RTP Payload Format for ISO/IEC 21122 (JPEG XS)

This document specifies a payload format for packetization of JPEG XS [ISO21122-1] encoded video signals into the Real-time Transport Protocol (RTP). The JPEG XS coding system offers compression and recompression of image sequences with very moderate computational resources while remaining robust under multiple compression and decompression cycles and mixing of content sources, e.g. embedding of subtitles, overlays or logos. Typical target compression ratios ensuring visually lossless quality are in the range of 2:1 to 10:1, depending on the nature of the source material. The end-to-end latency can be confined to a fraction of a frame, typically between a small number of lines down to below a single line.

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 RFC 2119. The unit of source data provided as payload to the transport layer, and corresponding, in this RTP payload definition, to a single JPEG XS frame. A ISO colour specification box defined in ISO/IEC 21122-3 that includes colour-related metadata required to correctly display JPEG XS frames, such as colour primaries, transfer characteristics and matrix coefficients. A marker that consists of the two bytes 0xff11 indicating the end of a JPEG XS codestream. A sequence of bytes representing a compressed image formatted according to JPEG XS Part-1. A sequence of bytes, starting with a SOC marker, at the beginning of each JPEG XS codestream encoded in multiple markers and marker segments that does not carry entropy coded data, but metadata such as the frame dimension and component precision. A JPEG XS picture segment in the case of a progressive frame, or, in the case of an interlaced frame, the concatenation of two JPEG XS picture segments. The concatenation of a video support box, as defined in ISO/IEC 21122-3, a colour specification box, as defined in ISO/IEC 21122-3 as well and a JPEG XS codestream header. The concatenation of a video support box, as defined in ISO/IEC 21122-3, a colour specification box, as defined in ISO/IEC 21122-3 as well and a JPEG XS codestream. A sequence of JPEG XS frames. A two-byte functional sequence that is part of a JPEG XS codestream starting with a 0xff byte and a subsequent byte defining its function. A marker along with a 16-bit marker size and payload data following the size. A portion of an Application Data Unit whose boundaries coincide with boundaries of RTP packet payloads (excluding payload header), i.e. the first (resp. last) byte of a packetization unit is the first (resp. last) byte of a RTP packet payload (excluding its payload header). The smallest independently decodable unit of a JPEG XS codestream, bearing in mind that it decodes to wavelet coefficients which still require inverse wavelet filtering to give an image. A marker that consists of the two bytes 0xff10 indicating the start of a JPEG XS codestream. A ISO video support box defined in ISO/IEC 21122-3 that includes metadata required to play back a JPEG XS stream, such as its maximum bitrate, its subsampling structure, its buffer model and its frame rate.

JPEG XS is a low-latency lightweight image coding system for coding continuous-tone grayscale or continuous-tone colour digital images. This coding system provides an efficient representation of image signals through the mathematical tool of wavelet analysis. The wavelet filter process separates each component into multiple bands, where each band consists of multiple coefficients describing the image signal of a given component within a frequency domain specific to the wavelet filter type, i.e. the particular filter corresponding to the band. Wavelet coefficients are grouped into precincts, where each precinct includes all coefficients over all bands that contribute to a spatial region of the image. One or multiple precincts are furthermore combined into slices consisting of an integer number of precincts. Precincts do not cross slice boundaries, and wavelet coefficients in precincts that are part of different slices can be decoded independently from each other. Note, however, that the wavelet transformation runs across slice boundaries. A slice always extends over the full width of the image, but may only cover parts of its height.

A JPEG XS codestream header, followed by several slices, and terminated by an EOC marker form a JPEG XS codestream. The overall codestream format, including the definition of all markers, is further defined in ISO/IEC 21122-1. It represents sample values of a single image, bare any interpretation relative to a colour space.

While the information defined in the codestream is sufficient to reconstruct the sample values of one image, the interpretation of the samples remains undefined by the codestream itself. This interpretation is given by the video support box and the colour specification box which contain significant information to correctly play the JPEG XS stream. The layout and syntax of these boxes, together with their content, are defined in ISO/IEC 21122-3. The video support box provides information on the maximum bitrate, the frame rate, the frame mode (progressive or interlaced), the subsampling image format, the timecode of the current JPEG XS frame, the profile, level and sublevel used (as defined in ISO/IEC 21122-2), and optionally on the buffer model and the mastering display metadata. The colour specification box indicates the colour primaries, transfer characteristics, matrix coefficients and video full range flag needed to specify the colour space of the video stream.

