial buffering 1882 before starting of decoding and playback. 1884 When sprop-max-don-diff is equal to 0 for all the received RTP 1885 streams, the de-packetization buffer size is zero bytes and the 1886 process described in the remainder of this paragraph applies. 1887 When there is only one RTP stream received, the NAL units carried 1888 in the single RTP stream are directly passed to the decoder in 1889 their transmission order, which is identical to their decoding 1890 order. When there is more than one RTP stream received, the NAL 1891 units carried in the multiple RTP streams are passed to the 1892 decoder in their NTP timestamp order. When there are several NAL 1893 units of different RTP streams with the same NTP timestamp, the 1894 order to pass them to the decoder is their dependency order, 1895 where NAL units of a dependee RTP stream are passed to the 1896 decoder prior to the NAL units of the dependent RTP stream. When 1897 there are several NAL units of the same RTP stream with the same 1898 NTP timestamp, the order to pass them to the decoder is their 1899 transmission order. 1901 Informative note: The mapping between RTP and NTP 1902 timestamps is conveyed in RTCP SR packets. In addition, 1903 the mechanisms for faster media timestamp synchronization 1904 discussed in [RFC6051] may be used to speed up the 1905 acquisition of the RTP-to-wall-clock mapping. 1907 When sprop-max-don-diff is greater than 0 for any the received 1908 RTP streams, the process described in the remainder of this 1909 section applies. 1911 There are two buffering states in the receiver: initial buffering 1912 and buffering while playing. Initial buffering starts when the 1913 reception is initialized. After initial buffering, decoding and 1914 playback are started, and the buffering-while-playing mode is 1915 used. 1917 Regardless of the buffering state, the receiver stores incoming 1918 NAL units, in reception order, into the de-packetization buffer. 1919 NAL units carried in RTP packets are stored in the de- 1920 packetization buffer individually, and the value of AbsDon is 1921 calculated and stored for each NAL unit. When MRST or MRMT is in 1922 use, NAL units of all RTP streams of a bitstream are stored in 1923 the same de-packetization buffer. When NAL units carried in any 1924 two RTP streams are available to be placed into the de- 1925 packetization buffer, those NAL units carried in the RTP stream 1926 that is lower in the dependency tree are placed into the buffer 1927 first. For example, if RTP stream A depends on RTP stream B, 1928 then NAL units carried in RTP stream B are placed into the buffer 1929 first. 1931 Initial buffering lasts until condition A (the difference between 1932 the greatest and smallest AbsDon values of the NAL units in the 1933 de-packetization buffer is greater than or equal to the value of 1934 sprop-max-don-diff of the highest RTP stream) or condition B (the 1935 number of NAL units in the de-packetization buffer is greater 1936 than the value of sprop-depack-buf-nalus) is true. 1938 After initial buffering, whenever condition A or condition B is 1939 true, the following operation is repeatedly applied until both 1940 condition A and condition B become false: 1942 o The NAL unit in the de-packetization buffer with the smallest 1943 value of AbsDon is removed from the de-packetization buffer 1944 and passed to the decoder. 1946 When no more NAL units are flowing into the de-packetization 1947 buffer, all NAL units remaining in the de-packetization buffer 1948 are removed from the buffer and passed to the decoder in the 1949 order of increasing AbsDon values. 1951 7 Payload Format Parameters 1953 This section specifies the parameters that MAY be used to select 1954 optional features of the payload format and certain features or 1955 properties of the bitstream or the RTP stream. The parameters 1956 are specified here as part of the media type registration for the 1957 HEVC codec. A mapping of the parameters into the Session 1958 Description Protocol (SDP) [RFC4566] is also provided for 1959 applications that use SDP. Equivalent parameters could be 1960 defined elsewhere for use with control protocols that do not use 1961 SDP. 1963 7.1 Media Type Registration 1965 The media subtype for the HEVC codec is allocated from the IETF 1966 tree. 1968 The receiver MUST ignore any unrecognized parameter. 1970 Media Type name: video 1972 Media subtype name: H265 1974 Required parameters: none 1976 OPTIONAL parameters: 1978 profile-space, tier-flag, profile-id, profile-compatibility- 1979 indicator, interop-constraints, and level-id: 1981 These parameters indicate the profile, tier, default level, 1982 and some constraints of the bitstream carried by the RTP 1983 stream and all RTP streams the RTP stream depends on, or a 1984 specific set of the profile, tier, default level, and some 1985 constraints the receiver supports. 1987 The profile and some constraints are indicated collectively 1988 by profile-space, profile-id, profile-compatibility- 1989 indicator, and interop-constraints. The profile specifies 1990 the subset of coding tools that may have been used to 1991 generate the bitstream or that the receiver supports. 1993 Informative note: There are 32 values of profile-id, and 1994 there are 32 flags in profile-compatibility-indicator, 1995 each flag corresponding to one value of profile-id. 1996 According to HEVC version 1 in [HEVC], when more than 1997 one of the 32 flags is set for a bitstream, the 1998 bitstream would comply with all the profiles 1999 corresponding to the set flags. However, in a draft of 2000 HEVC version 2 in [HEVC draft v2], subclause A.3.5, 19 2001 Format Range Extensions profiles have been specified, 2002 all using the same value of profile-id (4), 2003 differentiated by some of the 48 bits in interop- 2004 constraints - this (rather unexpected way of profile 2005 signalling) means that one of the 32 flags may 2006 correspond to multiple profiles. To be able to support 2007 whatever HEVC extension profile that might be specified 2008 and indicated using profile-space, profile-id, profile- 2009 compatibility-indicator, and interop-constraints in the 2010 future, it would be safe to require symmetric use of 2011 these parameters in SDP offer/answer unless recv-sub- 2012 layer-id is included in the SDP answer for choosing one 2013 of the sub-layers offered. 