idnits 2.17.1 draft-ietf-payload-vp9-13.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 : ---------------------------------------------------------------------------- No issues found here. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year -- The document date (7 May 2021) is 1082 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) -- Possible downref: Non-RFC (?) normative reference: ref. 'VP9-BITSTREAM' Summary: 0 errors (**), 0 flaws (~~), 1 warning (==), 2 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 AVTCore Working Group J. Uberti 3 Internet-Draft S. Holmer 4 Intended status: Standards Track M. Flodman 5 Expires: 8 November 2021 D. Hong 6 Google 7 J. Lennox 8 8x8 / Jitsi 9 7 May 2021 11 RTP Payload Format for VP9 Video 12 draft-ietf-payload-vp9-13 14 Abstract 16 This specification describes an RTP payload format for the VP9 video 17 codec. The payload format has wide applicability, as it supports 18 applications from low bit-rate peer-to-peer usage, to high bit-rate 19 video conferences. It includes provisions for temporal and spatial 20 scalability. 22 Status of This Memo 24 This Internet-Draft is submitted in full conformance with the 25 provisions of BCP 78 and BCP 79. 27 Internet-Drafts are working documents of the Internet Engineering 28 Task Force (IETF). Note that other groups may also distribute 29 working documents as Internet-Drafts. The list of current Internet- 30 Drafts is at https://datatracker.ietf.org/drafts/current/. 32 Internet-Drafts are draft documents valid for a maximum of six months 33 and may be updated, replaced, or obsoleted by other documents at any 34 time. It is inappropriate to use Internet-Drafts as reference 35 material or to cite them other than as "work in progress." 37 This Internet-Draft will expire on 8 November 2021. 39 Copyright Notice 41 Copyright (c) 2021 IETF Trust and the persons identified as the 42 document authors. All rights reserved. 44 This document is subject to BCP 78 and the IETF Trust's Legal 45 Provisions Relating to IETF Documents (https://trustee.ietf.org/ 46 license-info) in effect on the date of publication of this document. 47 Please review these documents carefully, as they describe your rights 48 and restrictions with respect to this document. Code Components 49 extracted from this document must include Simplified BSD License text 50 as described in Section 4.e of the Trust Legal Provisions and are 51 provided without warranty as described in the Simplified BSD License. 53 Table of Contents 55 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 2 56 2. Conventions, Definitions and Acronyms . . . . . . . . . . . . 3 57 3. Media Format Description . . . . . . . . . . . . . . . . . . 3 58 4. Payload Format . . . . . . . . . . . . . . . . . . . . . . . 5 59 4.1. RTP Header Usage . . . . . . . . . . . . . . . . . . . . 5 60 4.2. VP9 Payload Descriptor . . . . . . . . . . . . . . . . . 6 61 4.2.1. Scalability Structure (SS): . . . . . . . . . . . . . 11 62 4.3. Frame Fragmentation . . . . . . . . . . . . . . . . . . . 12 63 4.4. Scalable encoding considerations . . . . . . . . . . . . 13 64 4.5. Examples of VP9 RTP Stream . . . . . . . . . . . . . . . 13 65 4.5.1. Reference picture use for scalable structure . . . . 13 66 5. Feedback Messages and Header Extensions . . . . . . . . . . . 14 67 5.1. Reference Picture Selection Indication (RPSI) . . . . . . 14 68 5.2. Full Intra Request (FIR) . . . . . . . . . . . . . . . . 15 69 5.3. Layer Refresh Request (LRR) . . . . . . . . . . . . . . . 15 70 6. Payload Format Parameters . . . . . . . . . . . . . . . . . . 16 71 6.1. SDP Parameters . . . . . . . . . . . . . . . . . . . . . 17 72 6.1.1. Mapping of Media Subtype Parameters to SDP . . . . . 18 73 6.1.2. Offer/Answer Considerations . . . . . . . . . . . . . 18 74 7. Media Type Definition . . . . . . . . . . . . . . . . . . . . 19 75 8. Security Considerations . . . . . . . . . . . . . . . . . . . 21 76 9. Congestion Control . . . . . . . . . . . . . . . . . . . . . 21 77 10. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 21 78 11. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 22 79 12. References . . . . . . . . . . . . . . . . . . . . . . . . . 22 80 12.1. Normative References . . . . . . . . . . . . . . . . . . 22 81 12.2. Informative References . . . . . . . . . . . . . . . . . 23 82 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 24 84 1. Introduction 86 This specification describes an RTP payload specification applicable 87 to the transmission of video streams encoded using the VP9 video 88 codec [VP9-BITSTREAM]. The format described in this document can be 89 used both in peer-to-peer and video conferencing applications. 91 The VP9 video codec was developed by Google, and is the successor to 92 its earlier VP8 [RFC6386] codec. Above the compression improvements 93 and other general enhancements above VP8, VP9 is also designed in a 94 way that allows spatially-scalable video encoding. 96 2. Conventions, Definitions and Acronyms 98 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 99 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 100 "OPTIONAL" in this document are to be interpreted as described in BCP 101 14 [RFC2119] [RFC8174] when, and only when, they appear in all 102 capitals, as shown here. 104 3. Media Format Description 106 The VP9 codec can maintain up to eight reference frames, of which up 107 to three can be referenced by any new frame. 109 VP9 also allows a frame to use another frame of a different 110 resolution as a reference frame. (Specifically, a frame may use any 111 references whose width and height are between 1/16th that of the 112 current frame and twice that of the current frame, inclusive.) This 113 allows internal resolution changes without requiring the use of key 114 frames. 