The concatenation of a video support box, a colour specification box, and a JPEG XS codestream forms a JPEG XS picture segment. In the case of a progressive video stream, each JPEG XS frame consists of one single JPEG XS picture segment. In the case of an interlaced video stream, each JPEG XS frame is made of two concatenated JPEG XS picture segments. The codestream of each picture segment corresponds exclusively to one of the two fields of the interlaced frame. Both picture segments SHALL contain identical boxes (i.e. concatenation of the video support box and the colour specification box is byte exact the same for both picture segments of the frame). Note that the interlaced mode as signaled by the frat field in the video support box indicates either progressive, interlaced top-field first, or interlaced bottom-field first mode. Thus, its value too SHALL be identical in both picture segments.

This section specifies the payload format for JPEG XS streams over the Real-time Transport Protocol (RTP). In order to be transported over RTP, each JPEG XS stream is transported in a distinct RTP stream, identified by a distinct SSRC. A JPEG XS stream is divided into Application Data Units (ADUs), each ADU corresponding to a single JPEG XS frame.

An ADU is made of several packetization units. If a packetization unit is bigger than the maximum size of a RTP packet payload, the unit is split into multiple RTP packet payloads, as illustrated in . As seen there, each packet SHALL contain (part of) one and only one packetization unit. A packetization unit may extend over multiple packets. The payload of every packet SHALL have the same size (based e.g. on the Maximum Transfer Unit of the network), except (possibly) the last packet of a packetization unit. The boundaries of a packetization unit SHALL coincide with the boundaries of the payload of a packet (excluding the payload header), i.e. the first (resp. last) byte of the packetization unit SHALL be the first (resp. last) byte of the payload (excluding its header).

There are two different packetization modes defined for this RTP payload format. Codestream packetization mode: in this mode, the packetization unit SHALL be the entire JPEG XS picture segment (i.e. codestream preceeded by boxes). This means that a progressive frame will have a single packetization unit, while an interlaced frame will have two. The progressive case is illustrated in . Slice packetization mode: in this mode, the packetization unit SHALL be the slice, i.e. there SHALL be data from no more than one slice per RTP packet. The first packetization unit SHALL be made of the JPEG XS header segment (i.e. the concatenation of the VS box, the CS box and the JPEG XS codestream header). This first unit is then followed by successive units, each containing one and only one slice. The packetization unit containing the last slice of a JPEG XS codestream SHALL also contain the EOC marker immediately following this last slice. This is illustrated in . In the case of an interlaced frame, the JPEG XS header segment of the second field SHALL be in its own packetization unit.

Due to the constant bit-rate of JPEG XS, the codestream packetization mode guarantees that a JPEG XS RTP stream will produce a constant number of bytes per frame, and a constant number of RTP packets per frame. To reach the same guarantee with the slice packetization mode, an additional mechanism is required. This can involve a constraint at the rate allocation stage in the JPEG XS encoder to impose a constant bit-rate at the slice level, the usage of padding data, or the insertion of empty RTP packets (i.e. a RTP packet whose payload data is empty).

The format of the RTP header is specified in RFC 3550 and reprinted in for convenience. This RTP payload format uses the fields of the header in a manner consistent with that specification. The RTP payload (and the settings for some RTP header bits) for packetization units are specified in .

The version (V), padding (P), extension (X), CSRC count (CC), sequence number, synchronization source (SSRC) and contributing source (CSRC) fields follow their respective definitions in RFC 3550. The remaining RTP header information to be set according to this RTP payload format is set as follows: If progressive scan video is being transmitted, the marker bit denotes the end of a video frame. If interlaced video is being transmitted, it denotes the end of the field. The marker bit SHALL be set to 1 for the last packet of the video frame/field. It SHALL be set to 0 for all other packets. A dynamically allocated payload type field that designates the payload as JPEG XS video. The RTP timestamp is set to the sampling timestamp of the content. A 90 kHz clock rate SHALL be used. As per specified in RFC 3550 and RFC 4175, the RTP timestamp designates the sampling instant of the first octet of the frame to which the RTP packet belongs. Packets SHALL not include data from multiple frames, and all packets belonging to the same frame SHALL have the same timestamp. Several successive RTP packets will consequently have equal timestamps if they belong to the same frame (that is until the marker bit is set to 1, marking the last packet of the frame), and the timestamp is only increased when a new frame begins. If the sampling instant does not correspond to an integer value of the clock, the value SHALL be truncated to the next lowest integer, with no ambiguity.

