| Internet-Draft | media container | November 2024 |
| Zanaty, et al. | Expires 8 May 2025 | [Page] |
This specification describes a media container format for encoded and encrypted audio and video media data to be used primarily for interactive Media over QUIC Transport (MOQT) [MoQTransport], with the goal of it being a low-overhead format. It further defines the LOC Streaming Format for the MOQ Common Catalog format [MoQCatalog] for publishers to annouce and describe their LOC tracks and for subscribers to consume them. The specification also provides examples to aid application developers for building media applications over MOQT and intending to use LOC as the streaming format.¶
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 https://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 8 May 2025.¶
Copyright (c) 2024 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 (https://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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
This specification describes a low-overhead media container format for encoded and encrypted audio and video media data, as well as a MOQ Common Catalog streaming format called LOC to describe such tracks.¶
"Low-overhead" refers to minimal extra encapsulation as well as minimal application overhead when interfacing with WebCodecs [WebCodecs].¶
The container format description is specified for all audio and video codecs defined in the WebCodecs Codec Registry [WEBCODECS-CODEC-REGISTRY]. The audio and video payload bitstream is identical to the "internal data" inside an EncodedAudioChunk and EncodedVideoChunk, respectively, specified in the registry.¶
(Note: Do we need to support timed text tracks such as Web Video Text Tracks (WebVTT) ?)¶
In addition to the media payloads, critical metadata is also specified for audio and video payloads. (Note: Align with MOQT terminology of either "metadata" or "header".)¶
A primary motivation is to align with media formats used in WebCodecs to minimize extra encapsulation and application overhead when interfacing with WebCodecs. Other container formats like CMAF or RTP would require more extensive application overhead in format conversions, as well as larger encapsultion overhead which may burden some use cases like low bitrate audio scenarios.¶
This specification can also be used by applications outside the context of WebCodecs or a web browser. While the media payloads are defined by referring to the "internal data" of an EncodedAudioChunk or EncodedVideoChunk in the WebCodecs Codec Registry, this "internal data" is the elementary bitstream format of codecs without any encapsulation. Referring to the WebCodecs Codec Registry avoids duplicating it in an identical IANA registry.¶
Section 2.2 defines the metadata associated with audio and video payloads.¶
Section 3 describes the LOC Streaming Format bindings to the MoQ Common Catalog format including examples.¶
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD","SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119].¶
Track, Group, Subgroup, Object, and their corresponding identifiers (ID or alias) are defined in [MoQTransport] and used here to refer to those aspects of the MOQT Object Model.¶
The WebCodecs Codec Registry defines the contents of an EncodedAudioChunk and EncodedVideoChunk for the audio and video codec formats in the registry. The "internal data" in these chunks is used directly in this specification as the "LOC Payload" bitstream. This "internal data" is the elementary bitstream format of each codec without any encapsulation.¶
For video formats with multiple bitstream formats in the WebCodecs Registry, such as H.264/AVC or H.265/HEVC, the LOC Payload uses the "canonical" format ("avc" or "hevc", not "annexB") with the following additions:¶
Parameter sets can be sent in the bitstream before key frames, similar to "annexB" formats. (Note that newer "canonical" formats such as "avc3" and "hev1" codec strings support parameter sets in the bitstream or outside it.)¶
Parameter sets can be provided by means outside this specification, such as "extradata" in "canonical" ("avc" or "hevc") formats.¶
4 byte length codes can be sent before each NAL Unit, similar to "canonical" ("avc" or "hevc") formats. (Note that a length of 1 should be interpreted as a start code rather than a length.)¶