116 These features together enable an encoder to implement various forms 117 of coarse-grained scalability, including temporal, spatial and 118 quality scalability modes, as well as combinations of these, without 119 the need for explicit scalable coding tools. 121 Temporal layers define different frame rates of video; spatial and 122 quality layers define different and possibly dependent 123 representations of a single input frame. Spatial layers allow a 124 frame to be encoded at different resolutions, whereas quality layers 125 allow a frame to be encoded at the same resolution but at different 126 qualities (and thus with different amounts of coding error). VP9 127 supports quality layers as spatial layers without any resolution 128 changes; hereinafter, the term "spatial layer" is used to represent 129 both spatial and quality layers. 131 This payload format specification defines how such temporal and 132 spatial scalability layers can be described and communicated. 134 Temporal and spatial scalability layers are associated with non- 135 negative integer IDs. The lowest layer of either type has an ID of 136 0, and is sometimes referred to as the "base" temporal or spatial 137 layer. 139 Layers are designed (and MUST be encoded) such that if any layer, and 140 all higher layers, are removed from the bitstream along either of the 141 two dimensions, the remaining bitstream is still correctly decodable. 143 For terminology, this document uses the term "frame" to refer to a 144 single encoded VP9 frame for a particular resolution/quality, and 145 "picture" to refer to all the representations (frames) at a single 146 instant in time. A picture thus consists of one or more frames, 147 encoding different spatial layers. 149 Within a picture, a frame with spatial layer ID equal to SID, where 150 SID > 0, can depend on a frame of the same picture with a lower 151 spatial layer ID. This "inter-layer" dependency can result in 152 additional coding gain compared to the case where only traditional 153 "inter-picture" dependency is used, where a frame depends on 154 previously coded frame in time. For simplicity, this payload format 155 assumes that, within a picture and if inter-layer dependency is used, 156 a spatial layer SID frame can depend only on the immediately previous 157 spatial layer SID-1 frame, when S > 0. Additionally, if inter- 158 picture dependency is used, a spatial layer SID frame is assumed to 159 only depend on a previously coded spatial layer SID frame. 161 Given above simplifications for inter-layer and inter-picture 162 dependencies, a flag (the D bit described below) is used to indicate 163 whether a spatial layer SID frame depends on the spatial layer SID-1 164 frame. Given the D bit, a receiver only needs to additionally know 165 the inter-picture dependency structure for a given spatial layer 166 frame in order to determine its decodability. Two modes of 167 describing the inter-picture dependency structure are possible: 168 "flexible mode" and "non-flexible mode". An encoder can only switch 169 between the two on the first packet of a key frame with temporal 170 layer ID equal to 0. 172 In flexible mode, each packet can contain up to 3 reference indices, 173 which identify all frames referenced by the frame transmitted in the 174 current packet for inter-picture prediction. This (along with the D 175 bit) enables a receiver to identify if a frame is decodable or not 176 and helps it understand the temporal layer structure. Since this is 177 signaled in each packet it makes it possible to have very flexible 178 temporal layer hierarchies and patterns which are changing 179 dynamically. 181 In non-flexible mode, the inter-picture dependency (the reference 182 indices) of a Picture Group (PG) MUST be pre-specified as part of the 183 scalability structure (SS) data. In this mode, each packet has an 184 index to refer to one of the described pictures in the PG, from which 185 the pictures referenced by the picture transmitted in the current 186 packet for inter-picture prediction can be identified. 188 (Note: A "Picture Group", as used in this document, is not the same 189 thing as a the term "Group of Pictures" as it is traditionally used 190 in video coding, i.e. to mean an independently-decoadable run of 191 pictures beginning with a keyframe.) 193 The SS data can also be used to specify the resolution of each 194 spatial layer present in the VP9 stream for both flexible and non- 195 flexible modes. 197 4. Payload Format 199 This section describes how the encoded VP9 bitstream is encapsulated 200 in RTP. To handle network losses usage of RTP/AVPF [RFC4585] is 201 RECOMMENDED. All integer fields in the specifications are encoded as 202 unsigned integers in network octet order. 204 4.1. RTP Header Usage 206 The general RTP payload format for VP9 is depicted below. 208 0 1 2 3 209 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 1 210 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 211 |V=2|P|X| CC |M| PT | sequence number | 212 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 213 | timestamp | 214 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 215 | synchronization source (SSRC) identifier | 216 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 217 | contributing source (CSRC) identifiers | 218 | .... | 219 +=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+ 220 | VP9 payload descriptor (integer #octets) | 221 : : 222 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 223 | : VP9 pyld hdr | | 224 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | 225 | | 226 + | 227 : Bytes 2..N of VP9 payload : 228 | | 229 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 230 | : OPTIONAL RTP padding | 231 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ 232 Figure 1 234 The VP9 payload descriptor will be described in Section 4.2; the VP9 235 payload header is described in [VP9-BITSTREAM]. OPTIONAL RTP padding 236 MUST NOT be included unless the P bit is set. The figure 237 specifically shows the format for the first packet in a frame. 