| Payload Working Group | J. Uberti |
| Internet-Draft | S. Holmer |
| Intended status: Standards Track | M. Flodman |
| Expires: April 30, 2015 | |
| J. Lennox | |
| Vidyo | |
| October 27, 2014 |
RTP Payload Format for VP9 Video
draft-uberti-payload-vp9-00
This memo describes an RTP payload format for the VP9 video codec. The payload format has wide applicability, as it supports applications from low bit-rate peer-to-peer usage, to high bit-rate video conferences. It includes provisions for temporal and spatial scalability.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on April 30, 2015.
Copyright (c) 2014 IETF Trust and the persons identified as the document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
This memo describes an RTP payload specification applicable to the transmission of video streams encoded using the VP9 video codec [I-D.grange-vp9-bitstream]. The format described in this document can be used both in peer-to-peer and video conferencing applications.
TODO: VP9 description. Please see [I-D.grange-vp9-bitstream].
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].
The VP9 codec can maintain up to eight reference frames, of which up to three can be referenced or updated by any new frame.
VP9 also allows a reference frame to be resampled and used as a reference for another frame of a different resolution. This allows internal resolution changes without requiring the use of keyframes.
These features together enable an encoder to implement various forms of coarse-grained scalability, including temporal, spatial, and quality scalability modes, as well as combinations of these, without the need for explicit spatially scalabile encoding modes.
This payload format specification defines how such scalability modes can be encoded and communicated. In this payload, three separate types of layers are defined: temporal, spatial, and quality.
Temporal layers define different frame rates of video; spatial and quality layers define different, dependent representations of a single picture. Spatial layers allow a picture to be encoded at different resolutions, whereas quality layers allow a picture to be encoded at the same resolution but at different bitrates (and thus with different amounts of coding error).
Layers are designed (and MUST be encoded) such that if any layer, and all higher layers, are removed from the bitstream along any of the three dimensions, the remaining bitstream is still correctly decodable.
For terminology, this document uses the term "frame" to refer to a single encoded VP9 image, and "picture" to refer to all the representations of frames at a single instant in time. A picture thus can consist of multiple frames, encoding different spatial and/or quality layers.
[Editor's Note: Are separate spatial and quality layers necessary and useful? We could simplify by only defining a single sequence of frames within a picture.
Two modes of describing layer information are possible: "non-flexible mode" and "flexible mode". An encoder can freely switch between the two as appropriate.
In non-flexible mode, an SS message, which defines the layer hierarchy, is sent in the beginning of the stream together with the key frame. Each packet will have a picture id and reference indices, which in conjunction with the SS and the RTP sequence number can be used to determine if the packet is decodable or not. An SU message can be sent by the sending client, or an MCU, to notify the receiver about what subset of the SS it will actually be receiving.
In the flexible mode each packet contains 1-4 reference indices, which identifies all frames referenced by the frame transmitted in the current packet. This enables a receiver to identify if a frame is decodable or not and helps it understand the layer structure so that it can drop packets as it sees fit. Since this is signaled in each packet it makes it possible to have more flexible layer hierarchies and patterns which are changing dynamically.
This section describes how the encoded VP9 bitstream is encapsulated in RTP. To handle network losses usage of RTP/AVPF [RFC4585] is RECOMMENDED. All integer fields in the specifications are encoded as unsigned integers in network octet order.
The general RTP payload format for VP9 is depicted below.
0 1 2 3
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|V=2|P|X| CC |M| PT | sequence number |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| timestamp |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| synchronization source (SSRC) identifier |
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
| contributing source (CSRC) identifiers |
| .... |
+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+=+
| VP9 payload descriptor (integer #bytes) |
: :
| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| : VP9 pyld hdr | |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |
| |
+ |
: Bytes 2..N of VP9 payload :
| |
| +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| : OPTIONAL RTP padding |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
The VP9 payload descriptor and VP9 payload header will be described in the next section. OPTIONAL RTP padding MUST NOT be included unless the P bit is set.
Figure 1
The first octets after the RTP header are the VP9 payload descriptor, with the following structure.
0 1 2 3 4 5 6 7
+-+-+-+-+-+-+-+-+
|I|L|F|B|E|V|U|-| (REQUIRED)
+-+-+-+-+-+-+-+-+
I: |M|PICTURE ID | (RECOMMENDED)
+-+-+-+-+-+-+-+-+
M: | EXTENDED PID | (RECOMMENDED)
+-+-+-+-+-+-+-+-+
L: | T | S | Q | R | (CONDITIONALLY RECOMMENDED)
+-+-+-+-+-+-+-+-+ -\
F: | PID |X| RS| RQ| (OPTIONAL) .
+-+-+-+-+-+-+-+-+ . - R times
X: | EXTENDED PID | (OPTIONAL) .
+-+-+-+-+-+-+-+-+ -/
V: | SS |
| .. |
+-+-+-+-+-+-+-+-+
U: | SU |
| .. |
+-+-+-+-+-+-+-+-+
Figure 2
After the extension bit field follow the extension data fields that are enabled.