4 byte or longer start codes can be sent before each NAL Unit, similar to "annexB" formats.¶
Note that if length codes or start codes are less than 4 bytes, which is uncommon, it may not be possible to disambiguate them without information outside this specification.¶
An application object when transported as a [MoQTransport] object is composed of a MOQ Object Header, with optional Extensions, and a Payload. Media objects encoded using the container format defined in this specification populate the MOQ Object Payload with the LOC Payload, and the MOQ Object Header Extensions with the LOC Header Extensions, as shown below.¶
The LOC Payload is the "internal data" of an EncodedAudioChunk or EncodedVideoChunk.¶
The LOC Header Extensions carry optional metadata related to the Payload.¶
<----------- MOQ Object ------------>
+----------+--------------+-----------+
| MOQ | MOQ Header | MOQ |
| Header | Extensions | Payload |
+----------+--------------+-----------+
| |
| |
+--------------+-----------+
| LOC Header | LOC |
| Extensions | Payload |
+--------------+-----------+
LOC Header Extensions = some MOQ Object Header Extensions
LOC Payload = all MOQ Object Payload
LOC Payload = "internal data" of EncodedAudio/VideoChunk
¶
The LOC Header Extensions carry optional metadata for the corresponding LOC Payload. The LOC Header Extensions are contained within the MOQ Object Header Extensions. This metadata provides necessary information for end subscribers, relays and other intermediaries to perform their operations without accessing the media payload. For example, media switches can use this metadata to perform their media switching decisions without accessing the payload which may be encrypted end-to-end (from original publisher to end subscribers).¶
The following sections define specific metadata as LOC Header Extensions and register them in the IANA registry for MOQ Object Header Extensions.¶
Other specifications can define other metadata as LOC Header Extensions and register them in the same registry. Each extension must specify the following information in the IANA registry.¶
Name: Short name for the metadata. (Not sent on the wire.)¶
Description: Detailed description for the metadata. (Not sent on the wire.)¶
ID: Identifier assigned by the registry. (varint)¶
Length: Length of metadata Value in bytes. (varint)¶
Value: Value of metadata. (Length bytes)¶
Name: Capture Timestamp¶
Description: Wall-clock time in microseconds since the Unix epoch when the encoded media frame was captured, encoded as a 64 bit unsigned integer in network byte order (big endian).¶
ID: TBA (IANA, please assign from the MOQ Header Extensions Registry)¶
Length: 8 (bytes)¶
Value: Varies¶
Name: Video Frame Marking¶
Description: Flags for video frames which are independent, discardable, or base layer sync points, as well as temporal and spatial layer identification, as defined in [Framemarking].¶
ID: TBA (IANA, please assign from the MOQ Header Extensions Registry)¶
Length: Varies (1-3 bytes)¶
Value: Varies¶
A catalog track provides information about tracks from a given publisher. A catalog is used by subscribers for consuming tracks and by publishers to advertise and describe the tracks. The content of a catalog is opaque to the relays and may be end to end encrypted. A catalog describes the details of tracks such as Track IDs and corresponding media configuration details, for example, audio/video codec details.¶
The LOC Streaming Format uses the MoQ Common Catalog Format [MoQCatalog] to describe the content being produced by a publisher.¶
Per Sect 5.1 of [MoQCatalog], this document registers an entry in the "MoQ Streaming Format Type" table, with the type value 2, the name "LOC Streaming Format", and the RFC XXX.¶
Every LOC catalog track MUST declare a streaming format type (See Sect 3.2.1 of [MoQCatalog]) value of 2.¶
Every LOC catalog track MUST declare a streaming format version (See Sect 3.2.1 of [MoQCatalog]) value of 1, which is the version described in this document.¶
Every LOC catalog track MUST declare a packaging type (See Sect 3.2.9 of [MoQCatalog]) of "loc".¶
The catalog track MUST have a track name of "catalog". A catalog object MAY be independent of other catalog objects or it MAY represent a delta update of a prior catalog object. The first catalog object published within a new group MUST be independent. A catalog object SHOULD only be published only when the availability of tracks changes.¶
