RFC 4396






Network Working Group                                             J. Rey
Request for Comments: 4396                                     Y. Matsui
Category: Standards Track                                      Panasonic
                                                           February 2006


                           RTP Payload Format
       for 3rd Generation Partnership Project (3GPP) Timed Text

Status of This Memo



   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice



   Copyright (C) The Internet Society (2006).

Abstract



   This document specifies an RTP payload format for the transmission of
   3GPP (3rd Generation Partnership Project) timed text.  3GPP timed
   text is a time-lined, decorated text media format with defined
   storage in a 3GP file.  Timed Text can be synchronized with
   audio/video contents and used in applications such as captioning,
   titling, and multimedia presentations.  In the following sections,
   the problems of streaming timed text are addressed, and a payload
   format for streaming 3GPP timed text over RTP is specified.




















Rey & Matsui                Standards Track                     [Page 1]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


Table of Contents



   1. Introduction ....................................................3
   2. Motivation, Requirements, and Design Rationale ..................3
      2.1. Motivation .................................................3
      2.2. Basic Components of the 3GPP Timed Text Media Format .......4
      2.3. Requirements ...............................................5
      2.4. Limitations ................................................6
      2.5. Design Rationale ...........................................7
   3. Terminology ....................................................10
   4. RTP Payload Format for 3GPP Timed Text .........................12
      4.1. Payload Header Definitions ................................13
           4.1.1. Common Payload Header Fields .......................15
           4.1.2. TYPE 1 Header ......................................17
           4.1.3. TYPE 2 Header ......................................20
           4.1.4. TYPE 3 Header ......................................23
           4.1.5. TYPE 4 Header ......................................24
           4.1.6. TYPE 5 Header ......................................25
      4.2. Buffering of Sample Descriptions ..........................25
           4.2.1. Dynamic SIDX Wraparound Mechanism ..................26
      4.3. Finding Payload Header Values in 3GP Files ................28
      4.4. Fragmentation of Timed Text Samples .......................31
      4.5. Reassembling Text Samples at the Receiver .................33
      4.6. On Aggregate Payloads .....................................35
      4.7. Payload Examples ..........................................39
      4.8. Relation to RFC 3640 ......................................43
      4.9. Relation to RFC 2793 ......................................44
   5. Resilient Transport ............................................45
   6. Congestion Control .............................................46
   7. Scene Description ..............................................47
      7.1. Text Rendering Position and Composition ...................47
      7.2. SMIL Usage ................................................48
      7.3. Finding Layout Values in a 3GP File .......................48
   8. 3GPP Timed Text Media Type .....................................49
   9. SDP Usage ......................................................53
      9.1. Mapping to SDP ............................................53
      9.2. Parameter Usage in the SDP Offer/Answer Model .............53
           9.2.1. Unicast Usage ......................................54
           9.2.2. Multicast Usage ....................................57
      9.3. Offer/Answer Examples .....................................58
      9.4. Parameter Usage outside of Offer/Answer ...................60
   10. IANA Considerations ...........................................60
   11. Security Considerations .......................................60
   12. References ....................................................61
      12.1. Normative References .....................................61
      12.2. Informative References ...................................61
   13. Basics of the 3GP File Structure ..............................64
   14. Acknowledgements ..............................................65



Rey & Matsui                Standards Track                     [Page 2]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


1.  Introduction



   3GPP timed text is a media format for time-lined, decorated text
   specified in the 3GPP Technical Specification TS 26.245, "Transparent
   end-to-end packet switched streaming service (PSS); Timed Text Format
   (Release 6)" [1].  Besides plain text, the 3GPP timed text format
   allows the creation of decorated text such as that for karaoke
   applications, scrolling text for newscasts, or hyperlinked text.
   These contents may or may not be synchronized with other media, such
   as audio or video.

   The purpose of this document is to provide a means to stream 3GPP
   timed text contents using RTP [3].  This includes the streaming of
   timed text being read out of a (3GP) file, as well as the streaming
   of timed text generated in real-time, a.k.a. live streaming.

   Section 2 contains the motivation for this document, an overview of
   the media format, the requirements, and the design rationale.
   Section 3 defines the terminology used.  Section 4 specifies the
   payload headers, the fragmentation and re-assembly rules for text
   samples, the rules for payload aggregation, and the relations of this
   document to RFC 3640 [12] and RFC 2793 [22].  Section 5 specifies
   some simple schemes for resilient transport and gives pointers to
   other possible mechanisms.  Section 6 addresses congestion control.
   Section 7 specifies scene description.  Section 8 defines the media
   type.  Section 9 specifies SDP for unicast and multicast sessions,
   including usage in the Offer/Answer model [13].  Sections 10 and 11
   address IANA and security considerations.  Section 12 lists
   references.  Basics of the 3GP File Structure are in Section 13.