238 Subsequent packets will not contain the VP9 payload header, and will 239 have later octets in the frame payload. 241 Marker bit (M): MUST be set to 1 for the final packet of the highest 242 spatial layer frame (the final packet of the picture), and 0 243 otherwise. Unless spatial scalability is in use for this picture, 244 this will have the same value as the E bit described below. Note 245 this bit MUST be set to 1 for the target spatial layer frame if a 246 stream is being rewritten to remove higher spatial layers. 248 Payload Type (PT): In line with the policy in Section 3 of 249 [RFC3551], applications using the VP9 RTP payload profile MUST 250 assign a dynamic payload type number to be used in each RTP 251 session and provide a mechanism to indicate the mapping. See 252 Section 6.1 for the mechanism to be used with the Session 253 Description Protocol (SDP) [RFC8866]. 255 Timestamp: The RTP timestamp indicates the time when the input frame 256 was sampled, at a clock rate of 90 kHz. If the input picture is 257 encoded with multiple layer frames, all of the frames of the 258 picture MUST have the same timestamp. 260 If a frame has the VP9 show_frame field set to 0 (i.e., it is 261 meant only to populate a reference buffer, without being output) 262 its timestamp MAY alternatively be set to be the same as the 263 subsequent frame with show_frame equal to 1. (This will be 264 convenient for playing out pre-encoded content packaged with VP9 265 "superframes", which typically bundle show_frame==0 frames with a 266 subsequent show_frame==1 frame.) Every frame with show_frame==1, 267 however, MUST have a unique timestamp modulo the 2^32 wrap of the 268 field. 270 The remaining RTP Fixed Header Fields (V, P, X, CC, sequence number, 271 SSRC and CSRC identifiers) are used as specified in Section 5.1 of 272 [RFC3550]. 274 4.2. VP9 Payload Descriptor 276 In flexible mode (with the F bit below set to 1), the first octets 277 after the RTP header are the VP9 payload descriptor, with the 278 following structure. 280 0 1 2 3 4 5 6 7 281 +-+-+-+-+-+-+-+-+ 282 |I|P|L|F|B|E|V|Z| (REQUIRED) 283 +-+-+-+-+-+-+-+-+ 284 I: |M| PICTURE ID | (REQUIRED) 285 +-+-+-+-+-+-+-+-+ 286 M: | EXTENDED PID | (RECOMMENDED) 287 +-+-+-+-+-+-+-+-+ 288 L: | TID |U| SID |D| (Conditionally RECOMMENDED) 289 +-+-+-+-+-+-+-+-+ -\ 290 P,F: | P_DIFF |N| (Conditionally REQUIRED) - up to 3 times 291 +-+-+-+-+-+-+-+-+ -/ 292 V: | SS | 293 | .. | 294 +-+-+-+-+-+-+-+-+ 296 Figure 2 298 In non-flexible mode (with the F bit below set to 0), The first 299 octets after the RTP header are the VP9 payload descriptor, with the 300 following structure. 302 0 1 2 3 4 5 6 7 303 +-+-+-+-+-+-+-+-+ 304 |I|P|L|F|B|E|V|Z| (REQUIRED) 305 +-+-+-+-+-+-+-+-+ 306 I: |M| PICTURE ID | (RECOMMENDED) 307 +-+-+-+-+-+-+-+-+ 308 M: | EXTENDED PID | (RECOMMENDED) 309 +-+-+-+-+-+-+-+-+ 310 L: | TID |U| SID |D| (Conditionally RECOMMENDED) 311 +-+-+-+-+-+-+-+-+ 312 | TL0PICIDX | (Conditionally REQUIRED) 313 +-+-+-+-+-+-+-+-+ 314 V: | SS | 315 | .. | 316 +-+-+-+-+-+-+-+-+ 318 Figure 3 320 I: Picture ID (PID) present. When set to one, the OPTIONAL PID MUST 321 be present after the mandatory first octet and specified as below. 322 Otherwise, PID MUST NOT be present. If the SS field was present 323 in the stream's most recent start of a keyframe (i.e., non- 324 flexible scalability mode is in use), then the PID MUST also be 325 present in every packet. 327 P: Inter-picture predicted frame. When set to zero, the frame does 328 not utilize inter-picture prediction. In this case, up-switching 329 to a current spatial layer's frame is possible from directly lower 330 spatial layer frame. P SHOULD also be set to zero when encoding a 331 layer synchronization frame in response to an LRR 332 [I-D.ietf-avtext-lrr] message (see Section 5.3). When P is set to 333 zero, the TID field (described below) MUST also be set to 0 (if 334 present). Note that the P bit does not forbid intra-picture, 335 inter-layer prediction from earlier frames of the same picture, if 336 any. 338 L: Layer indices present. When set to one, the one or two octets 339 following the mandatory first octet and the PID (if present) is as 340 described by "Layer indices" below. If the F bit (described 341 below) is set to 1 (indicating flexible mode), then only one octet 342 is present for the layer indices. Otherwise if the F bit is set 343 to 0 (indicating non-flexible mode), then two octets are present 344 for the layer indices. 346 F: Flexible mode. F set to one indicates flexible mode and if the P 347 bit is also set to one, then the octets following the mandatory 348 first octet, the PID, and layer indices (if present) are as 349 described by "Reference indices" below. This MUST only be set to 350 1 if the I bit is also set to one; if the I bit is set to zero, 351 then this MUST also be set to zero and ignored by receivers. The 352 value of this F bit MUST only change on the first packet of a key 353 picture. A key picture is a picture whose base spatial layer 354 frame is a key frame, and which thus completely resets the encoder 355 state. This packet will have its P bit equal to zero, SID or D 356 bit (described below) equal to zero, and B bit (described below) 357 equal to 1. 359 B: Start of a frame. MUST be set to 1 if the first payload octet of 360 the RTP packet is the beginning of a new VP9 frame, and MUST NOT 361 be 1 otherwise. Note that this frame might not be the first frame 362 of a picture. 364 E: End of a frame. MUST be set to 1 for the final RTP packet of a 365 VP9 frame, and 0 otherwise. This enables a decoder to finish 366 decoding the frame, where it otherwise may need to wait for the 367 next packet to explicitly know that the frame is complete. Note 368 that, if spatial scalability is in use, more frames from the same 369 picture may follow; see the description of the M bit above. 