Each catalog update MUST be mapped to a discreet moq-transport object.¶
The MOQ Common Catalog defines the required base fields and optional extensions.¶
The LOC Streaming Format allows the following optional extensions for video media.¶
temporalId: Identifies the temporal layer/sub-layer encoded, starting with 0 for the base layer, and increasing with higher temporal fidelity.¶
spatialId: Identifies the spatial and quality layer encoded, starting with 0 for the base layer, and increasing with higher fidelity.¶
depends: Identifies track dependencies for a given track, usually for video media with scalable layers in separate tracks.¶
renderGroup: Identifies a group of time-aligned tracks which should be rendered simultaneously.¶
selectionParams: Selection parameters for media quality, fidelity, etc.; see next section.¶
Each video track can have the following associated Selection Parameters.¶
codec: Codec information (including profile, level, tier, etc.), as defined by the codec registrations listed in [WEBCODECS-CODEC-REGISTRY].¶
framerate: As defined in section 7.8 of [WEBCODECS-CODEC-REGISTRY].¶
bitrate: As defined in section 7.7 and 7.8 of [WEBCODECS-CODEC-REGISTRY].¶
width, height: As defined in section 7.8 of [WEBCODECS-CODEC-REGISTRY].¶
displayWidth, displayheight: As defined in section 7.7 of [WEBCODECS-CODEC-REGISTRY].¶
The LOC Streaming Format allows the following optional extensions for audio media.¶
Each audio track can have the following associated Selection Parameters.¶
codec: Codec information as defined by the codec registrations listed in [WEBCODECS-CODEC-REGISTRY].¶
bitrate: As defined in section 7.7 and 7.8 of [WEBCODECS-CODEC-REGISTRY].¶
samplerate: As defined in section 7.7 of [WEBCODECS-CODEC-REGISTRY].¶
chanelConfig: As defined in section 7.7 of [WEBCODECS-CODEC-REGISTRY].¶
lang: The primary language of the track, using standard tags from [RFC5646].¶
See section 3.4 of the MOQ Common Catalog [MoQCatalog].¶
When end to end encryption is supported, the encoded payload is encrypted with symmetric keys derived from key establishment mechanisms, such as [MOQ-MLS], and the payload itself is protected using mechanisms defined in [SecureObjects].¶
This section provides examples with details for building audio and video applications using MOQ and LOC; more specifically, it provides information on:¶
Using a catalog to describe track information,¶
Packaging media into LOC streaming format, and¶
Mapping application media objects to the MOQT object model and transport.¶
The figure below shows the conceptual model for mapping media application data to the MOQT object model and underlying QUIC transport.¶
+------------------------------+
| Media Application |
| Audio, Video Frames |
+---------------+--------------+
|
|
+---------------v--------------------+
| MOQT Object Model |
| Tracks, Groups, Subgroups, Objects |
+---------------+--------------------+
|
|
+---------------v--------------+
| QUIC |
| Streams, Datagrams |
+------------------------------+
¶
An example is shown below for an Opus mono channel audio track at 48Khz.¶
codec: "opus" bitrate: 24000 samplerate: 480000 channelConfig: "mono" lang: "en"¶
When ready for publishing, each encoded audio chunk, say 10ms, represents a
MOQT Object. In this setup, there is one MOQT Object
per MOQT Group, where the GroupID in the object header is
increment by one for each encoded audio chunk and the ObjectID
is defaulted to value 0.¶
These objects can be sent as QUIC streams or datagrams. When mapped to
QUIC datagrams, each object must fit entirely within a QUIC datagram, and
when mapped to QUIC Streams, each such unitary group is sent over
an individual unidirectional QUIC stream since there is just one SubGroup per
each MOQT Group.¶
An example is shown below for an H.264 video track with 1280x720p resolution and 30 fps frame rate at 1 Mbps bitrate.¶
codec: "avc3.42E01E" bitrate: 1000000 framerate: 30 width: 1280 height: 720¶
When ready for publishing, each encoded video chunk is considered as input
to MOQT Object payload. If encrypted, the output of encryption will serve as
the object's payload. The GroupID is incremented by 1 at IDR Frame boundaries.