2.  Motivation, Requirements, and Design Rationale



2.1.  Motivation



   The 3GPP timed text format was developed for use in the services
   specified in the 3GPP Transparent End-to-end Packet-switched
   Streaming Services (3GPP PSS) specification [16].

   As of today, PSS allows downloading 3GPP timed text contents stored
   in 3GP files.  However, due to the lack of a RTP payload format, it
   is not possible to stream 3GPP timed text contents over RTP.

   This document specifies such a payload format.








Rey & Matsui                Standards Track                     [Page 3]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


2.2.  Basic Components of the 3GPP Timed Text Media Format



   Before going into the details of the design, it is necessary to know
   how the media format is constructed.  We can identify four
   differentiated functional components: layout information, default
   formatting, text strings, and decoration.  In the following, we
   shortly explain these and match them to their designations in a 3GP
   file:

        o Initial spatial layout information related to the text
          strings: These are the height and width of the text region
          where text is displayed, the position of the text region in
          the display, and the layer or proximity of the text to the
          user.  In 3GP files, this information is contained in the
          Track Header Box (3GP file designations are capitalized for
          clarity).

        o Default settings for formatting and positioning of text: style
          (font, size, color,...), background color, horizontal and
          vertical justification, line width, scrolling, etc.  For 3GP
          files, this corresponds to the Sample Descriptions.

        o The actual text strings: encoded characters using either UTF-8
          [18] or UTF-16 [19] encoding.

        o The decoration: If some characters have different style,
          delay, blink, etc., this needs to be indicated.  The
          decoration is only present in the text samples if it is
          actually needed.  Otherwise, the default settings as above
          apply.  In 3GP files, within each Text Sample, the decoration
          (i.e., Modifier Boxes) is appended to the text strings, if
          needed.  At the time of writing this payload format, the
          following modifiers are specified in the 3GPP timed text media
          format specification [1]:

           - text highlight
           - highlight color
           - blinking text
           - karaoke feature
           - hyperlink
           - text delay
           - text style
           - positioning of the text box
           - text wrap indication







Rey & Matsui                Standards Track                     [Page 4]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


2.3.  Requirements



   Once the basic components are known, it is necessary to define which
   requirements the payload format shall fulfill:

     1. It shall enable both live streaming and streaming from a 3GP
        file.

                Informative note: For the purpose of this document, the
                term "live streaming" refers to those scenarios where
                the timed text stream is sent from a live encoder.  Upon
                reception, the content may or may not be stored in a 3GP
                file.  Typically, in live streaming applications, the
                sender encapsulates the timed text content in RTP
                packets following the guidelines given in this document.
                At the receiving side, a buffer is used to cancel the
                network delay and delay jitter.  If receiver and sender
                support packet loss resilience mechanisms (see Section
                5), it may also be possible to recover from packet
                losses.  Note that how sender and receiver actually
                manage and dimension the buffers is an implementation
                design choice.

     2. Furthermore, it shall be possible for an RTP receiver using this
        payload format, and capable of storing in 3GP format, to obtain
        all necessary information from the RTP packets for storing the
        received text contents according to the 3GP file format.  This
        file may or may not be the same as the original file.

                Informative note: The 3GP file format itself is based on
                the ISO Base Media File Format recommendation [2].
                Section 13.1 gives some insight into the 3GP file
                structure.  Further, Sections 4.3 and 7.3 specify where
                the information needed for filling in payload headers is
                found in a 3GP file.  For live streaming, appropriate
                values complying with the format and units described in
                [1] shall be used.  Where needed, clarifications on
                appropriate values are given in this document.

     3. It shall enable efficient and resilient transport of timed text
        contents over RTP.  In particular:



          a. Enable the transmission of the sample descriptions by both
             out-of-band and in-band means.  Sample descriptions are
             important information, which potentially apply to several
             text samples.  These default formatting settings are
             typically transmitted out-of-band (reliably) once at the
             initialization phase.  If additional sample descriptions



Rey & Matsui                Standards Track                     [Page 5]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


             are needed in the course of a session, these may also be
             sent out-of-band or in-band.  In-band transmission,
             although unreliable, may be more appropriate for sending
             sample descriptions if these should be sent frequently, as
             opposed to establishing an additional communication channel
             for SDP, for example.  It is also useful in cases where an
             out-of-band channel may not be available and for live
             streaming, where contents are not known a priori.  Thus,
             the payload format shall enable out-of-band and in-band
             transmission of sample descriptions.  Section 4.1.6
             specifies a payload header for transmitting sample
             descriptions in-band.  Section 9 specifies how sample
             descriptions are mapped to SDP.

          b. Enable the fragmentation of a text sample into several RTP
             packets in order to cover a wide range of applications and
             network environments.  In general, fragmentation should be
             a rare event, given the low bit rates and relatively small
             text sample sizes.  However, the 3GPP Timed Text media
             format does allow for larger text samples.  Therefore, the
             payload format shall take this into account and provide a
             means for coping with fragmentation and reassembly. Section
             4.4 deals with fragmentation.

          c. Enable the aggregation of units into an RTP packet for
             making the transport more efficient.  In a mobile
             communication environment, a typical text sample size is
             around 100-200 bytes.  If the available bit rate and the
             packet size allow it, units should be aggregated into one
             RTP packet.  Section 4.6 deals with aggregation.

          d. Enable the use of resilient transport mechanisms, such as
             repetition, retransmission [11], and FEC [7] (see Section
             5).  For a more general discussion, refer to RFC 2354 [8],
             which discusses available mechanisms for stream repair.