371 V: Scalability structure (SS) data present. When set to one, the 372 OPTIONAL SS data MUST be present in the payload descriptor. 373 Otherwise, the SS data MUST NOT be present. 375 Z: Not a reference frame for upper spatial layers. If set to 1, 376 indicates that frames with higher spatial layers SID+1 of the 377 current and following pictures do not depend on the current 378 spatial layer SID frame. This enables a decoder which is 379 targeting a higher spatial layer to know that it can safely 380 discard this packet's frame without processing it, without having 381 to wait for the "D" bit in the higher-layer frame (see below). 383 The mandatory first octet is followed by the extension data fields 384 that are enabled: 386 M: The most significant bit of the first octet is an extension flag. 387 The field MUST be present if the I bit is equal to one. If set, 388 the PID field MUST contain 15 bits; otherwise, it MUST contain 7 389 bits. See PID below. 391 Picture ID (PID): Picture ID represented in 7 or 15 bits, depending 392 on the M bit. This is a running index of the pictures. The field 393 MUST be present if the I bit is equal to one. If M is set to 394 zero, 7 bits carry the PID; else if M is set to one, 15 bits carry 395 the PID in network byte order. The sender may choose between a 7- 396 or 15-bit index. The PID SHOULD start on a random number, and 397 MUST wrap after reaching the maximum ID (0x7f or 0x7fff depending 398 on the index size chosen). The receiver MUST NOT assume that the 399 number of bits in PID stay the same through the session. 401 In the non-flexible mode (when the F bit is set to 0), this PID is 402 used as an index to the picture group (PG) specified in the SS 403 data below. In this mode, the PID of the key frame corresponds to 404 the first specified frame in the PG. Then subsequent PIDs are 405 mapped to subsequently specified frames in the PG (modulo N_G, 406 specified in the SS data below), respectively. 408 All frames of the same picture MUST have the same PID value. 410 Frames (and their corresponding pictures) with the VP9 show_frame 411 field equal to 0 MUST have distinct PID values from subsequent 412 pictures with show_frame equal to 1. Thus, a Picture as defined 413 in this specification is different than a VP9 Superframe. 415 All frames of the same picture MUST have the same value for 416 show_frame. 418 Layer indices: This information is optional but RECOMMENDED whenever 419 encoding with layers. For both flexible and non-flexible modes, 420 one octet is used to specify a layer frame's temporal layer ID 421 (TID) and spatial layer ID (SID) as shown both in Figure 2 and 422 Figure 3. Additionally, a bit (U) is used to indicate that the 423 current frame is a "switching up point" frame. Another bit (D) is 424 used to indicate whether inter-layer prediction is used for the 425 current frame. 427 In the non-flexible mode (when the F bit is set to 0), another 428 octet is used to represent temporal layer 0 index (TL0PICIDX), as 429 depicted in Figure 3. The TL0PICIDX is present so that all 430 minimally required frames - the base temporal layer frames - can 431 be tracked. 433 The TID and SID fields indicate the temporal and spatial layers 434 and can help middleboxes and endpoints quickly identify which 435 layer a packet belongs to. 437 TID: The temporal layer ID of current frame. In the case of non- 438 flexible mode, if PID is mapped to a picture in a specified PG, 439 then the value of TID MUST match the corresponding TID value of 440 the mapped picture in the PG. 442 U: Switching up point. If this bit is set to 1 for the current 443 picture with temporal layer ID equal to TID, then "switch up" 444 to a higher frame rate is possible as subsequent higher 445 temporal layer pictures will not depend on any picture before 446 the current picture (in coding order) with temporal layer ID 447 greater than TID. 449 SID: The spatial layer ID of current frame. Note that frames 450 with spatial layer SID > 0 may be dependent on decoded spatial 451 layer SID-1 frame within the same picture. Different frames of 452 the same picture MUST have distinct spatial layer IDs, and 453 frames' spatial layers MUST appear in increasing order within 454 the frame. 456 D: Inter-layer dependency used. MUST be set to one if and only 457 if the current spatial layer SID frame depends on spatial layer 458 SID-1 frame of the same picture, otherwise MUST set to zero. 459 For the base layer frame (with SID equal to 0), this D bit MUST 460 be set to zero. 462 TL0PICIDX: 8 bits temporal layer zero index. TL0PICIDX is only 463 present in the non-flexible mode (F = 0). This is a running 464 index for the temporal base layer pictures, i.e., the pictures 465 with TID set to 0. If TID is larger than 0, TL0PICIDX 466 indicates which temporal base layer picture the current picture 467 depends on. TL0PICIDX MUST be incremented when TID is equal to 468 0. The index SHOULD start on a random number, and MUST restart 469 at 0 after reaching the maximum number 255. 471 Reference indices: When P and F are both set to one, indicating a 472 non-key frame in flexible mode, then at least one reference index 473 MUST be specified as below. Additional reference indices (total 474 of up to 3 reference indices are allowed) may be specified using 475 the N bit below. When either P or F is set to zero, then no 476 reference index is specified. 