2015 The tier is indicated by tier-flag. The default level is 2016 indicated by level-id. The tier and the default level 2017 specify the limits on values of syntax elements or 2018 arithmetic combinations of values of syntax elements that 2019 are followed when generating the bitstream or that the 2020 receiver supports. 2022 A set of profile-space, tier-flag, profile-id, profile- 2023 compatibility-indicator, interop-constraints, and level-id 2024 parameters ptlA is said to be consistent with another set 2025 of these parameters ptlB if any decoder that conforms to 2026 the profile, tier, level, and constraints indicated by ptlB 2027 can decode any bitstream that conforms to the profile, 2028 tier, level, and constraints indicated by ptlA. 2030 In SDP offer/answer, when the SDP answer does not include 2031 the recv-sub-layer-id parameter that is less than the 2032 sprop-sub-layer-id parameter in the SDP offer, the 2033 following applies: 2035 o The profile-space, tier-flag, profile-id, profile- 2036 compatibility-indicator, and interop-constraints 2037 parameters MUST be used symmetrically, i.e. the value 2038 of each of these parameters in the offer MUST be the 2039 same as that in the answer, either explicitly 2040 signalled or implicitly inferred. 2042 o The level-id parameter is changeable as long as the 2043 highest level indicated by the answer is either equal 2044 to or lower than that in the offer. Note that the 2045 highest level is indicated by level-id and max-recv- 2046 level-id together. 2048 In SDP offer/answer, when the SDP answer does include the 2049 recv-sub-layer-id parameter that is less than the sprop- 2050 sub-layer-id parameter in the SDP offer, the set of 2051 profile-space, tier-flag, profile-id, profile- 2052 compatibility-indicator, interop-constraints, and level-id 2053 parameters included in the answer MUST be consistent with 2054 that for the chosen sub-layer representation as indicated 2055 in the SDP offer, with the exception that the level-id 2056 parameter in the SDP answer is changable as long as the 2057 highest level indicated by the answer is either lower than 2058 or equal to that in the offer. 2060 More specifications of these parameters, including how they 2061 relate to the values of the profile, tier, and level syntax 2062 elements specified in [HEVC] are provided below. 2064 profile-space, profile-id: 2066 The value of profile-space MUST be in the range of 0 to 3, 2067 inclusive. The value of profile-id MUST be in the range of 2068 0 to 31, inclusive. 2070 When profile-space is not present, a value of 0 MUST be 2071 inferred. When profile-id is not present, a value of 1 2072 (i.e. the Main profile) MUST be inferred. 2074 When used to indicate properties of a bitstream, profile- 2075 space and profile-id are derived from the profile, tier, 2076 and level syntax elements in SPS or VPS NAL units as 2077 follows, where general_profile_space, general_profile_idc, 2078 sub_layer_profile_space[j], and sub_layer_profile_idc[j] 2079 are specified in [HEVC]: 2081 If the RTP stream is the highest RTP stream, the 2082 following applies: 2084 o profile_space = general_profile_space 2085 o profile_id = general_profile_idc 2087 Otherwise (the RTP stream is a dependee RTP stream), the 2088 following applies, with j being the value of the sprop- 2089 sub-layer-id parameter: 2091 o profile_space = sub_layer_profile_space[j] 2092 o profile_id = sub_layer_profile_idc[j] 2094 tier-flag, level-id: 2096 The value of tier-flag MUST be in the range of 0 to 1, 2097 inclusive. The value of level-id MUST be in the range of 0 2098 to 255, inclusive. 2100 If the tier-flag and level-id parameters are used to 2101 indicate properties of a bitstream, they indicate the tier 2102 and the highest level the bitstream complies with. 2104 If the tier-flag and level-id parameters are used for 2105 capability exchange, the following applies. If max-recv- 2106 level-id is not present, the default level defined by 2107 level-id indicates the highest level the codec wishes to 2108 support. Otherwise, max-recv-level-id indicates the 2109 highest level the codec supports for receiving. For either 2110 receiving or sending, all levels that are lower than the 2111 highest level supported MUST also be supported. 2113 If no tier-flag is present, a value of 0 MUST be inferred 2114 and if no level-id is present, a value of 93 (i.e. level 2115 3.1) MUST be inferred. 2117 When used to indicate properties of a bitstream, the tier- 2118 flag and level-id parameters are derived from the profile, 2119 tier, and level syntax elements in SPS or VPS NAL units as 2120 follows, where general_tier_flag, general_level_idc, 2121 sub_layer_tier_flag[j], and sub_layer_level_idc[j] are 2122 specified in [HEVC]: 2124 If the RTP stream is the highest RTP stream, the 2125 following applies: 2127 o tier-flag = general_tier_flag 2128 o level-id = general_level_idc 2130 Otherwise (the RTP stream is a dependee RTP stream), the 2131 following applies, with j being the value of the sprop- 2132 sub-layer-id parameter: 2134 o tier-flag = sub_layer_tier_flag[j] 2135 o level-id = sub_layer_level_idc[j] 2137 interop-constraints: 2139 A base16 [RFC4648] (hexadecimal) representation of six 2140 bytes of data, consisting of progressive_source_flag, 2141 interlaced_source_flag, non_packed_constraint_flag, 2142 frame_only_constraint_flag, and reserved_zero_44bits. 2144 If the interop-constraints parameter is not present, the 2145 following MUST be inferred: 2147 o progressive_source_flag = 1 2148 o interlaced_source_flag = 0 2149 o non_packed_constraint_flag = 1 2150 o frame_only_constraint_flag = 1 2151 o reserved_zero_44bits = 0 2153 When the interop-constraints parameter is used to indicate 2154 properties of a bitstream, the following applies, where 2155 general_progressive_source_flag, 2156 general_interlaced_source_flag, 2157 general_non_packed_constraint_flag, 2158 general_non_packed_constraint_flag, 2159 general_frame_only_constraint_flag, 2160 general_reserved_zero_44bits, 2161 sub_layer_progressive_source_flag[j], 2162 sub_layer_interlaced_source_flag[j], 2163 sub_layer_non_packed_constraint_flag[j], 2164 sub_layer_frame_only_constraint_flag[j], and 2165 sub_layer_reserved_zero_44bits[j] are specified in [HEVC]: 2167 If the RTP stream is the highest RTP stream, the 2168 following applies: 2170 o progressive_source_flag = 2171 general_progressive_source_flag 2172 o interlaced_source_flag = 2173 general_interlaced_source_flag 2174 o non_packed_constraint_flag = 2175 general_non_packed_constraint_flag 2176 o frame_only_constraint_flag = 2177 general_frame_only_constraint_flag 2178 o reserved_zero_44bits = general_reserved_zero_44bits 2180 Otherwise (the RTP stream is a dependee RTP stream), the 2181 following applies, with j being the value of the sprop- 2182 sub-layer-id parameter: 2184 o progressive_source_flag = 2185 sub_layer_progressive_source_flag[j] 2186 o interlaced_source_flag = 2187 sub_layer_interlaced_source_flag[j] 2188 o non_packed_constraint_flag = 2190 sub_layer_non_packed_constraint_flag[j] 2191 o frame_only_constraint_flag = 2193 sub_layer_frame_only_constraint_flag[j] 2194 o reserved_zero_44bits = 2195 sub_layer_reserved_zero_44bits[j] 2197 Using interop-constraints for capability exchange results 2198 in a requirement on any bitstream to be compliant with the 2199 interop-constraints. 