2.4.  Limitations



     The payload headers have been optimized in size for RTP.  Instead
     of using 32-bit (S)LEN, SDUR, and SIDX header fields, which would
     carry many unused bits much of the time, it has been a design
     choice to reduce the size of these fields.  As a consequence, this
     payload format has reduced maximum values with respect to sizes and
     durations of (text) samples and sample descriptions.  These maximum
     values differ from those allowed in 3GP files, where they are
     expressed using 32-bit (unsigned) integers.  In some cases,





Rey & Matsui                Standards Track                     [Page 6]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     extension mechanisms are provided to deal with larger values.
     However, it is noted that the values used here should be enough for
     the streaming applications targeted.

     The following limitations apply:

     1. The maximum size of text samples carried in RTP packets is
        restricted to be a 16-bit (unsigned) integer (this includes the
        text strings and modifiers).  This means a maximum size for the
        unit would be about 64 Kbytes.  No extension mechanism is
        provided.

     2. The sample description index values are restricted to be an 8-
        bit (unsigned) integer.  An extension mechanism is given in
        Section 4.3.

     3. The text sample duration is restricted to be a 24-bit (unsigned)
        integer.  This yields a maximum duration at a timestamp
        clockrate of 1000 Hz of about 4.6 hours.  Nevertheless, an
        extension mechanism is provided in Section 4.3.

     4. Sample descriptions are also restricted in size: If the size
        cannot be expressed as a 16-bit (unsigned) integer, the sample
        description shall not be conveyed.  As in the case of the sample
        size, no extension mechanism is provided.

     5. A further limitation concerns the UTF-16 encodings supported:
        Only transport of text strings following big endian byte order
        is supported.  See Section 4.1.1 for details.

2.5.  Design Rationale



   The following design choices were made:

     1. 'Unit' approach: The payload formats specified in this document
        follow a simple scheme: a 3-byte common header (Common Payload
        Header) followed by a specific header for each text sample
        (fragment) type.  Following these headers, the text sample
        contents are placed (Section 4.1.1 and following).  This
        structure is called a 'unit'.

        The following units have been devised to comply with the
        requirements mentioned in Section 2.3:

          a. A TYPE 1 unit that contains one complete text sample,

          b. A TYPE 2 unit that contains a complete text string or a
             fragment thereof,



Rey & Matsui                Standards Track                     [Page 7]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


          c. A TYPE 3 unit that contains the complete modifiers or only
             the first fragment thereof,

          d. A TYPE 4 unit that contains one modifier fragment other
             than the first, and

          e. A TYPE 5 unit that contains one sample description.

        This 'unit' approach was motivated by the following reasons:

              1. Allows a simple classification of the text samples and
                 text sample fragments that can be conveyed by the
                 payload format.

              2. Enables easy interoperability with RFC 3640 [12].
                 During the development of this payload format, interest
                 was shown from MPEG-4 standardization participants in
                 developing a common payload structure for the transport
                 of 3GPP Timed Text.  While interoperability is not
                 strictly necessary for this payload format to work, it
                 has been pursued in this payload format.  Section 4.8
                 explains how this is done.

     2. Character count is not implemented.  This payload format does
        detect lost text samples fragments, but it does not enable an
        RTP receiver to find out the exact number of text characters
        lost.  In fact, the fragment size included in the payload
        headers does not help in finding the number of lost characters
        because the UTF-8/UTF-16 [18][19] encodings used yield a
        variable number of bytes per character.

        For finding the exact number of lost characters, an additional
        field reflecting the character count (and possibly the character
        offset) upon fragmentation would be required.  This would
        additionally require that the entity performing fragmentation
        count the characters included in each text fragment.

        One benefit of having a character count would be that the
        display application would be able to replace missing characters
        through some other character representing character loss.  For
        example:

             If we take the "Some text is lost now" and assume the loss
             of a packet containing the text in the middle, this could
             be displayed (with a character count):

             "Some ############now"




Rey & Matsui                Standards Track                     [Page 8]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


             As opposed to:

             "Some #now"

             which is what this payload format enables ("#" indicates a
             missing character or packet, respectively).

        However, it is the consensus of the working group that for
        applications such as subtitling applications and multimedia
        presentations that use this payload format, such partial error
        correction is not worth the cost of including two additional
        fields; namely, character count and character offset.  Instead,
        it is recommended that some more overhead be invested to provide
        full error correction by protecting the less text sample
        fragments using the measures outlined in Section 5.