The first four bytes of the payload of an RTP packet in this RTP payload format are referred to as the payload header. illustrates the structure of this payload header.

The payload header consists of the following fields: The T bit is set to indicate that packets are sent sequentially by the transmitter. This information allows a receiver to dimension its input buffer(s) accordingly. If T=0, nothing can be assumed about the transmission order and packets may be sent out-of-order by the transmitter. If T=1, packets SHALL be sent sequentially by the transmitter. The K bit is set to indicate which packetization mode is used. K=0 indicates codestream packetization mode, while K=1 indicates slice packetization mode. In the case that the Transmission mode (T) is set to 0, the slice packetization mode SHALL be used and K SHALL be set to 1. The L bit is set to indicate the last packet of a packetization unit. As the end of the frame also ends the packet containing the last unit of the frame, the L bit is set whenever the M bit is set. If codestream packetization mode is used, L bit and M bit are equivalent. These 2 bits are used to indicate how the JPEG XS frame is scanned (progressive or interlaced). In case of an interlaced frame, they also indicate which JPEG XS picture segment the payload is part of (first or second). The payload is progressively scanned. Reserved for future use. The payload is part of the first JPEG XS picture segment of an interlaced video frame. The height specified in the included JPEG XS codestream header is half of the height of the entire displayed image. The payload is part of the second JPEG XS picture segment of an interlaced video frame. The height specified in the included JPEG XS codestream header is half of the height of the entire displayed image. The frame (F) counter identifies the frame number modulo 32 to which a packet belongs. Frame numbers are incremented by 1 for each frame transmitted. The frame number, in addition to the timestamp, may help the decoder manage its input buffer and bring packets back into their natural order. The Slice and Extended Packet (SEP) counter is used differently depending on the packetization mode. In the case of codestream packetization mode (K=0), this counter resets whenever the Packet counter resets (see hereunder), and increments by 1 whenever the Packet counter overruns. In the case of slice packetization mode (K=1), this counter identifies the slice modulo 2047 to which the packet contributes. If the data belongs to the JPEG XS header segment, this field SHALL have its maximal value, namely 2047 = 0x07ff. Otherwise, it is the slice index modulo 2047. Slice indices are counted from 0 (corresponding to the top of the frame). The packet (P) counter identifies the packet number modulo 2048 within the current packetization unit. It is set to 0 at the start of the packetization unit and incremented by 1 for every subsequent packet (if any) belonging to the same unit. Practically, if codestream packetization mode is enabled, this field counts the packets within a JPEG XS picture segment and is extended by the SEP counter when it overruns. If slice packetization mode is enabled, this field counts the packets within a slice or within the JPEG XS header segment.

The payload data of a JPEG XS RTP stream consists of a concatenation of multiple JPEG XS frames. Within the RTP stream, all of the video support boxes and all of the colour specification boxes SHALL retain their respective layouts for each JPEG XS frame. Thus, each video support box in the RTP stream SHALL define the same sub boxes. The effective values in the boxes are allowed to change under the condition that their relative byte offsets SHALL not change. Each JPEG XS frame is the concatenation of one or more packetization unit(s), as explained in . depicts this layout for a progressive frame in the codestream packetization mode, depicts this layout for an interlaced frame in the codestream packetization mode, depicts this layout for a progressive frame in the slice packetization mode and depicts this layout for an interlaced frame in the slice packetization mode. The Frame counter value is not indicated because the value is constant for all packetization units of a given frame.

The traffic shaping and delivery timing SHALL be in accordance with the Network Compatibility Model compliance definitions specified in SMPTE ST 2110-21 for either Narrow Linear Senders (Type NL) or Wide Senders (Type W). The session description SHALL include a format-specific parameter of either TP=2110TPNL or TP=2110TPW to indicate compliance with Type NL or Type W respectively. NOTE: The Virtual Receiver Buffer Model compliance definitions of ST 2110-21 do not apply.