478 P_DIFF: The reference index (in 7 bits) specified as the relative 479 PID from the current picture. For example, when P_DIFF=3 on a 480 packet containing the picture with PID 112 means that the 481 picture refers back to the picture with PID 109. This 482 calculation is done modulo the size of the PID field, i.e., 483 either 7 or 15 bits. 485 N: 1 if there is additional P_DIFF following the current P_DIFF. 487 4.2.1. Scalability Structure (SS): 489 The scalability structure (SS) data describes the resolution of each 490 frame within a picture as well as the inter-picture dependencies for 491 a picture group (PG). If the VP9 payload descriptor's "V" bit is 492 set, the SS data is present in the position indicated in Figure 2 and 493 Figure 3. 495 +-+-+-+-+-+-+-+-+ 496 V: | N_S |Y|G|-|-|-| 497 +-+-+-+-+-+-+-+-+ -\ 498 Y: | WIDTH | (OPTIONAL) . 499 + + . 500 | | (OPTIONAL) . 501 +-+-+-+-+-+-+-+-+ . - N_S + 1 times 502 | HEIGHT | (OPTIONAL) . 503 + + . 504 | | (OPTIONAL) . 505 +-+-+-+-+-+-+-+-+ -/ 506 G: | N_G | (OPTIONAL) 507 +-+-+-+-+-+-+-+-+ -\ 508 N_G: | TID |U| R |-|-| (OPTIONAL) . 509 +-+-+-+-+-+-+-+-+ -\ . - N_G times 510 | P_DIFF | (OPTIONAL) . - R times . 511 +-+-+-+-+-+-+-+-+ -/ -/ 513 Figure 4 515 N_S: N_S + 1 indicates the number of spatial layers present in the 516 VP9 stream. 518 Y: Each spatial layer's frame resolution present. When set to one, 519 the OPTIONAL WIDTH (2 octets) and HEIGHT (2 octets) MUST be 520 present for each layer frame. Otherwise, the resolution MUST NOT 521 be present. 523 G: PG description present flag. 525 -: Bit reserved for future use. MUST be set to zero and MUST be 526 ignored by the receiver. 528 N_G: N_G indicates the number of pictures in a Picture Group (PG). 529 If N_G is greater than 0, then the SS data allows the inter- 530 picture dependency structure of the VP9 stream to be pre-declared, 531 rather than indicating it on the fly with every packet. If N_G is 532 greater than 0, then for N_G pictures in the PG, each picture's 533 temporal layer ID (TID), switch up point (U), and the R reference 534 indices (P_DIFFs) are specified. 536 The first picture specified in the PG MUST have TID set to 0. 538 G set to 0 or N_G set to 0 indicates that either there is only one 539 temporal layer or no fixed inter-picture dependency information is 540 present going forward in the bitstream. 542 Note that for a given picture, all frames follow the same inter- 543 picture dependency structure. However, the frame rate of each 544 spatial layer can be different from each other and this can be 545 controlled with the use of the D bit described above. The 546 specified dependency structure in the SS data MUST be for the 547 highest frame rate layer. 549 In a scalable stream sent with a fixed pattern, the SS data SHOULD be 550 included in the first packet of every key frame. This is a packet 551 with P bit equal to zero, SID or D bit equal to zero, and B bit equal 552 to 1. The SS data MUST only be changed on the picture that 553 corresponds to the first picture specified in the previous SS data's 554 PG (if the previous SS data's N_G was greater than 0). 556 4.3. Frame Fragmentation 558 VP9 frames are fragmented into packets, in RTP sequence number order, 559 beginning with a packet with the B bit set, and ending with a packet 560 with the E bit set. There is no mechanism for finer-grained access 561 to parts of a VP9 frame. 563 4.4. Scalable encoding considerations 565 In addition to the use of reference frames, VP9 has several 566 additional forms of inter-frame dependencies, largely involving 567 probability tables for the entropy and tree encoders. In VP9 syntax, 568 the syntax element "error_resilient_mode" resets this additional 569 inter-frame data, allowing a frame's syntax to be decoded 570 independently. 572 Due to the requirements of scalable streams, a VP9 encoder producing 573 a scalable stream needs to ensure that a frame does not depend on a 574 previous frame (of the same or a previous picture) that can 575 legitimately be removed from the stream. Thus, a frame that follows 576 a removable frame (in full decode order) MUST be encoded with 577 "error_resilient_mode" set to true. 579 For spatially-scalable streams, this means that 580 "error_resilient_mode" needs to be turned on for the base spatial 581 layer; it can however be turned off for higher spatial layers, 582 assuming they are sent with inter-layer dependency (i.e. with the "D" 583 bit set). For streams that are only temporally-scalable without 584 spatial scalability, "error_resilient_mode" can additionally be 585 turned off for any picture that immediately follows a temporal layer 586 0 frame. 588 4.5. Examples of VP9 RTP Stream 590 4.5.1. Reference picture use for scalable structure 592 As discussed in Section 3, the VP9 codec can maintain up to eight 593 reference frames, of which up to three can be referenced or updated 594 by any new frame. This section illustrates one way that a scalable 595 structure (with three spatial layers and three temporal layers) can 596 be constructed using these reference frames. 598 +==========+=========+============+=========+ 599 | Temporal | Spatial | References | Updates | 600 +==========+=========+============+=========+ 601 | 0 | 0 | 0 | 0 | 602 +----------+---------+------------+---------+ 603 | 0 | 1 | 0,1 | 1 | 604 +----------+---------+------------+---------+ 605 | 0 | 2 | 1,2 | 2 | 606 +----------+---------+------------+---------+ 607 | 2 | 0 | 0 | 6 | 608 +----------+---------+------------+---------+ 609 | 2 | 1 | 1,6 | 7 | 610 +----------+---------+------------+---------+ 611 | 2 | 2 | 2,7 | - | 612 +----------+---------+------------+---------+ 613 | 1 | 0 | 0 | 3 | 614 +----------+---------+------------+---------+ 615 | 1 | 1 | 1,3 | 4 | 616 +----------+---------+------------+---------+ 617 | 1 | 2 | 2,4 | 5 | 618 +----------+---------+------------+---------+ 619 | 2 | 0 | 3 | 6 | 620 +----------+---------+------------+---------+ 621 | 2 | 1 | 4,6 | 7 | 622 +----------+---------+------------+---------+ 623 | 2 | 2 | 5,7 | - | 624 +----------+---------+------------+---------+ 626 Table 1: Example scalability