     3. Fragment re-assembly: In order to re-assemble the text samples,
        offset information is needed.  Instead of a character or byte
        offset, a single byte, TOTAL/THIS, is used.  These two values
        indicate the total number and current index of fragments of a
        text sample.  This is simpler than having a character offset
        field in each fragment.  Details in Section 4.1.3.

     4. A length field, LEN, is present in the common header fields.
        While the length in the RTP payload format is not needed by most
        RTP applications (typically lower layers, like UDP, provide this
        information), it does ease interoperability with RFC 3640.  This
        is because the Access Units (AUs) used for carriage of data in
        RFC 3640 must include a length indication.  Details are in
        Section 4.8.

     5. The header fields in the specific payload headers (TYPE headers
        in Sections 4.1.2 to 4.1.6) have been arranged for easy
        processing on 32-bit machines.  For this reason, the fields SIDX
        and SDUR are swapped in TYPE 1 unit, compared to the other
        units.















Rey & Matsui                Standards Track                     [Page 9]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


3.  Terminology

   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 [5].

   Furthermore, the following terms are used and have specific meaning
   within the context of this document:

   text sample or whole text sample

        In the 3GPP Timed Text media format [1], these terms refer to a
        unit of timed text data as contained in the source (3GP) file.
        This includes the text string byte count, possibly a Byte Order
        Mark, the text string and any modifiers that may follow.  Its
        equivalent in audio/video would be a frame.

        In this document, however, a text sample contains only text
        strings followed by zero or more modifiers.  This definition of
        text sample excludes the 16-bit text string byte count and the
        16-bit Byte Order Mark (BOM) present in 3GP file text samples
        (see Section 4.3 and Figure 9).  The 16-bit BOM is not
        transported in RTP, as explained in Section 4.1.1.

   text strings

        The actual text characters encoded either as UTF-8 or UTF-16.
        When using this payload format, the text string does not contain
        any byte order mark (BOM).  See Figure 9 for details.

   fragment or text sample fragment

        A fraction of a text sample.  A fragment may contain either text
        strings or modifier (decoration) contents, but not both at the
        same time.

   sample contents

        General term to identify timed text data transported when using
        this payload format.  Sample contents may be one or several text
        samples, sample descriptions, and sample fragments (note that,
        as per Section 4.6, there is only one case in which more than
        one fragment may be included in a payload).








Rey & Matsui                Standards Track                    [Page 10]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   decoration or modifiers

        These terms are used interchangeably throughout the document to
        denote the contents of the text sample that modify the default
        text formatting.  Modifiers may, for example, specify different
        font size for a particular sequence of characters or define
        karaoke timing for the sample.

   sample description

        Information that is potentially shared by more than one text
        sample.  In a 3GP file, a sample description is stored in a
        place where it can be shared.  It contains setup and default
        information such as scrolling direction, text box position,
        delay value, default font, background color, etc.

   units or transport units

        The payload headers specified in this document encapsulate text
        samples, fragments thereof, and sample descriptions by placing a
        common header and specific payload header (Sections 4.1.1 to
        4.1.6) before them, thus building what is here called a
        (transport) unit.

   aggregation or aggregate packet

        The payload of an aggregate (RTP) packet consists of several
        (transport) units.

   track or stream

        3GP files contain audio/video and text tracks.  This document
        enables streaming of text tracks using RTP.  Therefore, these
        terms are used interchangeably in this document in the context
        of 3GP files.

   Media Header Box / Track Header Box / ...

        The 3GP file format makes use of these structures defined in the
        ISO Base File Format [2].  When referring to these in this
        document, initials are capitalized for clarity.










Rey & Matsui                Standards Track                    [Page 11]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.  RTP Payload Format for 3GPP Timed Text



   The format of an RTP packet containing 3GPP timed text is shown
   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            |
     /+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    | |U|   R   | TYPE|             LEN               |               :
    | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+               :
   U| :           (variable header fields depending on TYPE           :
   N| :                                                               :
   I< +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   T| |                                                               |
    | :                    SAMPLE CONTENTS                            :
    | |                                               +-+-+-+-+-+-+-+-+
    | |                                               |
     \+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

               Figure 1. 3GPP Timed Text RTP Packet Format

   Marker bit (M): The marker bit SHALL be set to 1 if the RTP packet
   includes one or more whole text samples or the last fragment of a
   text sample; otherwise, it is set to zero (0).

   Timestamp: The timestamp MUST indicate the sampling instant of the
   earliest (or only) unit contained in the RTP packet.  The initial
   value SHOULD be randomly determined, as specified in RTP [3].

        The timestamp value should provide enough timing resolution for
        expressing the duration of text samples, for synchronizing text
        with other media, and for performing RTP Control Protocol (RTCP)
        measurements such as the interarrival delay jitter or the RTCP
        Packet Receipt Times Report Block (Section 4.3 of RFC 3611
        [20]).  This is compliant to RTP, Section 5.1:

             "The resolution of the clock MUST be sufficient for the
             desired synchronization accuracy and for measuring packet
             arrival jitter (one tick per video frame is typically not
             sufficient)".