Congestion control for RTP SHALL be used in accordance with RFC 3550, and with any applicable RTP profile: e.g., RFC 3551. An additional requirement if best-effort service is being used is users of this payload format SHALL monitor packet loss to ensure that the packet loss rate is within acceptable parameters. Circuit Breakers is an update to RTP that defines criteria for when one is required to stop sending RTP Packet Streams and applications implementing this standard SHALL comply with it. RFC 8085 provides additional information on the best practices for applying congestion control to UDP streams.

video jxsv The RTP timestamp clock rate. Applications using this payload format SHALL use a value of 90000. This parameter specifies the configured transmission mode as defined by the Transmission mode (T) bit in the payload header of . This value SHALL be equal to the T bit value configured in the RTP stream (i.e. 0 for out-of-order-allowed or 1 for sequential). This parameter specifies the configured packetization mode as defined by the pacKetization mode (K) bit in the payload header of . If specified, this value SHALL be equal to the K bit value configured in the RTP stream (i.e. 0 for codestream or 1 for slice). The JPEG XS profile in use, as defined in ISO/IEC 21122-2 (JPEG XS Part 2). Any white space in the profile name SHALL be replaced by a dash (-). Examples are 'Main-444.12' or 'High-444.12'. The JPEG XS level in use, as defined in ISO/IEC 21122-2 (JPEG XS Part 2). Any white space in the level name SHALL be replaced by a dash (-). Examples are '2k-1' or '4k-2'. The JPEG XS sublevel in use, as defined in ISO/IEC 21122-2 (JPEG XS Part 2). Any white space in the sublevel name SHALL be replaced by a dash (-). Examples are 'Sublev3bpp' or 'Sublev6bpp'. Determines the number of bits per sample. This is an integer with typical values including 8, 10, 12, and 16. Determines the number of pixels per line. This is an integer between 1 and 32767. Determines the number of lines per frame. This is an integer between 1 and 32767. Signals the frame rate in frames per second. Integer frame rates SHALL be signaled as a single decimal number (e.g. "25") whilst non-integer frame rates SHALL be signaled as a ratio of two integer decimal numbers separated by a "forward-slash" character (e.g. "30000/1001"), utilizing the numerically smallest numerator value possible. If this parameter name is present, it indicates that the video is interlaced, or that the video is Progressive segmented Frame (PsF). If this parameter name is not present, the progressive video format SHALL be assumed. If this parameter name is present, and the interlace parameter name is also present, then the video is a Progressive segmented Frame (PsF). Signaling of this parameter without the interlace parameter is forbidden. Signals the colour difference signal sub-sampling structure. Signals utilizing the non-constant luminance Y'C'B C'R signal format of Recommendation ITU-R BT.601-7, Recommendation ITU-R BT.709-6, Recommendation ITU-R BT.2020-2, or Recommendation ITU-R BT.2100 SHALL use the appropriate one of the following values for the Media Type Parameter "sampling":

Signals utilizing the Constant Luminance Y'C C'BC C'RC signal format of Recommendation ITU-R BT.2020-2 SHALL use the appropriate one of the following values for the Media Type Parameter "sampling":

Signals utilizing the constant intensity I CT CP signal format of Recommendation ITU-R BT.2100 SHALL use the appropriate one of the following values for the Media Type Parameter "sampling":

Signals utilizing the 4:4:4 R' G' B' or RGB signal format (such as that of Recommendation ITU-R BT.601, Recommendation ITU-R BT.709, Recommendation ITU-R BT.2020, Recommendation ITU-R BT.2100, SMPTE ST 2065-1 or ST 2065-3) SHALL use the following value for the Media Type Parameter sampling.

Signals utilizing the 4:4:4 X' Y' Z' signal format (such as defined in SMPTE ST 428-1) SHALL use the following value for the Media Type Parameter sampling.

Key signals as defined in SMPTE RP 157 SHALL use the value key for the Media Type Parameter sampling. The Key signal is represented as a single component.

Signals utilizing a colour sub-sampling other than what is defined here SHALL use the following value for the Media Type Parameter sampling.

Specifies the system colorimetry used by the image samples. Valid values and their specification are:

Signals utilizing the Recommendation ITU-R BT.2100 colorimetry SHOULD also signal the representational range using the optional parameter RANGE defined below. Signals utilizing the UNSPECIFIED colorimetry might require manual coordination between the sender and the receiver. Transfer Characteristic System. This parameter specifies the transfer characteristic system of the image samples. Valid values and their specification are:

This parameter SHOULD be used to signal the encoding range of the sample values within the stream. When paired with ITU Rec BT.2100 colorimetry, this parameter has two allowed values NARROW and FULL, corresponding to the ranges specified in table 9 of ITU Rec BT.2100. In any other context, this parameter has three allowed values: NARROW, FULLPROTECT, and FULL, which correspond to the ranges specified in SMPTE RP 2077. In the absence of this parameter, and for all but the UNSPECIFIED colometries, NARROW SHALL be the assumed value. When paired with the UNSPECIFIED colometry, FULL SHALL be the default assumed value. This media type is framed and binary; see Section 4.8 in RFC 6838. Please see the Security Considerations section in RFC XXXX

A Session Description Protocol (SDP) media description SHALL be created for each RTP stream and it SHALL be in accordance with the provisions of SMPTE ST 2110-10. The information carried in the media type specification has a specific mapping to the SDP fields, used to describe RTP sessions. This information is redundant with the information found in the payload data (namely, in the JPEG XS header segment) and SHALL be consistent with it. In case of discrepancy between parameters values found in the payload data and in the SDP fields, the values from the payload data SHALL prevail.