2201 profile-compatibility-indicator: 2203 A base16 [RFC4648] representation of four bytes of data. 2205 When profile-compatibility-indicator is used to indicate 2206 properties of a bitstream, the following applies, where 2207 general_profile_compatibility_flag[j] and 2208 sub_layer_profile_compatibility_flag[i][j] are specified in 2209 [HEVC]: 2211 The profile-compatibility-indicator in this case 2212 indicates additional profiles to the profile defined by 2213 profile_space, profile_id, and interop-constraints the 2214 bitstream conforms to. A decoder that conforms to any 2215 of all the profiles the bitstream conforms to would be 2216 capable of decoding the bitstream. These additional 2217 profiles are defined by profile-space, each set bit of 2218 profile-compatibility-indicator, and interop- 2219 constraints. 2221 If the RTP stream is the highest RTP stream, the 2222 following applies for each value of j in the range of 0 2223 to 31, inclusive: 2225 o bit j of profile-compatibility-indicator = 2226 general_profile_compatibility_flag[j] 2228 Otherwise (the RTP stream is a dependee RTP stream), the 2229 following applies for i equal to sprop-sub-layer-id and 2230 for each value of j in the range of 0 to 31, inclusive: 2232 o bit j of profile-compatibility-indicator = 2233 sub_layer_profile_compatibility_flag[i][j] 2235 Using profile-compatibility-indicator for capability 2236 exchange results in a requirement on any bitstream to be 2237 compliant with the profile-compatibility-indicator. This 2238 is intended to handle cases where any future HEVC profile 2239 is defined as an intersection of two or more profiles. 2241 If this parameter is not present, this parameter defaults 2242 to the following: bit j, with j equal to profile-id, of 2243 profile-compatibility-indicator is inferred to be equal to 2244 1, and all other bits are inferred to be equal to 0. 2246 sprop-sub-layer-id: 2248 This parameter MAY be used to indicate the highest allowed 2249 value of TID in the bitstream. When not present, the value 2250 of sprop-sub-layer-id is inferred to be equal to 6. 2252 The value of sprop-sub-layer-id MUST be in the range of 0 2253 to 6, inclusive. 2255 recv-sub-layer-id: 2257 This parameter MAY be used to signal a receiver's choice of 2258 the offered or declared sub-layer representations in the 2259 sprop-vps. The value of recv-sub-layer-id indicates the 2260 TID of the highest sub-layer of the bitstream that a 2261 receiver supports. When not present, the value of recv- 2262 sub-layer-id is inferred to be equal to the value of the 2263 sprop-sub-layer-id parameter in the SDP offer. 2265 The value of recv-sub-layer-id MUST be in the range of 0 to 2266 6, inclusive. 2268 max-recv-level-id: 2270 This parameter MAY be used to indicate the highest level a 2271 receiver supports. The highest level the receiver supports 2272 is equal to the value of max-recv-level-id divided by 30. 2274 The value of max-recv-level-id MUST be in the range of 0 2275 to 255, inclusive. 2277 When max-recv-level-id is not present, the value is 2278 inferred to be equal to level-id. 2280 max-recv-level-id MUST NOT be present when the highest 2281 level the receiver supports is not higher than the default 2282 level. 2284 tx-mode: 2286 This parameter indicates whether the transmission mode is 2287 SRST, MRST, or MRMT. 2289 The value of tx-mode MUST be equal to "SRST", "MRST" or 2290 "MRMT". When not present, the value of tx-mode is inferred 2291 to be equal to "SRST". 2293 If the value is equal to "MRST", MRST MUST be in use. 2294 Otherwise, if the value is equal to "MRMT", MRMT MUST be in 2295 use. Otherwise (the value is equal to "SRST"), SRST MUST be 2296 in use. 2298 The value of tx-mode MUST be equal to "MRST" for all RTP 2299 streams in an MRST. 2301 The value of tx-mode MUST be equal to "MRMT" for all RTP 2302 streams in an MRMT. 2304 sprop-vps: 2306 This parameter MAY be used to convey any video parameter 2307 set NAL unit of the bitstream for out-of-band transmission 2308 of video parameter sets. The parameter MAY also be used 2309 for capability exchange and to indicate sub-stream 2310 characteristics (i.e. properties of sub-layer 2311 representations as defined in [HEVC]). The value of the 2312 parameter is a comma-separated (',') list of base64 2313 [RFC4648] representations of the video parameter set NAL 2314 units as specified in Section 7.3.2.1 of [HEVC]. 2316 The sprop-vps parameter MAY contain one or more than one 2317 video parameter set NAL unit. However, all other video 2318 parameter sets contained in the sprop-vps parameter MUST be 2319 consistent with the first video parameter set in the sprop- 2320 vps parameter. A video parameter set vpsB is said to be 2321 consistent with another video parameter set vpsA if any 2322 decoder that conforms to the profile, tier, level, and 2323 constraints indicated by the 12 bytes of data starting from 2324 the syntax element general_profile_space to the syntax 2325 element general_level_id, inclusive, in the first 2326 profile_tier_level( ) syntax structure in vpsA can decode 2327 any bitstream that conforms to the profile, tier, level, 2328 and constraints indicated by the 12 bytes of data starting 2329 from the syntax element general_profile_space to the syntax 2330 element general_level_id, inclusive, in the first 2331 profile_tier_level( ) syntax structure in vpsB. 2333 sprop-sps: 2335 This parameter MAY be used to convey sequence parameter set 2336 NAL units of the bitstream for out-of-band transmission of 2337 sequence parameter sets. The value of the parameter is a 2338 comma-separated (',') list of base64 [RFC4648] 2339 representations of the sequence parameter set NAL units as 2340 specified in Section 7.3.2.2 of [HEVC]. 