structure 628 This structure is constructed such that the "U" bit can always be 629 set. 631 5. Feedback Messages and Header Extensions 633 5.1. Reference Picture Selection Indication (RPSI) 635 The reference picture selection index is a payload-specific feedback 636 message defined within the RTCP-based feedback format. The RPSI 637 message is generated by a receiver and can be used in two ways. 638 Either it can signal a preferred reference picture when a loss has 639 been detected by the decoder -- preferably then a reference that the 640 decoder knows is perfect -- or, it can be used as positive feedback 641 information to acknowledge correct decoding of certain reference 642 pictures. The positive feedback method is useful for VP9 used for 643 point to point (unicast) communication. The use of RPSI for VP9 is 644 preferably combined with a special update pattern of the codec's two 645 special reference frames -- the golden frame and the altref frame -- 646 in which they are updated in an alternating leapfrog fashion. When a 647 receiver has received and correctly decoded a golden or altref frame, 648 and that frame had a PictureID in the payload descriptor, the 649 receiver can acknowledge this simply by sending an RPSI message back 650 to the sender. The message body (i.e., the "native RPSI bit string" 651 in [RFC4585]) is simply the PictureID of the received frame. 653 Note: because all frames of the same picture must have the same 654 inter-picture reference structure, there is no need for a message to 655 specify which frame is being selected. 657 5.2. Full Intra Request (FIR) 659 The Full Intra Request (FIR) [RFC5104] RTCP feedback message allows a 660 receiver to request a full state refresh of an encoded stream. 662 Upon receipt of an FIR request, a VP9 sender MUST send a picture with 663 a keyframe for its spatial layer 0 layer frame, and then send frames 664 without inter-picture prediction (P=0) for any higher layer frames. 666 5.3. Layer Refresh Request (LRR) 668 The Layer Refresh Request [I-D.ietf-avtext-lrr] allows a receiver to 669 request a single layer of a spatially or temporally encoded stream to 670 be refreshed, without necessarily affecting the stream's other 671 layers. 673 +---------------+---------------+ 674 |0|1|2|3|4|5|6|7|0|1|2|3|4|5|6|7| 675 +---------------+---------+-----+ 676 | RES | TID | RES | SID | 677 +---------------+---------+-----+ 679 Figure 5 681 Figure 5 shows the format of LRR's layer index fields for VP9 682 streams. The two "RES" fields MUST be set to 0 on transmission and 683 ingnored on reception. See Section 4.2 for details on the TID and 684 SID fields. 686 Identification of a layer refresh frame can be derived from the 687 reference IDs of each frame by backtracking the dependency chain 688 until reaching a point where only decodable frames are being 689 referenced. Therefore it's recommended for both the flexible and the 690 non-flexible mode that, when upgrade frames are being encoded in 691 response to a LRR, those packets should contain layer indices and the 692 reference fields so that the decoder or an MCU can make this 693 derivation. 695 Example: 697 LRR {1,0}, {2,1} is sent by an MCU when it is currently relaying 698 {1,0} to a receiver and which wants to upgrade to {2,1}. In response 699 the encoder should encode the next frames in layers {1,1} and {2,1} 700 by only referring to frames in {1,0}, or {0,0}. 702 In the non-flexible mode, periodic upgrade frames can be defined by 703 the layer structure of the SS, thus periodic upgrade frames can be 704 automatically identified by the picture ID. 706 6. Payload Format Parameters 708 This payload format has three optional parameters, "max-fr", "max- 709 fs", and "profile-id". 711 The max-fr and max-fs parameters are used to signal the capabilities 712 of a receiver implementation. If the implementation is willing to 713 receive media, both parameters MUST be provided. These parameters 714 MUST NOT be used for any other purpose. A media sender SHOULD NOT 715 send media with a frame rate or frame size exceeding the max-fr and 716 max-fs values signaled. (There may be scenarios, such as pre-encoded 717 media or selective forwarding middleboxes [RFC7667], where a media 718 sender does not have media available that fits within a receivers 719 max-fs and max-fr value; in such scenarios, a sender MAY exceed the 720 signaled values.) 722 max-fr: The value of max-fr is an integer indicating the maximum 723 frame rate in units of frames per second that the decoder is 724 capable of decoding. 726 max-fs: The value of max-fs is an integer indicating the maximum 727 frame size in units of macroblocks that the decoder is capable of 728 decoding. 730 The decoder is capable of decoding this frame size as long as the 731 width and height of the frame in macroblocks are less than 732 int(sqrt(max-fs * 8)) - for instance, a max-fs of 1200 (capable of 733 supporting 640x480 resolution) will support widths and heights up 734 to 1552 pixels (97 macroblocks). 736 profile-id: The value of profile-id is an integer indicating the 737 default coding profile, the subset of coding tools that may have 738 been used to generate the stream or that the receiver supports). 739 Table 2 lists all of the profiles defined in section 7.2 of 740 [VP9-BITSTREAM] and the corresponding integer values to be used. 742 If no profile-id is present, Profile 0 MUST be inferred. (The 743 profile-id parameter was added relatively late in the development 744 of this specification, so some existing implementations may not 745 send it.) 747 Informative note: See Table 3 for capabilities of coding profiles 748 defined in section 7.2 of [VP9-BITSTREAM]. 750 A receiver MUST ignore any parameter unspecified in this 751 specification. 