Rey & Matsui                Standards Track                    [Page 12]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


        The above observation applies to both timed text tracks included
        in a 3GP file and live streaming sessions.  In the case of a 3GP
        timed text track, the timestamp clockrate is the value of the
        "timescale" parameter in the Media Header Box for that text
        track.  Each track in a 3GP file MAY have its own clockrate as
        specified in the Media Header Box.  Likewise, live streaming
        applications SHALL use an appropriate timestamp clockrate.  A
        default value of 1000 Hz is RECOMMENDED.  Other timestamp
        clockrates MAY be used.  In this case, the typical behavior here
        is to match the 3GPP timed text clockrate to that used by an
        associated audio or video stream.

        In an aggregate payload, units MUST be placed in play-out order,
        i.e., earliest first in the payload.  If TYPE 1 units are
        aggregated, the timestamp of the subsequent units MUST be
        obtained by adding the timed text sample duration of previous
        samples to the RTP timestamp value.  There are two exceptions to
        this rule: TYPE 5 units and an aggregate payload containing two
        fragments of the same text sample.  The details of the timestamp
        calculation are given in Section 4.6.

        Finally, timestamp clockrates MUST be signaled by out-of-band
        means at session setup, e.g., using the media type "rate"
        parameter in SDP.  See Section 9 for details.

   Payload Type (PT): The payload type is set dynamically and sent by
   out-of-band means.

   The usage of the remaining RTP header fields (namely, V, P, X, CC, SN
   and SSRC) follows the rules of RTP and the profile in use.

4.1.  Payload Header Definitions



   The (transport) units specified in this document consist of a set of
   common fields (U, R, TYPE, LEN), followed by specific header fields
   (TYPES 1-5) and text sample contents.  See Figure 1 and Figure 2.

   In Figure 2, two example RTP packets are depicted.  The first
   contains an aggregate RTP payload with two complete text samples, and
   the second contains one text sample fragment.  After each unit header
   is explained, detailed payload examples follow in Section 4.7.










Rey & Matsui                Standards Track                    [Page 13]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


                                        +----------------------+
                                        |                      |
                                        |   RTP Header         |
                                        |                      |
                               ---------+----------------------+
                               |        |                      |
                               |        |COMMON + TYPE 1 Header|
                               |        ........................
                        UNIT 1 -        |                      |
                               |        |    Text Sample       |
                               |        |                      |
                               |-------\........................
                                -------/|                      |
                               |        |COMMON + TYPE 1 Header|
                               |        ........................
                        UNIT 2 -        |                      |
                               |        |    Text Sample       |
                               |        |                      |
                               |        |                      |
                               ---------+----------------------+

                                        +----------------------+
                                        |                      |
                                        |   RTP Header         |
                                        |                      |
                               ---------+----------------------+
                               |        |  COMMON + TYPE 2     |
                               |        |    (or 3 or 4) Hdr   |
                               |        ........................
                        UNIT 3 -        |                      |
                               |        | Text Sample Fragment |
                               |        |                      |
                               |        |                      |
                               ---------+----------------------+

                     Figure 2.  Example RTP packets















Rey & Matsui                Standards Track                    [Page 14]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.1.1.  Common Payload Header Fields



   The fields common to all payload headers have the following format:

            0                   1                   2
            0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3
           +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
           |U|   R   |TYPE |             LEN               |
           +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                Figure 3.  Common payload header fields

   Where:

   o U (1 bit) "UTF Transformation flag": This is used to inform RTP
     receivers whether UTF-8 (U=0) or UTF-16 (U=1) was used to encode
     the text string.  UTF-16 text strings transported by this payload
     format MUST be serialized in big endian order, a.k.a. network byte
     order.

        Informative note: Timed text clients complying with the 3GPP
        Timed Text format [1] are only required to understand the big
        endian serialization.  Thus, in order to ease interoperability,
        the reverse serialization (little endian) is not supported by
        this payload format.

     For the payload formats defined in this document, the U bit is only
     used in TYPE 1 and TYPE 2 headers.  Senders MUST set the U bit to
     zero in TYPE 3, TYPE 4, and TYPE 5 headers.  Consequently,
     receivers MUST ignore the U bit in TYPE 3, TYPE 4, and TYPE 5
     headers.

   o R (4 bits) "Reserved bits": for future extensions.  This field MUST
     be set to zero (0x0) and MUST be ignored by receivers.

   o TYPE (3 bits) "Type Field": This field specifies which specific
     header fields follow.  The following TYPE values are defined:

        - TYPE 1, for a whole text sample.
        - TYPE 2, for a text string fragment (without modifiers).
        - TYPE 3, for a whole modifier box or the first fragment of a
          modifier box.
        - TYPE 4, for a modifier fragment other than first.
        - TYPE 5, for a sample description.  Exactly one header per
          sample description.
        - TYPE 0, 6, and 7 are reserved for future extensions.  Note
          that future extensions are possible, e.g., a unit that
          explicitly signals the number of characters present in a



Rey & Matsui                Standards Track                    [Page 15]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


          fragment (see Section 2.5).  In order to guarantee backwards-
          compatibility, it SHALL be possible that older clients ignore
          (newer) units they do not understand, without invalidating the
          timestamp calculation mechanisms or otherwise preventing them
          from decoding the other units.

   o Finally, the LEN (16 bits) "Length Field": indicates the size (in
     bytes) of this header field and all the fields following, i.e., the
     LEN field followed by the unit payload: text strings and modifiers
     (if any).  This definition only excludes the initial U/R/TYPE byte
     of the common header.  The LEN field follows network byte order.

     The way in which LEN is obtained when streaming out of a 3GP file
     depends on the particular unit type.  This is explained for each
     unit in the sections below.

     For live streaming, both sample length and the LEN value for the
     current fragment MUST be calculated during the sampling process or
     during fragmentation.

     In general, LEN may take the following values:

      - TYPE = 1, LEN >= 8
      - TYPE = 2, LEN > 9
      - TYPE = 3, LEN > 6
      - TYPE = 4, LEN > 6
      - TYPE = 5, LEN > 3

     Receivers MUST discard units that do not comply with these values.
     However, the RTP header fields and the rest of the units in the
     payload (if any) are still useful, as guaranteed by the requirement
     for future extensions above.

     In the following subsections the different payload headers for the
     values of TYPE are specified.
















Rey & Matsui                Standards Track                    [Page 16]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.1.2.  TYPE 1 Header



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE |       LEN  (always >=8)       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |     TLEN      |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |      TLEN     |
      +-+-+-+-+-+-+-+-+

                    Figure 4.  TYPE 1 Header Format

   This header type is used to transport whole text samples.  This unit
   should be the most common case, i.e., the text sample should usually
   be small enough to be transported in one unit without having to
   separate text strings from modifiers.  In an aggregate (RTP packet)
   payload containing several text samples, every sample is preceded by
   its own TYPE 1 header (see Figure 12).

        Informative note: As indicated in Section 3, "Terminology", a
        text sample is composed of the text strings followed by the
        modifiers (if any).  This is also how text samples are stored in
        3GP files.  The separation of a text sample into text strings
        and modifiers is only needed for large samples (or small
        available IP MTU sizes; see Section 4.4), and it is accomplished
        with TYPE 2 and TYPE 3 headers, as explained in the sections
        below.

   Note also that empty text samples are considered whole text samples,
   although they do not contain sample contents.  Empty text samples may
   be used to clear the display or to put an end to samples of unknown
   duration, for example.  Units without sample contents SHALL have a
   LEN field value of 8 (0x0008).

   The fields above have the following meaning:

   o U, R, and TYPE, as defined in Section 4.1.1.

   o LEN, in this case, represents the length of the (complete) text
     sample plus eight (8) bytes of headers.  For finding the length of
     the text sample in the Sample Size Box of 3GP files, see Section
     4.3.

   o SIDX (8 bits) "Text Sample Entry Index": This is an index used to
     identify the sample descriptions.




Rey & Matsui                Standards Track                    [Page 17]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     The SIDX field is used to find the sample description corresponding
     to the unit's payload.  There are two types of SIDX values: static
     and dynamic.

     Static SIDX values are used to identify sample descriptions that
     MUST be sent out-of-band and MUST remain active during the whole
     session.  A static SIDX value is unequivocally linked to one
     particular sample description during the whole session.  Carrying
     many sample descriptions out-of-band SHOULD be avoided, since these
     may become large and, ultimately, transport is not the goal of the
     out-of-band channel.  Thus, this feature is RECOMMENDED for
     transporting those sample descriptions that provide a set of
     minimum default format settings.  Static SIDX values MUST fall in
     the (closed) interval [129,254].

     Dynamic SIDX values are used for sample descriptions sent in-band.
     Sample descriptions MAY be sent in-band for several reasons:
     because they are generated in real time, for transport resiliency,
     or both.  A dynamic SIDX value is unequivocally linked to one
     particular sample description during the period in which this is
     active in the session, and it SHALL NOT be modified during that
     period.  This period MAY be smaller than or equal to the session
     duration.  This period is not known a priori.  A maximum of 64
     dynamic simultaneously active SIDX values is allowed at any moment.
     Dynamic SIDX values MUST fall in the closed interval [0,127].  This
     should be enough for both recorded content and live streaming
     applications.  Nevertheless, a wraparound mechanism is provided in
     Section 4.2.1 to handle streaming sessions where more than 64 SIDX
     values might be needed.  Servers MAY make use of dynamic sample
     descriptions.  Clients MUST be able to receive and interpret
     dynamic sample descriptions.