2342 sprop-pps: 2344 This parameter MAY be used to convey picture parameter set 2345 NAL units of the bitstream for out-of-band transmission of 2346 picture parameter sets. The value of the parameter is a 2347 comma-separated (',') list of base64 [RFC4648] 2348 representations of the picture parameter set NAL units as 2349 specified in Section 7.3.2.3 of [HEVC]. 2351 sprop-sei: 2353 This parameter MAY be used to convey one or more SEI 2354 messages that describe bitstream characteristics. When 2355 present, a decoder can rely on the bitstream 2356 characteristics that are described in the SEI messages for 2357 the entire duration of the session, independently from the 2358 persistence scopes of the SEI messages as specified in 2359 [HEVC]. 2361 The value of the parameter is a comma-separated (',') list 2362 of base64 [RFC4648] representations of SEI NAL units as 2363 specified in Section 7.3.2.4 of [HEVC]. 2365 Informative note: Intentionally, no list of applicable 2366 or inapplicable SEI messages is specified here. 2367 Conveying certain SEI messages in sprop-sei may be 2368 sensible in some application scenarios and meaningless 2369 in others. However, a few examples are described below: 2371 1) In an environment where the bitstream was created 2372 from film-based source material, and no splicing is 2373 going to occur during the lifetime of the session, 2374 the film grain characteristics SEI message or the 2375 tone mapping information SEI message are likely 2376 meaningful, and sending them in sprop-sei rather than 2377 in the bitstream at each entry point may help saving 2378 bits and allows to configure the renderer only once, 2379 avoiding unwanted artifacts. 2380 2) The structure of pictures information SEI message in 2381 sprop-sei can be used to inform a decoder of 2382 information on the NAL unit types, picture order 2383 count values, and prediction dependencies of a 2384 sequence of pictures. Having such knowledge can be 2385 helpful for error recovery. 2386 3) Examples for SEI messages that would be meaningless 2387 to be conveyed in sprop-sei include the decoded 2388 picture hash SEI message (it is close to impossible 2389 that all decoded pictures have the same hash-tag), 2390 the display orientation SEI message when the device 2391 is a handheld device (as the display orientation may 2392 change when the handheld device is turned around), or 2393 the filler payload SEI message (as there is no point 2394 in just having more bits in SDP). 2396 max-lsr, max-lps, max-cpb, max-dpb, max-br, max-tr, max-tc: 2398 These parameters MAY be used to signal the capabilities of 2399 a receiver implementation. These parameters MUST NOT be 2400 used for any other purpose. The highest level (specified 2401 by max-recv-level-id) MUST be such that the receiver is 2402 fully capable of supporting. max-lsr, max-lps, max-cpb, 2403 max-dpb, max-br, max-tr, and max-tc MAY be used to indicate 2404 capabilities of the receiver that extend the required 2405 capabilities of the highest level, as specified below. 2407 When more than one parameter from the set (max-lsr, max- 2408 lps, max-cpb, max-dpb, max-br, max-tr, max-tc) is present, 2409 the receiver MUST support all signaled capabilities 2410 simultaneously. For example, if both max-lsr and max-br 2411 are present, the highest level with the extension of both 2412 the picture rate and bitrate is supported. That is, the 2413 receiver is able to decode bitstreams in which the luma 2414 sample rate is up to max-lsr (inclusive), the bitrate is up 2415 to max-br (inclusive), the coded picture buffer size is 2416 derived as specified in the semantics of the max-br 2417 parameter below, and the other properties comply with the 2418 highest level specified by max-recv-level-id. 2420 Informative note: When the OPTIONAL media type 2421 parameters are used to signal the properties of a 2422 bitstream, and max-lsr, max-lps, max-cpb, max-dpb, max- 2423 br, max-tr, and max-tc are not present, the values of 2424 profile-space, tier-flag, profile-id, profile- 2425 compatibility-indicator, interop-constraints, and level- 2426 id must always be such that the bitstream complies fully 2427 with the specified profile, tier, and level. 2429 max-lsr: 2430 The value of max-lsr is an integer indicating the maximum 2431 processing rate in units of luma samples per second. The 2432 max-lsr parameter signals that the receiver is capable of 2433 decoding video at a higher rate than is required by the 2434 highest level. 2436 When max-lsr is signaled, the receiver MUST be able to 2437 decode bitstreams that conform to the highest level, with 2438 the exception that the MaxLumaSR value in Table A-2 of 2439 [HEVC] for the highest level is replaced with the value of 2440 max-lsr. Senders MAY use this knowledge to send pictures 2441 of a given size at a higher picture rate than is indicated 2442 in the highest level. 2444 When not present, the value of max-lsr is inferred to be 2445 equal to the value of MaxLumaSR given in Table A-2 of 2446 [HEVC] for the highest level. 2448 The value of max-lsr MUST be in the range of MaxLumaSR to 2449 16 * MaxLumaSR, inclusive, where MaxLumaSR is given in 2450 Table A-2 of [HEVC] for the highest level. 2452 max-lps: 2453 The value of max-lps is an integer indicating the maximum 2454 picture size in units of luma samples. The max-lps 2455 parameter signals that the receiver is capable of decoding 2456 larger picture sizes than are required by the highest 2457 level. When max-lps is signaled, the receiver MUST be able 2458 to decode bitstreams that conform to the highest level, 2459 with the exception that the MaxLumaPS value in Table A-1 of 2460 [HEVC] for the highest level is replaced with the value of 2461 max-lps. Senders MAY use this knowledge to send larger 2462 pictures at a proportionally lower picture rate than is 2463 indicated in the highest level. 2465 When not present, the value of max-lps is inferred to be 2466 equal to the value of MaxLumaPS given in Table A-1 of 2467 [HEVC] for the highest level. 2469 The value of max-lps MUST be in the range of MaxLumaPS to 2470 16 * MaxLumaPS, inclusive, where MaxLumaPS is given in 2471 Table A-1 of [HEVC] for the highest level. 