753 +=========+============+ 754 | Profile | profile-id | 755 +=========+============+ 756 | 0 | 0 | 757 +---------+------------+ 758 | 1 | 1 | 759 +---------+------------+ 760 | 2 | 2 | 761 +---------+------------+ 762 | 3 | 3 | 763 +---------+------------+ 765 Table 2: Table of 766 profile-id integer 767 values representing 768 the VP9 profile 769 corresponding to the 770 set of coding tools 771 supported. 773 +=========+===========+=================+==========================+ 774 | Profile | Bit Depth | SRGB Colorspace | Chroma Subsampling | 775 +=========+===========+=================+==========================+ 776 | 0 | 8 | No | YUV 4:2:0 | 777 +---------+-----------+-----------------+--------------------------+ 778 | 1 | 8 | Yes | YUV 4:2:0,4:4:0 or 4:4:4 | 779 +---------+-----------+-----------------+--------------------------+ 780 | 2 | 10 or 12 | No | YUV 4:2:0 | 781 +---------+-----------+-----------------+--------------------------+ 782 | 3 | 10 or 12 | Yes | YUV 4:2:0,4:4:0 or 4:4:4 | 783 +---------+-----------+-----------------+--------------------------+ 785 Table 3: Table of profile capabilities. 787 6.1. SDP Parameters 788 6.1.1. Mapping of Media Subtype Parameters to SDP 790 The media type video/VP9 string is mapped to fields in the Session 791 Description Protocol (SDP) [RFC8866] as follows: 793 * The media name in the "m=" line of SDP MUST be video. 795 * The encoding name in the "a=rtpmap" line of SDP MUST be VP9 (the 796 media subtype). 798 * The clock rate in the "a=rtpmap" line MUST be 90000. 800 * The parameters "max-fr" and "max-fs" MUST be included in the 801 "a=fmtp" line of SDP if the receiver wishes to declare its 802 receiver capabilities. These parameters are expressed as a media 803 subtype string, in the form of a semicolon separated list of 804 parameter=value pairs. 806 * The OPTIONAL parameter profile-id, when present, SHOULD be 807 included in the "a=fmtp" line of SDP. This parameter is expressed 808 as a media subtype string, in the form of a parameter=value pair. 809 When the parameter is not present, a value of 0 MUST be inferred 810 for profile-id. 812 6.1.1.1. Example 814 An example of media representation in SDP is as follows: 816 m=video 49170 RTP/AVPF 98 817 a=rtpmap:98 VP9/90000 818 a=fmtp:98 max-fr=30;max-fs=3600;profile-id=0 820 6.1.2. Offer/Answer Considerations 822 When VP9 is offered over RTP using SDP in an Offer/Answer model 823 [RFC3264] for negotiation for unicast usage, the following 824 limitations and rules apply: 826 * The parameter identifying a media format configuration for VP9 is 827 profile-id. This media format configuration parameter MUST be 828 used symmetrically; that is, the answerer MUST either maintain 829 this configuration parameter or remove the media format (payload 830 type) completely if it is not supported. 832 * The max-fr and max-fs parameters are used declaratively to 833 describe receiver capabilities, even in the Offer/Answer model. 834 The values in an answer are used to describe the answerer's 835 capabilities, and thus their values are set independently of the 836 values in the offer. 838 * To simplify the handling and matching of these configurations, the 839 same RTP payload type number used in the offer SHOULD also be used 840 in the answer and in a subsequent offer, as specified in 841 [RFC3264]. An answer or subsequent offer MUST NOT contain the 842 payload type number used in the offer unless the profile-id value 843 is exactly the same as in the original offer. However, max-fr and 844 max-fs parameters MAY be changed in subsequent offers and answers, 845 with the same payload type number, if an endpoint wishes to change 846 its declared receiver capabilties. 848 7. Media Type Definition 850 This registration is done using the template defined in [RFC6838] and 851 following [RFC4855]. 853 Type name: 854 video 856 Subtype name: 857 VP9 859 Required parameters: 860 N/A. 862 Optional parameters: 863 There are three optional parameters, "max-fr", "max-fs", and 864 "profile-id". See Section 6 for their definition. 866 Encoding considerations: 867 This media type is framed in RTP and contains binary data; see 868 Section 4.8 of [RFC6838]. 870 Security considerations: 871 See Section 8 of RFC xxxx. 873 [RFC Editor: Upon publication as an RFC, please replace "XXXX" 874 with the number assigned to this document and remove this note.] 876 Interoperability considerations: 877 None. 879 Published specification: 881 VP9 bitstream format [VP9-BITSTREAM] and RFC XXXX. 883 [RFC Editor: Upon publication as an RFC, please replace "XXXX" 884 with the number assigned to this document and remove this note.] 886 Applications which use this media type: 887 For example: Video over IP, video conferencing. 889 Fragment identifier considerations: 890 N/A. 892 Additional information: 893 None. 895 Person & email address to contact for further information: 896 Jonathan Lennox 898 Intended usage: 899 COMMON 901 Restrictions on usage: 902 This media type depends on RTP framing, and hence is only defined 903 for transfer via RTP [RFC3550]. 905 Author: 906 Jonathan Lennox 908 Change controller: 909 IETF AVTCore Working Group delegated from the IESG. 911 8. Security Considerations 913 RTP packets using the payload format defined in this specification 914 are subject to the security considerations discussed in the RTP 915 specification [RFC3550], and in any applicable RTP profile such as 916 RTP/AVP [RFC3551], RTP/AVPF [RFC4585], RTP/SAVP [RFC3711], or RTP/ 917 SAVPF [RFC5124]. SAVPF [RFC5124]. However, as "Securing the RTP 918 Protocol Framework: Why RTP Does Not Mandate a Single Media Security 919 Solution" [RFC7202] discusses, it is not an RTP payload format's 920 responsibility to discuss or mandate what solutions are used to meet 921 the basic security goals like confidentiality, integrity and source 922 authenticity for RTP in general. This responsibility lays on anyone 923 using RTP in an application. They can find guidance on available 924 security mechanisms in Options for Securing RTP Sessions [RFC7201]. 