The media type ("video") goes in SDP "m=" as the media name. The media subtype ("jxsv") goes in SDP "a=rtpmap" as the encoding name, followed by a slash ("/") and the required parameter "rate" corresponding to the RTP timestamp clock rate (which for the payload format defined in this document SHALL be 90000). The required parameter "transmode" and the additional optional parameters go in the SDP "a=fmtp" attribute by copying them directly from the MIME media type string as a semicolon-separated list of parameter=value pairs. A sample SDP mapping for JPEG XS video is as follows:

In this example, a JPEG XS RTP stream is being sent to UDP destination port 30000, with an RTP dynamic payload type of 112 and a media clock rate of 90000 Hz. Note that the "a=fmtp:" line has been wrapped to fit this page, and will be a single long line in the SDP file.

The SDP media description SHALL include the TP parameter (either 2110TPNL or 2110TPW as specified in ) and may include the CMAX parameter as specified in SMPTE ST 2110-21.

When XS is offered using An Offer/Answer Model with Session Description Protocol (SDP) for negotiation for unicast usage, the following limitations and rules apply: All parameters are declarative, i.e. apply only to media sent by the entity that generated the SDP RFC 4568. Thus, a declarative parameter in an offer applies to media sent by the offeror, whereas a declarative parameter in an answer applies to media sent by the answerer. All parameters must be supported by both sides, i.e. the answerer SHALL either maintain all parameters or remove the media format (payload type) completely if one or more of the parameter values are not supported.

This memo requests that IANA registers video/jxsv as specified in . The media type is also requested to be added to the IANA registry for "RTP Payload Format MIME types".

RTP packets using the payload format defined in this specification are subject to the security considerations discussed in the RTP specification and in any applicable RTP profile such as RTP/AVP, RTP/AVPF, RTP/SAVP, or RTP/SAVPF. This implies that confidentiality of the media streams is achieved by encryption. However, as "Securing the RTP Framework: Why RTP Does Not Mandate a Single Media Security Solution" discusses, it is not an RTP payload format's responsibility to discuss or mandate what solutions are used to meet the basic security goals like confidentiality, integrity, and source authenticity for RTP in general. This responsibility lies on anyone using RTP in an application. They can find guidance on available security mechanisms and important considerations in "Options for Securing RTP Sessions". Applications SHOULD use one or more appropriate strong security mechanisms. This payload format and the JPEG XS encoding do not exhibit any substantial non-uniformity, either in output or in complexity to perform the decoding operation and thus are unlikely to pose a denial-of-service threat due to the receipt of pathological datagrams. It is important to note that HD or UHDTV JPEG XS-encoded video can have significant bandwidth requirements (typically more than 1 Gbps for ultra high-definition video, especially if using high framerate). This is sufficient to cause potential for denial-of-service if transmitted onto most currently available Internet paths. Accordingly, if best-effort service is being used, users of this payload format SHALL monitor packet loss to ensure that the packet loss rate is within acceptable parameters. Packet loss is considered acceptable if a TCP flow across the same network path, and experiencing the same network conditions, would achieve an average throughput, measured on a reasonable timescale, that is not less than the RTP flow is achieving. This condition can be satisfied by implementing congestion control mechanisms to adapt the transmission rate (or the number of layers subscribed for a layered multicast session), or by arranging for a receiver to leave the session if the loss rate is unacceptably high. This payload format may also be used in networks that provide quality-of-service guarantees. If enhanced service is being used, receivers SHOULD monitor packet loss to ensure that the service that was requested is actually being delivered. If it is not, then they SHOULD assume that they are receiving best-effort service and behave accordingly.

The authors would like to thank the following people for their valuable contributions to this specification: Arnaud Germain, Alexandre Willème, Gaël Rouvroy, and Jean-Baptise Lorent.

Note to RFC Editor: This section may be removed after carrying out all the instructions of this section. RFC XXXX is to be replaced by the RFC number this specification receives when published.