2473 max-cpb: 2474 The value of max-cpb is an integer indicating the maximum 2475 coded picture buffer size in units of CpbBrVclFactor bits 2476 for the VCL HRD parameters and in units of CpbBrNalFactor 2477 bits for the NAL HRD parameters, where CpbBrVclFactor and 2478 CpbBrNalFactor are defined in Section A.4 of [HEVC]. The 2479 max-cpb parameter signals that the receiver has more memory 2480 than the minimum amount of coded picture buffer memory 2481 required by the highest level. When max-cpb is signaled, 2482 the receiver MUST be able to decode bitstreams that conform 2483 to the highest level, with the exception that the MaxCPB 2484 value in Table A-1 of [HEVC] for the highest level is 2485 replaced with the value of max-cpb. Senders MAY use this 2486 knowledge to construct coded bitstreams with greater 2487 variation of bitrate than can be achieved with the MaxCPB 2488 value in Table A-1 of [HEVC]. 2490 When not present, the value of max-cpb is inferred to be 2491 equal to the value of MaxCPB given in Table A-1 of [HEVC] 2492 for the highest level. 2494 The value of max-cpb MUST be in the range of MaxCPB to 2495 16 * MaxCPB, inclusive, where MaxLumaCPB is given in Table 2496 A-1 of [HEVC] for the highest level. 2498 Informative note: The coded picture buffer is used in 2499 the hypothetical reference decoder (Annex C of HEVC). 2500 The use of the hypothetical reference decoder is 2501 recommended in HEVC encoders to verify that the produced 2502 bitstream conforms to the standard and to control the 2503 output bitrate. Thus, the coded picture buffer is 2504 conceptually independent of any other potential buffers 2505 in the receiver, including de-packetization and de- 2506 jitter buffers. The coded picture buffer need not be 2507 implemented in decoders as specified in Annex C of HEVC, 2508 but rather standard-compliant decoders can have any 2509 buffering arrangements provided that they can decode 2510 standard-compliant bitstreams. Thus, in practice, the 2511 input buffer for a video decoder can be integrated with 2512 de-packetization and de-jitter buffers of the receiver. 2514 max-dpb: 2515 The value of max-dpb is an integer indicating the maximum 2516 decoded picture buffer size in units decoded pictures at 2517 the MaxLumaPS for the highest level, i.e. the number of 2518 decoded pictures at the maximum picture size defined by the 2519 highest level. The value of max-dpb MUST be in the range 2520 of 1 to 16, respectively. The max-dpb parameter signals 2521 that the receiver has more memory than the minimum amount 2522 of decoded picture buffer memory required by default, which 2523 is MaxDpbPicBuf as defined in [HEVC] (equal to 6). When 2524 max-dpb is signaled, the receiver MUST be able to decode 2525 bitstreams that conform to the highest level, with the 2526 exception that the MaxDpbPicBuff value defined in [HEVC] as 2527 6 is replaced with the value of max-dpb. Consequently, a 2528 receiver that signals max-dpb MUST be capable of storing 2529 the following number of decoded pictures (MaxDpbSize) in 2530 its decoded picture buffer: 2532 if( PicSizeInSamplesY <= ( MaxLumaPS >> 2 ) ) 2533 MaxDpbSize = Min( 4 * max-dpb, 16 ) 2534 else if ( PicSizeInSamplesY <= ( MaxLumaPS >> 1 ) ) 2535 MaxDpbSize = Min( 2 * max-dpb, 16 ) 2536 else if ( PicSizeInSamplesY <= ( ( 3 * MaxLumaPS ) >> 2 2537 ) ) 2538 MaxDpbSize = Min( (4 * max-dpb) / 3, 16 ) 2539 else 2540 MaxDpbSize = max-dpb 2542 Wherein MaxLumaPS given in Table A-1 of [HEVC] for the 2543 highest level and PicSizeInSamplesY is the current size of 2544 each decoded picture in units of luma samples as defined in 2545 [HEVC]. 2547 The value of max-dpb MUST be greater than or equal to the 2548 value of MaxDpbPicBuf (i.e. 6) as defined in [HEVC]. 2549 Senders MAY use this knowledge to construct coded 2550 bitstreams with improved compression. 2552 When not present, the value of max-dpb is inferred to be 2553 equal to the value of MaxDpbPicBuf (i.e. 6) as defined in 2554 [HEVC]. 2556 Informative note: This parameter was added primarily to 2557 complement a similar codepoint in the ITU-T 2558 Recommendation H.245, so as to facilitate signaling 2559 gateway designs. The decoded picture buffer stores 2560 reconstructed samples. There is no relationship between 2561 the size of the decoded picture buffer and the buffers 2562 used in RTP, especially de-packetization and de-jitter 2563 buffers. 2565 max-br: 2566 The value of max-br is an integer indicating the maximum 2567 video bitrate in units of CpbBrVclFactor bits per second 2568 for the VCL HRD parameters and in units of CpbBrNalFactor 2569 bits per second for the NAL HRD parameters, where 2570 CpbBrVclFactor and CpbBrNalFactor are defined in Section 2571 A.4 of [HEVC]. 2573 The max-br parameter signals that the video decoder of the 2574 receiver is capable of decoding video at a higher bitrate 2575 than is required by the highest level. 2577 When max-br is signaled, the video codec of the receiver 2578 MUST be able to decode bitstreams that conform to the 2579 highest level, with the following exceptions in the limits 2580 specified by the highest level: 2582 o The value of max-br replaces the MaxBR value in Table A- 2583 2 of [HEVC] for the highest level. 2584 o When the max-cpb parameter is not present, the result of 2585 the following formula replaces the value of MaxCPB in 2586 Table A-1 of [HEVC]: 2588 (MaxCPB of the highest level) * max-br / (MaxBR of 2589 the highest level) 2591 For example, if a receiver signals capability for Main 2592 profile Level 2 with max-br equal to 2000, this indicates a 2593 maximum video bitrate of 2000 kbits/sec for VCL HRD 2594 parameters, a maximum video bitrate of 2200 kbits/sec for 2595 NAL HRD parameters, and a CPB size of 2000000 bits (2000000 2596 / 1500000 * 1500000). 2598 Senders MAY use this knowledge to send higher bitrate video 2599 as allowed in the level definition of Annex A of HEVC to 2600 achieve improved video quality. 2602 When not present, the value of max-br is inferred to be 2603 equal to the value of MaxBR given in Table A-2 of [HEVC] 2604 for the highest level. 2606 The value of max-br MUST be in the range of MaxBR to 2607 16 * MaxBR, inclusive, where MaxBR is given in Table A-2 of 2608 [HEVC] for the highest level. 