     Finally, SIDX values 128 and 255 are reserved for future use.

   o SDUR (24 bits) "Text Sample Duration": indicates the sample
     duration in RTP timestamp units of the text sample.  For this
     field, a length of 3 bytes is preferred to 2 bytes.  This is
     because, for a typical clockrate of 1000 Hz, 16 bits would allow
     for a maximum duration of just 65 seconds, which might be too short
     for some streams.  On the other hand, 24 bits at 1000 Hz allow for
     a maximum duration of about 4.6 hours, while for 90 KHz, this value
     is about 3 minutes.  These values should be enough for streaming
     applications.  However, if a larger duration is needed, the
     extension mechanism specified in Section 4.3 SHALL be used.

     Apart from defining the time period during which the text is
     displayed, the duration field is also used to find the timestamp of
     subsequent units within the aggregate RTP packet payload (if any).



Rey & Matsui                Standards Track                    [Page 18]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     This is explained in Section 4.6.

     Text samples have generally a known duration at the time of
     transmission.  However, in some cases such as live streaming, the
     time for which a text piece shall be presented might not be known a
     priori.  Thus, the value zero SDUR=0 (0x000000) is reserved to
     signal unknown duration.  The amount of time that a sample of
     unknown duration is presented is determined by the timestamp of the
     next sample that shall be displayed at the receiver: Text samples
     of unknown duration SHALL be displayed until the next text sample
     becomes active, as indicated by its timestamp.

     The next example illustrates how units of unknown duration MUST be
     presented.  If no text sample following is available, it is an
     implementation issue what should be displayed.  For example, a
     server could send an empty sample to clear the text box.

        Example: Imagine you are in an airport watching the latest news
        report while you wait for your plane.  Airports are loud, so the
        news report is transcribed in the lower area of the screen.
        This area displays two lines of text: the headlines and the
        words spoken by the news speaker.  As usual, the headlines are
        shown for a longer time than the rest.  This time is, in
        principle, unknown to the stream server, which is streaming
        live.  A headline is just replaced when the next headline is
        received.

     However, upon storing a text sample with SDUR=0 in a 3GP file, the
     SDUR value MUST be changed to the effective duration of the text
     sample, which MUST be always greater than zero (note that the ISO
     file format [2] explicitly forbids a sample duration of zero).  The
     effective duration MUST be calculated as the timestamp difference
     between the current sample (with unknown duration) and the next
     text sample that is displayed.

     Note that samples of unknown duration SHALL NOT use features, which
     require knowledge of the duration of the sample up front.  Such
     features are scrolling and karaoke in [1].  This also applies for
     future extensions of the Timed Text format.  Furthermore, only
     sample descriptions (TYPE 5 units) MAY follow units of unknown
     duration in the same aggregate payload.  Otherwise, it would not be
     possible to calculate the timestamp of these other units.

     For text contents stored in 3GP files, see Section 4.3 for details
     on how to extract the duration value.  For live streaming, live
     encoders SHALL assign appropriate values and units according to [1]
     and later releases.




Rey & Matsui                Standards Track                    [Page 19]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   o TLEN (16 bits), "Text String Length", is a byte count of the text
     string.  The decoder needs the text string length in order to know
     where the modifiers in the payload start.  TLEN is not present in
     text string fragments (TYPE 2) since it can be deductively
     calculated from the LEN values of each fragment.

     The TLEN value is obtained from the text samples as contained in
     3GP files.  Refer to Section 4.3.  For live content, the TLEN MUST
     be obtained during the sampling process.

   o Finally, the actual text sample is placed after the TLEN field.  As
     defined in Section 3, a text sample consists of a string of
     characters encoded using either UTF-8 or UTF-16, followed by zero
     or more modifiers.  Note also that no BOM and no byte count are
     included in the strings carried in the payload (as opposed to text
     samples stored in 3GP files [1]).

4.1.3.  TYPE 2 Header



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE |          LEN( always >9)      | TOTAL | THIS  |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                    SDUR                       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |               SLEN            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                      Figure 5.  TYPE 2 Header Format

   This header type is used to transport either a whole text string or a
   fragment of it.  TYPE 2 units SHALL NOT contain modifiers.  In
   detail:

   o U, R, and TYPE, as defined in Section 4.1.1.

   o SIDX and SDUR, as defined in Section 4.1.2.

        Note that the U, SIDX, and SDUR fields are meaningful since
        partial text strings can also be displayed.

   o The LEN field (16 bits) indicates the length of the text string
     fragment plus nine (9) bytes of headers.  Its value is calculated
     upon fragmentation.  LEN MUST always be greater than nine (0x0009).
     Otherwise, the unit MUST be discarded.