925 Applications SHOULD use one or more appropriate strong security 926 mechanisms. The rest of this security consideration section 927 discusses the security impacting properties of the payload format 928 itself. 930 This RTP payload format and its media decoder do not exhibit any 931 significant non-uniformity in the receiver-side computational 932 complexity for packet processing, and thus are unlikely to pose a 933 denial-of-service threat due to the receipt of pathological data. 934 Nor does the RTP payload format contain any active content. 936 9. Congestion Control 938 Congestion control for RTP SHALL be used in accordance with RFC 3550 939 [RFC3550], and with any applicable RTP profile; e.g., RFC 3551 940 [RFC3551]. The congestion control mechanism can, in a real-time 941 encoding scenario, adapt the transmission rate by instructing the 942 encoder to encode at a certain target rate. Media aware network 943 elements MAY use the information in the VP9 payload descriptor in 944 Section 4.2 to identify non-reference frames and discard them in 945 order to reduce network congestion. Note that discarding of non- 946 reference frames cannot be done if the stream is encrypted (because 947 the non-reference marker is encrypted). 949 10. IANA Considerations 951 The IANA is requested to register the media type registration "video/ 952 vp9" as specified in Section 7. The media type is also requested to 953 be added to the IANA registry for "RTP Payload Format MIME types" 954 . 956 11. Acknowledgments 958 Alex Eleftheriadis, Yuki Ito, Won Kap Jang, Sergio Garcia Murillo, 959 Roi Sasson, Timothy Terriberry, Emircan Uysaler, and Thomas Volkert 960 commented on the development of this document and provided helpful 961 comments and feedback. 963 12. References 965 12.1. Normative References 967 [I-D.ietf-avtext-lrr] 968 Lennox, J., Hong, D., Uberti, J., Holmer, S., and M. 969 Flodman, "The Layer Refresh Request (LRR) RTCP Feedback 970 Message", Work in Progress, Internet-Draft, draft-ietf- 971 avtext-lrr-07, 2 July 2017, 972 . 975 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 976 Requirement Levels", BCP 14, RFC 2119, 977 DOI 10.17487/RFC2119, March 1997, 978 . 980 [RFC3264] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model 981 with Session Description Protocol (SDP)", RFC 3264, 982 DOI 10.17487/RFC3264, June 2002, 983 . 985 [RFC3550] Schulzrinne, H., Casner, S., Frederick, R., and V. 986 Jacobson, "RTP: A Transport Protocol for Real-Time 987 Applications", STD 64, RFC 3550, DOI 10.17487/RFC3550, 988 July 2003, . 990 [RFC4585] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey, 991 "Extended RTP Profile for Real-time Transport Control 992 Protocol (RTCP)-Based Feedback (RTP/AVPF)", RFC 4585, 993 DOI 10.17487/RFC4585, July 2006, 994 . 996 [RFC4855] Casner, S., "Media Type Registration of RTP Payload 997 Formats", RFC 4855, DOI 10.17487/RFC4855, February 2007, 998 . 1000 [RFC5104] Wenger, S., Chandra, U., Westerlund, M., and B. Burman, 1001 "Codec Control Messages in the RTP Audio-Visual Profile 1002 with Feedback (AVPF)", RFC 5104, DOI 10.17487/RFC5104, 1003 February 2008, . 1005 [RFC6838] Freed, N., Klensin, J., and T. Hansen, "Media Type 1006 Specifications and Registration Procedures", BCP 13, 1007 RFC 6838, DOI 10.17487/RFC6838, January 2013, 1008 . 1010 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1011 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1012 May 2017, . 1014 [RFC8866] Begen, A., Kyzivat, P., Perkins, C., and M. Handley, "SDP: 1015 Session Description Protocol", RFC 8866, 1016 DOI 10.17487/RFC8866, January 2021, 1017 . 1019 [VP9-BITSTREAM] 1020 Grange, A., de Rivaz, P., and J. Hunt, "VP9 Bitstream & 1021 Decoding Process Specification", Version 0.6, 31 March 1022 2016, 1023 . 1027 12.2. Informative References 1029 [RFC3551] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and 1030 Video Conferences with Minimal Control", STD 65, RFC 3551, 1031 DOI 10.17487/RFC3551, July 2003, 1032 . 1034 [RFC3711] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K. 1035 Norrman, "The Secure Real-time Transport Protocol (SRTP)", 1036 RFC 3711, DOI 10.17487/RFC3711, March 2004, 1037 . 1039 [RFC5124] Ott, J. and E. Carrara, "Extended Secure RTP Profile for 1040 Real-time Transport Control Protocol (RTCP)-Based Feedback 1041 (RTP/SAVPF)", RFC 5124, DOI 10.17487/RFC5124, February 1042 2008, . 1044 [RFC6386] Bankoski, J., Koleszar, J., Quillio, L., Salonen, J., 1045 Wilkins, P., and Y. Xu, "VP8 Data Format and Decoding 1046 Guide", RFC 6386, DOI 10.17487/RFC6386, November 2011, 1047 . 1049 [RFC7201] Westerlund, M. and C. Perkins, "Options for Securing RTP 1050 Sessions", RFC 7201, DOI 10.17487/RFC7201, April 2014, 1051 . 1053 [RFC7202] Perkins, C. and M. Westerlund, "Securing the RTP 1054 Framework: Why RTP Does Not Mandate a Single Media 1055 Security Solution", RFC 7202, DOI 10.17487/RFC7202, April 1056 2014, . 1058 [RFC7667] Westerlund, M. and S. Wenger, "RTP Topologies", RFC 7667, 1059 DOI 10.17487/RFC7667, November 2015, 1060 . 1062 Authors' Addresses 1064 Justin Uberti 1065 Google, Inc. 1066 747 6th Street South 1067 Kirkland, WA 98033 1068 United States of America 1070 Email: justin@uberti.name 1072 Stefan Holmer 1073 Google, Inc. 1074 Kungsbron 2 1075 SE-111 22 Stockholm 1076 Sweden 1078 Email: holmer@google.com 1080 Magnus Flodman 1081 Google, Inc. 1082 Kungsbron 2 1083 SE-111 22 Stockholm 1084 Sweden 1086 Email: mflodman@google.com 1088 Danny Hong 1089 Google, Inc. 1090 1585 Charleston Road 1091 Mountain View, CA 94043 1092 United States of America 1094 Email: dannyhong@google.com 1095 Jonathan Lennox 1096 8x8, Inc. / Jitsi 1097 1350 Broadway 1098 New York, NY 10018 1099 United States of America 1101 Email: jonathan.lennox@8x8.com