2610 Informative note: This parameter was added primarily to 2611 complement a similar codepoint in the ITU-T 2612 Recommendation H.245, so as to facilitate signaling 2613 gateway designs. The assumption that the network is 2614 capable of handling such bitrates at any given time 2615 cannot be made from the value of this parameter. In 2616 particular, no conclusion can be drawn that the signaled 2617 bitrate is possible under congestion control 2618 constraints. 2620 max-tr: 2621 The value of max-tr is an integer indication the maximum 2622 number of tile rows. The max-tr parameter signals that the 2623 receiver is capable of decoding video with a larger number 2624 of tile rows than the value allowed by the highest level. 2626 When max-tr is signaled, the receiver MUST be able to 2627 decode bitstreams that conform to the highest level, with 2628 the exception that the MaxTileRows value in Table A-1 of 2629 [HEVC] for the highest level is replaced with the value of 2630 max-tr. 2632 Senders MAY use this knowledge to send pictures utilizing a 2633 larger number of tile rows than the value allowed by the 2634 highest level. 2636 When not present, the value of max-tr is inferred to be 2637 equal to the value of MaxTileRows given in Table A-1 of 2638 [HEVC] for the highest level. 2640 The value of max-tr MUST be in the range of MaxTileRows to 2641 16 * MaxTileRows, inclusive, where MaxTileRows is given in 2642 Table A-1 of [HEVC] for the highest level. 2644 max-tc: 2645 The value of max-tc is an integer indication the maximum 2646 number of tile columns. The max-tc parameter signals that 2647 the receiver is capable of decoding video with a larger 2648 number of tile columns than the value allowed by the 2649 highest level. 2651 When max-tc is signaled, the receiver MUST be able to 2652 decode bitstreams that conform to the highest level, with 2653 the exception that the MaxTileCols value in Table A-1 of 2654 [HEVC] for the highest level is replaced with the value of 2655 max-tc. 2657 Senders MAY use this knowledge to send pictures utilizing a 2658 larger number of tile columns than the value allowed by the 2659 highest level. 2661 When not present, the value of max-tc is inferred to be 2662 equal to the value of MaxTileCols given in Table A-1 of 2663 [HEVC] for the highest level. 2665 The value of max-tc MUST be in the range of MaxTileCols to 2666 16 * MaxTileCols, inclusive, where MaxTileCols is given in 2667 Table A-1 of [HEVC] for the highest level. 2669 max-fps: 2671 The value of max-fps is an integer indicating the maximum 2672 picture rate in units of pictures per 100 seconds that can 2673 be effectively processed by the receiver. The max-fps 2674 parameter MAY be used to signal that the receiver has a 2675 constraint in that it is not capable of processing video 2676 effectively at the full picture rate that is implied by the 2677 highest level and, when present, one or more of the 2678 parameters max-lsr, max-lps, and max-br. 2680 The value of max-fps is not necessarily the picture rate at 2681 which the maximum picture size can be sent, it constitutes 2682 a constraint on maximum picture rate for all resolutions. 2684 Informative note: The max-fps parameter is semantically 2685 different from max-lsr, max-lps, max-cpb, max-dpb, max- 2686 br, max-tr, and max-tc in that max-fps is used to signal 2687 a constraint, lowering the maximum picture rate from 2688 what is implied by other parameters. 2690 The encoder MUST use a picture rate equal to or less than 2691 this value. In cases where the max-fps parameter is absent 2692 the encoder is free to choose any picture rate according to 2693 the highest level and any signaled optional parameters. 2695 The value of max-fps MUST be smaller than or equal to the 2696 full picture rate that is implied by the highest level and, 2697 when present, one or more of the parameters max-lsr, max- 2698 lps, and max-br. 2700 sprop-max-don-diff: 2702 If tx-mode is equal to "SRST" and there is no NAL unit 2703 naluA that is followed in transmission order by any NAL 2704 unit preceding naluA in decoding order (i.e. the 2705 transmission order of the NAL units is the same as the 2706 decoding order), the value of this parameter MUST be equal 2707 to 0. 2709 Otherwise, if tx-mode is equal to "MRST" or "MRMT", the 2710 decoding order of the NAL units of all the RTP streams is 2711 the same as the NAL unit transmission order and the NAL 2712 unit output order, the value of this parameter MUST be 2713 equal to either 0 or 1. 2715 Otherwise, if tx-mode is equal to "MRST" or "MRMT" and the 2716 decoding order of the NAL units of all the RTP streams is 2717 the same as the NAL unit transmission order but not the 2718 same as the NAL unit output order, the value of this 2719 parameter MUST be equal to 1. 2721 Otherwise, this parameter specifies the maximum absolute 2722 difference between the decoding order number (i.e., AbsDon) 2723 values of any two NAL units naluA and naluB, where naluA 2724 follows naluB in decoding order and precedes naluB in 2725 transmission order. 2727 The value of sprop-max-don-diff MUST be an integer in the 2728 range of 0 to 32767, inclusive. 2730 When not present, the value of sprop-max-don-diff is 2731 inferred to be equal to 0. 2733 sprop-depack-buf-nalus: 2735 This parameter specifies the maximum number of NAL units 2736 that precede a NAL unit in transmission order and follow 2737 the NAL unit in decoding order. 2739 The value of sprop-depack-buf-nalus MUST be an integer in 2740 the range of 0 to 32767, inclusive. 2742 When not present, the value of sprop-depack-buf-nalus is 2743 inferred to be equal to 0. 2745 When sprop-max-don-diff is present and greater than 0, this 2746 parameter MUST be present and the value MUST be greater 2747 than 0. 2749 sprop-depack-buf-bytes: 2751 This parameter signals the required size of the de- 2752 packetization buffer in units of bytes. The value of the 2753 parameter MUST be greater than or equal to the maximum 2754 buffer occupancy (in units of bytes) of the de- 2755 packetization buffer as specified in Section 6. 2757 The value of sprop-depack-buf-bytes MUST be an integer in 2758 the range of 0 to 4294967295, inclusive. 2760 When sprop-max-don-diff is present and greater than 0, this 2761 parameter MUST be present and the value MUST be greater 2762 than 0. When not present, the value of sprop-depack-buf- 2763 bytes is inferred to be equal to 0. 