These 1-2 bytes are repeated R times, defined by the two R bits in the layer indices field.
The Scalability Structure data describes the pattern of scalable frames that will be used in a scalable stream. If the VP9 payload header's "V" bit is set, the scalability structure (SS) is present in the position indicated in Figure 2.
+-+-+-+-+-+-+-+-+
V: | PATTERN LENGTH|
+-+-+-+-+-+-+-+-+ -\
| T | S | Q | R | (OPTIONAL) .
+-+-+-+-+-+-+-+-+ -\ .
| PID |X| RS| RQ| (OPTIONAL) . . - PAT. LEN. times
+-+-+-+-+-+-+-+-+ . - R times .
X: | EXTENDED PID | (OPTIONAL) . .
+-+-+-+-+-+-+-+-+ -/ -/
Figure 3
The scalability structure allows the structure of the VP9 stream to be predeclared, rather than indicating it on the fly with every frame as with the layer indices.
Its structure consists of a sequence of frames, encoded as with the layer indices. It begins with PATTERN LENGTH, indicating the number of frames in the pattern; it is then followed by that many instances of data encoded using the same semantics as the layer indices.
TODO: add frame resolution information.
In a scalable stream sent with a fixed pattern, the scalability structure SHOULD be included in the first packet of every keyframe picture, and also in the first packet of the first picture in which the scalability structure changes. If a SS is included in a picture with TID not equal to 0, it MUST also be repeated in the first packet the first frame with a lower TID, until TID equals 0.
If PATTERN LENGTH is 0, it indicates that no fixed scalability information is present going forward in the bitstream. An SS with a PATTERN LENGTH of 0 allows a bitstream to be changed from non-flexible to flexible mode.
TODO
TODO: need to describe VP9 payload header.
VP9 frames are fragmented into packets, in RTP sequence number order, beginning with a packet with the B bit set, and ending with a packet with the RTP marker bit set. There is no mechanism for finer-grained access to parts of a VP9 frame.
TODO
The VP9 payload descriptor defined in Section 4.2 above contains an optional PictureID parameter. One use of this parameter is included to enable use of reference picture selection index (RPSI) and slice loss indication (SLI), both defined in [RFC4585].
TODO: Update to indicate which frame within the picture.
The reference picture selection index is a payload-specific feedback message defined within the RTCP-based feedback format. The RPSI message is generated by a receiver and can be used in two ways. Either it can signal a preferred reference picture when a loss has been detected by the decoder -- preferably then a reference that the decoder knows is perfect -- or, it can be used as positive feedback information to acknowledge correct decoding of certain reference pictures. The positive feedback method is useful for VP9 used as unicast. The use of RPSI for VP9 is preferably combined with a special update pattern of the codec's two special reference frames -- the golden frame and the altref frame -- in which they are updated in an alternating leapfrog fashion. When a receiver has received and correctly decoded a golden or altref frame, and that frame had a PictureID in the payload descriptor, the receiver can acknowledge this simply by sending an RPSI message back to the sender. The message body (i.e., the "native RPSI bit string" in [RFC4585]) is simply the PictureID of the received frame.
TODO: Update to indicate which frame within the picture.
The slice loss indication is another payload-specific feedback message defined within the RTCP-based feedback format. The SLI message is generated by the receiver when a loss or corruption is detected in a frame. The format of the SLI message is as follows [RFC4585]:
0 1 2 3
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
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
| First | Number | PictureID |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
Figure 4
Here, First is the macroblock address (in scan order) of the first lost block and Number is the number of lost blocks. PictureID is the six least significant bits of the codec-specific picture identifier in which the loss or corruption has occurred. For VP9, this codec-specific identifier is naturally the PictureID of the current frame, as read from the payload descriptor. If the payload descriptor of the current frame does not have a PictureID, the receiver MAY send the last received PictureID+1 in the SLI message. The receiver MAY set the First parameter to 0, and the Number parameter to the total number of macroblocks per frame, even though only parts of the frame is corrupted. When the sender receives an SLI message, it can make use of the knowledge from the latest received RPSI message. Knowing that the last golden or altref frame was successfully received, it can encode the next frame with reference to that established reference.
TODO: this example is copied from the VP8 payload format specification, and has not been updated for VP9. It may be incorrect.