The ObjectID is increment by 1 for each encoded video frame, starting at 0
and resetting to 0 at the start of a new group. The first encoded video frame,
MOQT Object with ObjectID 0, shall be the Independent (IDR) frame and
the rest of the encoded video frames corresponds to dependent (delta) frames,
organized in the decode order.¶
When mapping to QUIC for sending, one unidirectional QUIC stream is setup to deliver all the encoded video chunks within a MOQT group.¶
When decoding at the 'End Consumer', the objects from each of the QUIC streams are fed in the GroupID then ObjectID order to the decoder for the track.¶
An example is shown below for an H.264 video track with 1280x720p resolution and 2 temporal layers at 30 fps and 60 fps frame rate.¶
codec: "avc3.42E01F" bitrate: 1500000 framerate: 60 width: 1280 height: 720¶
When ready for publishing, each encoded video chunk is considered as input
to MOQT Object payload. If encrypted, the output of encryption will serve as
the object's payload. The GroupID is incremented by 1 at Independent (IDR)
frame boundaries. Each MOQT group shall contain 2 SubGroups corresponding
to the 2 temporal layers as shown below:¶
Layer:0/30fps Subgroup: 0 ObjectID: even Layer:1/60fps Subgroup: 1 ObjectID: odd¶
Within the MOQT group, ObjectID is increment by 1 for each encoded video
frame, starting at 0 and resetting to 0 at the start of a new group. The
first encoded video frame, MOQT Object with ObjectID 0, shall be the
Indepedent (IDR) frame and the rest of the encoded video frames corresponds to
dependent (delta) frames, organized in the decode order. When mapping to
QUIC for sending, one unidirectional QUIC stream is used per SubGroup,
thus resulting in 2 QUIC streams per MOQT group.¶
When decoding at the 'End Consumer' for a given MOQT group, the objects must be fed in the GroupID then ObjectID order. This implies that the consumer media application needs to order objects across the SubGroup QUIC streams.¶
An example is shown below for an H.264 video track with 2 spatial qualities at 360p and 720p each at 30 fps¶
Video Track 1 codec: "avc3.42E01E" bitrate: 500000 framerate: 30 width: 640 height: 360 Video Track 2 codec: "svc1.56401F" bitrate: 1000000 framerate: 30 width: 1280 height: 720¶
When ready for publishing, the mapping to the MOQT object model and to underlying QUIC, follows the same procedures as described in Section 5.2 for each video track.¶
When decoding at the 'End Consumer' for a given MOQT group, the objects must be fed in the GroupID then ObjectID order in the ascending quality track order.¶
For the example in the section, this would imply following pattern when decoding group 5.¶
Track 1 Group 5 Object 0 Track 2 Group 5 Object 0 Track 1 Group 5 Object 1 Track 2 Group 5 Object 1 ....¶
An example is shown below for an H.264 video track with 2 spatial qualities at 360p and 720p, however, the framerate between tracks vary dyadically.¶
Video Track 1 codec: "avc3.42E01E" bitrate: 500000 framerate: 30 width: 640 height: 360 Video Track 2 codec: "svc1.56E01F" bitrate: 1000000 framerate: 60 width: 1280 height: 720¶
When ready for publishing, the mapping to the MOQT object model and to underlying QUIC, follows the same procedures as described in Section 5.2 for each video track.¶
When decoding at the 'End Consumer' for a given MOQT group, the objects from across the tracks must be fed in the timestamp order to the decoder, if no frame reordering is present in the encoding.¶
If the encoding uses frame reordering, or if timestamp cannot be obtained, the object to choose next shall follow the below formula.¶
Object Decode Order = ObjectID * multiplier + offset multiplier = 2^(maxlayer-max(0,layer-1)) offset = 2^(maxlayer-layer) MOD multiplier¶
An example is shown below for an H.264 video track with 2 simulcast spatial qualities at 360p and 720p each at 30 fps.¶
Video Track 1 codec: "avc3.42E01E" bitrate: 500000 framerate: 30 width: 640 height: 360 Video Track 2 codec: "avc3.42E01F" bitrate: 1000000 framerate: 30 width: 1280 height: 720¶
When ready for publishing, the mapping to the MOQT object model and to underlying QUIC, follows the same procedures as described in Section 5.2 for each video track.¶
When decoding at the 'End Consumer', the objects from the QUIC stream are fed in the GroupID then ObjectID order to the decoders setup for the corresponding video tracks.¶
The metadata in LOC Header Extensions is visible to relays, since the MOQ Object Header Extensions are often not encrypted end-to-end (from original publisher to end subscribers) in common schemes. In some cases, this may be an intentional design intent for proper relay operation. In other cases, this may be unintentional or undesirable leaking of the metadata to relays. Each metadata that is defined should consider the security and privacy aspects of granting relays visibility to the metadata. End-to-end encyption schemes should support end-to-end encryption of sensitive metadata.¶
The metadata defined and registered in this specification (Capture Timestamp, Video Frame Marking, and Audio Level) may be sensitive metadata that should be encrypted end-to-end. They are used by media switches, which are not merely relays, and likely have access to some media keys. This may require end-to-end encryption schemes with multiple different security key contexts for payload versus metadata.¶
The IANA registry for MOQ Object Header Extensions is populated with the entries specified in section Section 2.2, referencing this specification.¶
This document creates a new entry in the "MoQ Streaming Format" Registry (see [MoQTransport] Sect 8). The type value is 0x002, the name is "LOC Streaming Format" and the RFC is XXX.¶
Thanks to Cullen Jennings for suggestions and review.¶