2765 Informative note: The value of sprop-depack-buf-bytes 2766 indicates the required size of the de-packetization 2767 buffer only. When network jitter can occur, an 2768 appropriately sized jitter buffer has to be available as 2769 well. 2771 depack-buf-cap: 2773 This parameter signals the capabilities of a receiver 2774 implementation and indicates the amount of de-packetization 2775 buffer space in units of bytes that the receiver has 2776 available for reconstructing the NAL unit decoding order 2777 from NAL units carried in one or more RTP streams. A 2778 receiver is able to handle any RTP stream, and all RTP 2779 streams the RTP stream depends on, when present, for which 2780 the value of the sprop-depack-buf-bytes parameter is 2781 smaller than or equal to this parameter. 2783 When not present, the value of depack-buf-cap is inferred 2784 to be equal to 4294967295. The value of depack-buf-cap 2785 MUST be an integer in the range of 1 to 4294967295, 2786 inclusive. 2788 Informative note: depack-buf-cap indicates the maximum 2789 possible size of the de-packetization buffer of the 2790 receiver only. When network jitter can occur, an 2791 appropriately sized jitter buffer has to be available as 2792 well. 2794 sprop-segmentation-id: 2796 This parameter MAY be used to signal the segmentation tools 2797 present in the bitstream and that can be used for 2798 parallelization. The value of sprop-segmentation-id MUST 2799 be an integer in the range of 0 to 3, inclusive. When not 2800 present, the value of sprop-segmentation-id is inferred to 2801 be equal to 0. 2803 When sprop-segmentation-id is equal to 0, no information 2804 about the segmentation tools is provided. When sprop- 2805 segmentation-id is equal to 1, it indicates that slices are 2806 present in the bitstream. When sprop-segmentation-id is 2807 equal to 2, it indicates that tiles are present in the 2808 bitstream. When sprop-segmentation-id is equal to 3, it 2809 indicates that WPP is used in the bitstream. 2811 sprop-spatial-segmentation-idc: 2813 A base16 [RFC4648] representation of the syntax element 2814 min_spatial_segmentation_idc as specified in [HEVC]. This 2815 parameter MAY be used to describe parallelization 2816 capabilities of the bitstream. 2818 dec-parallel-cap: 2820 This parameter MAY be used to indicate the decoder's 2821 additional decoding capabilities given the presence of 2822 tools enabling parallel decoding, such as slices, tiles, 2823 and WPP, in the bitstream. The decoding capability of the 2824 decoder may vary with the setting of the parallel decoding 2825 tools present in the bitstream, e.g. the size of the tiles 2826 that are present in a bitstream. Therefore, multiple 2827 capability points may be provided, each indicating the 2828 minimum required decoding capability that is associated 2829 with a parallelism requirement, which is a requirement on 2830 the bitstream that enables parallel decoding. 2832 Each capability point is defined as a combination of 1) a 2833 parallelism requirement, 2) a profile (determined by 2834 profile-space and profile-id), 3) a highest level, and 4) a 2835 maximum processing rate, a maximum picture size, and a 2836 maximum video bitrate that may be equal to or greater than 2837 that determined by the highest level. The parameter's 2838 syntax in ABNF [RFC5234] is as follows: 2840 dec-parallel-cap = "dec-parallel-cap={" cap-point *("," 2841 cap-point) "}" 2843 cap-point = ("w" / "t") ":" spatial-seg-idc 1*(";" 2844 cap-parameter) 2846 spatial-seg-idc = 1*4DIGIT ; (1-4095) 2848 cap-parameter = tier-flag / level-id / max-lsr 2849 / max-lps / max-br 2851 tier-flag = "tier-flag" EQ ("0" / "1") 2853 level-id = "level-id" EQ 1*3DIGIT ; (0-255) 2855 max-lsr = "max-lsr" EQ 1*20DIGIT ; (0- 2856 18,446,744,073,709,551,615) 2858 max-lps = "max-lps" EQ 1*10DIGIT ; (0-4,294,967,295) 2859 max-br = "max-br" EQ 1*20DIGIT ; (0- 2860 18,446,744,073,709,551,615) 2862 EQ = "=" 2864 The set of capability points expressed by the dec-parallel- 2865 cap parameter is enclosed in a pair of curly braces ("{}"). 2866 Each set of two consecutive capability points is separated 2867 by a comma (','). Within each capability point, each set 2868 of two consecutive parameters, and when present, their 2869 values, is separated by a semicolon (';'). 2871 The profile of all capability points is determined by 2872 profile-space and profile-id that are outside the dec- 2873 parallel-cap parameter. 2875 Each capability point starts with an indication of the 2876 parallelism requirement, which consists of a parallel tool 2877 type, which may be equal to 'w' or 't', and a decimal value 2878 of the spatial-seg-idc parameter. When the type is 'w', 2879 the capability point is valid only for H.265 bitstreams 2880 with WPP in use, i.e. entropy_coding_sync_enabled_flag 2881 equal to 1. When the type is 't', the capability point is 2882 valid only for H.265 bitstreams with WPP not in use (i.e. 2883 entropy_coding_sync_enabled_flag equal to 0). The 2884 capability-point is valid only for H.265 bitstreams with 2885 min_spatial_segmentation_idc equal to or greater than 2886 spatial-seg-idc. 2888 After the parallelism requirement indication, each 2889 capability point continues with one or more pairs of 2890 parameter and value in any order for any of the following 2891 parameters: 2893 o tier-flag 2894 o level-id 2895 o max-lsr 2896 o max-lps 2897 o max-br 2899 At most one occurrence of each of the above five parameters 2900 is allowed within each capability point. 2902 The values of dec-parallel-cap.tier-flag and dec-parallel- 2903 cap.level-id for a capability point indicate the highest 2904 level of the capability point. The values of dec-parallel- 2905 cap.max-lsr, dec-parallel-cap.max-lps, and dec-parallel- 2906 cap.max-br for a capability point indicate the maximum 2907 processing rate in units of luma samples per second, the 2908 maximum picture size in units of luma samples, and the 2909 maximum video bitrate (in units of CpbBrVclFactor bits per 2910 second for the VCL HRD parameters and in units of 2911 CpbBrNalFactor bits per second for the NAL HRD parameters 2912 where CpbBrVclFactor and CpbBrNalFactor are defined in 2913 Section A.4 of [HEVC]). 2915 When not present, the value of dec-parallel-cap.tier-flag 2916 is inferred to be equal to the value of tier-flag outside 2917 the dec-parallel-cap parameter. When not present, the 2918 value of dec-parallel-cap.level-id is inferred to be equal 2919 to the value of max-recv-level-id outside the dec-parallel- 2920 cap parameter. When not present, the value of dec- 2921 parallel-cap.max-lsr, dec-parallel-cap.max-lps, or dec- 2922 parallel-cap.max-br is inferred to be equal to the value of 2923 max-lsr, max-lps, or max-br, respectively, outside the dec- 2924 parallel-cap parameter. 