The use of RPSI and SLI is best illustrated in an example. In this example, the encoder may not update the altref frame until the last sent golden frame has been acknowledged with an RPSI message. If an update is not received within some time, a new golden frame update is sent instead. Once the new golden frame is established and acknowledged, the same rule applies when updating the altref frame.
| Event | Sender | Receiver | Established reference |
|---|---|---|---|
| 1000 | Send golden frame PictureID = 0 | ||
| Receive and decode golden frame | |||
| 1001 | Send RPSI(0) | ||
| 1002 | Receive RPSI(0) | golden | |
| ... | (sending regular frames) | ||
| 1100 | Send altref frame PictureID = 100 | ||
| Altref corrupted or lost | golden | ||
| 1101 | Send SLI(100) | golden | |
| 1102 | Receive SLI(100) | ||
| 1103 | Send frame with reference to golden | ||
| Receive and decode frame (decoder state restored) | golden | ||
| ... | (sending regular frames) | ||
| 1200 | Send altref frame PictureID = 200 | ||
| Receive and decode altref frame | golden | ||
| 1201 | Send RPSI(200) | ||
| 1202 | Receive RPSI(200) | altref | |
| ... | (sending regular frames) | ||
| 1300 | Send golden frame PictureID = 300 | ||
| Receive and decode golden frame | altref | ||
| 1301 | Send RPSI(300) | altref | |
| 1302 | RPSI lost | ||
| 1400 | Send golden frame PictureID = 400 | ||
| Receive and decode golden frame | altref | ||
| 1401 | Send RPSI(400) | ||
| 1402 | Receive RPSI(400) | golden |
Note that the scheme is robust to loss of the feedback messages. If the RPSI is lost, the sender will try to update the golden (or altref) again after a while, without releasing the established reference. Also, if an SLI is lost, the receiver can keep sending SLI messages at any interval allowed by the RTCP sending timing restrictions as specified in [RFC4585], as long as the picture is corrupted.
Editor's Note: The message described in this section is applicable to other codecs beyond just VP9. In the future it will be likely be split out into another document.
TODO: details of how this is encoded in RTCP.
A synchronization frame can be requested by sending a LIR, which is an RTCP feedback message asking the encoder to encode a frame which makes it possible to upgrade to a higher layer. The LIR message contains two tuples, {T1,S1,Q1} and {T2,S2,Q2}, where the first tuple is the currently highest layer the decoder can decode, while the second tuple is the layer the decoder wants to upgrade to.
Identification of an upgrade frame can be derived from the reference IDs of each frame by backtracking the dependency chain until reaching a point where only decodable frames are being referenced. Therefore it's recommended both for both the flexible and the non-flexible mode that, when upgrade frames are being encoded in response to a LIR, those packets should contain layer indices and the reference fields so that the decoder or an MCU can make this derivation.
Example:
LIR {1,1,0}, {1,2,1} is sent by an MCU when it is currently relaying {1,1,0} to a receiver and which wants to upgrade to {1,2,1}. In response the encoder should encode the next frames in layers {1,1,1} and {1,2,1} by only referring to frames in {1,1,0}, {1,0,0} or {0,0,0}.
In the non-flexible mode, periodic upgrade frames can be defined by the layer structure of the SS, thus periodic upgrade frames can be automatically identified by the picture ID.
This payload format has two required parameters.
This registration is done using the template defined in [RFC6838] and following [RFC4855].
The receiver MUST ignore any fmtp parameter unspecified in this memo.
The media type video/VP9 string is mapped to fields in the Session Description Protocol (SDP) [RFC4566] as follows:
An example of media representation in SDP is as follows:
m=video 49170 RTP/AVPF 98
a=rtpmap:98 VP9/90000
a=fmtp:98 max-fr=30; max-fs=3600;
TODO: Update this for VP9
RTP packets using the payload format defined in this specification are subject to the security considerations discussed in the RTP specification [RFC3550], and in any applicable RTP profile. The main security considerations for the RTP packet carrying the RTP payload format defined within this memo are confidentiality, integrity and source authenticity. Confidentiality is achieved by encryption of the RTP payload. Integrity of the RTP packets through suitable cryptographic integrity protection mechanism. Cryptographic system may also allow the authentication of the source of the payload. A suitable security mechanism for this RTP payload format should provide confidentiality, integrity protection and at least source authentication capable of determining if an RTP packet is from a member of the RTP session or not. Note that the appropriate mechanism to provide security to RTP and payloads following this memo may vary. It is dependent on the application, the transport, and the signaling protocol employed. Therefore a single mechanism is not sufficient, although if suitable the usage of SRTP [RFC3711] is recommended. This RTP payload format and its media decoder do not exhibit any significant non-uniformity in the receiver-side computational complexity for packet processing, and thus are unlikely to pose a denial-of-service threat due to the receipt of pathological data. Nor does the RTP payload format contain any active content.
Congestion control for RTP SHALL be used in accordance with RFC 3550 [RFC3550], and with any applicable RTP profile; e.g., RFC 3551 [RFC3551]. The congestion control mechanism can, in a real-time encoding scenario, adapt the transmission rate by instructing the encoder to encode at a certain target rate. Media aware network elements MAY use the information in the VP9 payload descriptor in Section 4.2 to identify non-reference frames and discard them in order to reduce network congestion. Note that discarding of non-reference frames cannot be done if the stream is encrypted (because the non-reference marker is encrypted).
The IANA is requested to register the following values:
- Media type registration as described in Section 7.1.