2926 The general decoding capability, expressed by the set of 2927 parameters outside of dec-parallel-cap, is defined as the 2928 capability point that is determined by the following 2929 combination of parameters: 1) the parallelism requirement 2930 corresponding to the value of sprop-segmentation-id equal 2931 to 0 for a bitstream, 2) the profile determined by profile- 2932 space, profile-id, profile-compatibility-indicator, and 2933 interop-constraints, 3) the tier and the highest level 2934 determined by tier-flag and max-recv-level-id, and 4) the 2935 maximum processing rate, the maximum picture size, and the 2936 maximum video bitrate determined by the highest level. The 2937 general decoding capability MUST NOT be included as one of 2938 the set of capability points in the dec-parallel-cap 2939 parameter. 2941 For example, the following parameters express the general 2942 decoding capability of 720p30 (Level 3.1) plus an 2943 additional decoding capability of 1080p30 (Level 4) given 2944 that the spatially largest tile or slice used in the 2945 bitstream is equal to or less than 1/3 of the picture size: 2947 a=fmtp:98 level-id=93;dec-parallel-cap={t:8;level- 2948 id=120} 2950 For another example, the following parameters express an 2951 additional decoding capability of 1080p30, using dec- 2952 parallel-cap.max-lsr and dec-parallel-cap.max-lps, given 2953 that WPP is used in the bitstream: 2955 a=fmtp:98 level-id=93;dec-parallel-cap={w:8; 2956 max-lsr=62668800;max-lps=2088960} 2958 Informative note: When min_spatial_segmentation_idc is 2959 present in a bitstream and WPP is not used, [HEVC] 2960 specifies that there is no slice or no tile in the 2961 bitstream containing more than 4 * PicSizeInSamplesY / 2962 ( min_spatial_segmentation_idc + 4 ) luma samples. 2964 include-dph: 2966 This parameter is used to indicate the capability and 2967 preference to utilize or include decoded picture hash (DPH) 2968 SEI messages (See Section D.3.19 of [HEVC]) in the 2969 bitstream. DPH SEI messages can be used to detect picture 2970 corruption so the receiver can request picture repair, see 2971 Section 8. The value is a comma separated list of hash 2972 types that is supported or requested to be used, each hash 2973 type provided as an unsigned integer value (0-255), with 2974 the hash types listed from most preferred to the least 2975 preferred. Example: "include-dph=0,2", which indicates the 2976 capability for MD5 (most preferred) and Checksum (less 2977 preferred). If the parameter is not included or the value 2978 contains no hash types, then no capability to utilize DPH 2979 SEI messages is assumed. Note that DPH SEI messages MAY 2980 still be included in the bitstream even when there is no 2981 declaration of capability to use them, as in general SEI 2982 messages do not affect the normative decoding process and 2983 decoders are allowed to ignore SEI messages. 2985 Encoding considerations: 2987 This type is only defined for transfer via RTP (RFC 3550). 2989 Security considerations: 2991 See Section 9 of RFC XXXX. 2993 Public specification: 2995 Please refer to Section 13 of RFC XXXX. 2997 Additional information: None 2999 File extensions: none 3001 Macintosh file type code: none 3003 Object identifier or OID: none 3005 Person & email address to contact for further information: 3007 Ye-Kui Wang (yekuiw@qti.qualcomm.com). 3009 Intended usage: COMMON 3011 Author: See Section 14 of RFC XXXX. 3013 Change controller: 3015 IETF Audio/Video Transport Payloads working group delegated 3016 from the IESG. 3018 7.2 SDP Parameters 3020 The receiver MUST ignore any parameter unspecified in this memo. 3022 7.2.1 Mapping of Payload Type Parameters to SDP 3024 The media type video/H265 string is mapped to fields in the 3025 Session Description Protocol (SDP) [RFC4566] as follows: 3027 o The media name in the "m=" line of SDP MUST be video. 3029 o The encoding name in the "a=rtpmap" line of SDP MUST be H265 3030 (the media subtype). 3032 o The clock rate in the "a=rtpmap" line MUST be 90000. 3034 o The OPTIONAL parameters "profile-space", "profile-id", "tier- 3035 flag", "level-id", "interop-constraints", "profile- 3036 compatibility-indicator", "sprop-sub-layer-id", "recv-sub- 3037 layer-id", "max-recv-level-id", "tx-mode", "max-lsr", "max- 3038 lps", "max-cpb", "max-dpb", "max-br", "max-tr", "max-tc", 3039 "max-fps", "sprop-max-don-diff", "sprop-depack-buf-nalus", 3040 "sprop-depack-buf-bytes", "depack-buf-cap", "sprop- 3041 segmentation-id", "sprop-spatial-segmentation-idc", "dec- 3042 parallel-cap", and "include-dph", when present, MUST be 3043 included in the "a=fmtp" line of SDP. This parameter is 3044 expressed as a media type string, in the form of a semicolon 3045 separated list of parameter=value pairs. 3047 o The OPTIONAL parameters "sprop-vps", "sprop-sps", and "sprop- 3048 pps", when present, MUST be included in the "a=fmtp" line of 3049 SDP or conveyed using the "fmtp" source attribute as specified 3050 in Section 6.3 of [RFC5576]. For a particular media format 3051 (i.e. RTP payload type), "sprop-vps" "sprop-sps", or "sprop- 3052 pps" MUST NOT be both included in the "a=fmtp" line of SDP and 3053 conveyed using the "fmtp" source attribute. When included in 3054 the "a=fmtp" line of SDP, these parameters are expressed as a 3055 media type string, in the form of a semicolon separated list 3056 of parameter=value pairs. When conveyed in the "a=fmtp" line 3057 of SDP for a particular payload type, the parameters "sprop- 3058 vps", "sprop-sps", and "sprop-pps" MUST be applied to each 3059 SSRC with the payload type. When conveyed using the "fmtp" 3060 source attribute, these parameters are only associated with 3061 the given source and payload type as parts of the "fmtp" 3062 source attribute. 3064 Informative note: Conveyance of "sprop-vps", "sprop-sps", 3065 and "sprop-pps" using the "fmtp" source attribute allows 3066 for out-of-band transport of parameter sets in topologies 3067 like Topo-Video-switch-MCU as specified in [RFC5117]. 3069 An example of media representation in SDP is as follows: 3071 m=video 49170 RTP/AVP 98 3072 a=rtpmap:98 H265/90000 3073 a=fmtp:98 profile-id=1; 3074 sprop-vps=