Rey & Matsui                Standards Track                    [Page 20]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     According to the guidelines in Section 4.4, text strings MUST be
     split at character boundaries for allowing the display of text
     fragments.  Therefore, a text fragment MUST contain at least one
     character in either UTF-8 or UTF-16.  Actually, this is just a
     formalism since by observing the guidelines, much larger fragments
     should be created.

     Note also that TYPE 2 units do not contain an explicit text string
     length, TLEN (see TYPE 1).  This is because TYPE 2 units do not
     contain any modifiers after the text string.  If needed, the length
     of the received string can be obtained using the LEN values of the
     TYPE 2 units.

   o The SLEN field (16 bits) indicates the size (in bytes) of the
     original (whole) text sample to which this fragment belongs.  This
     length comprises the text string plus any modifier boxes present
     (and includes neither the byte order mark nor the text string
     length as mentioned in Section 3, "Terminology").

     Regarding the text sample length: Timed text samples are not
     generated at regular intervals, nor is there a default sample size.
     If 3GP files are streamed, the length of the text samples is
     calculated beforehand and included in the track itself, while for
     live encoding it is the real time encoder that SHALL choose an
     appropriate size for each text sample.  In this case, the amount of
     text 'captured' in a sample depends on the text source and the
     particular application (see examples below).  Samples may, e.g., be
     tailored to match the packet MTU as closely as possible or to
     provide a given redundancy for the available bit rate.  The
     encoding application MUST also take into account the delay
     constraints of the real-time session and assess whether FEC,
     retransmission, or other similar techniques are reasonable options
     for stream repair.

     The following examples shall illustrate how a real-time encoder may
     choose its settings to adapt to the scenario constraints.

          Example: Imagine a newscast scenario, where the spoken news is
          transcribed and synchronized with the image and voice of the
          reporter.  We assume that the news speaker talks at an average
          speed of 5 words per second with an average word length of 5
          characters plus one space per word, i.e., 30 characters per
          second.  We assume an available IP MTU of 576 bytes and an
          available bitrate of 576*8 bits per second = 4.6 Kbps.  We
          assume each character can be encoded using 2 bytes in UTF-16.
          In this scenario, several constraints may apply; for example:
          available IP MTU, available bandwidth, allowable delay, and
          required redundancy.  If the target were to minimize the



Rey & Matsui                Standards Track                    [Page 21]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


          packet overhead, a text sample covering 8 seconds of text
          would be closest to the IP MTU:

       IP/UDP/RTP/TYPE1 Header + (8-second text sample)
     = 20 + 8 + 12 + 8 + (~6 chars/word * 5 word/s * 8 s * 2 chars/word)
     = 528 bytes < 576 bytes

    For other scenarios, like lossy networks, it may happen that just
    one packet per sample is too low a redundancy.  In this case, a
    choice could be that the encoder 'collects' text every second, thus
    yielding text samples (TYPE 1 units) of 68 bytes, TYPE 1 header
    included.  We can, e.g., include three contiguous text samples in
    one RTP payload: the current and last two text samples (see below).
    This accounts to a total IP packet size of 20 + 8 + 12 + 3*(8 + 60)
    = 244 bytes.  Now, with the same available bitrate of 4.6 Kbps,
    these 244-byte packets can be sent redundantly up two times per
    second:

          RTP payload (1,2,3)(1,2,3) (2,3,4)(2,3,4) (3,4,5)(3,4,5) ...
          Time:       <----1s------> <----1s------> <-----1s-----> ...

          This means that each text sample is sent at least six times,
          which should provide enough redundancy.  Although not as
          bandwidth efficient (488*8 < 528*8  < 576*8 bps) as the
          previous packetization, this option increases the stream
          redundancy while still meeting the delay and bandwidth
          constraints.

          Another example would be a user sending timed text from a
          type-in area in the display.  In this case, the text sample is
          created as soon as the user clicks the 'send' button.
          Depending on the packet length, fragmentation may be needed.

          In a video conferencing application, text is synchronized with
          audio and video.  Thus, the text samples shall be displayed
          long enough to be read by a human, shall fit in the video
          screen, and shall 'capture' the audio contents rendered during
          the time the corresponding video and audio is rendered.

     For stored content, see Section 4.3 for details on how to find the
     SLEN value in a 3GP file.  For live content, the SLEN MUST be
     obtained during the sampling process.

     Finally, note that clients MAY use SLEN to buffer space for the
     remaining fragments of a text sample.

   o The fields TOTAL (4 bits) and THIS (4 bits) indicate the total
     number of fragments in which the original text sample (i.e., the



Rey & Matsui                Standards Track                    [Page 22]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     text string and its modifiers) has been fragmented and which order
     occupies the current fragment in that sequence, respectively.  Note
     that the sequence number alone cannot replace the functionality of
     the THIS field, since packets (and fragments) may be repeated,
     e.g., as in repeated transmission (see Section 5).  Thus, an
     indication for "fragment offset" is needed.

     The usual "byte offset" field is not used here for two reasons: a)
     it would take one more byte and b) it does not provide any
     information on the character offset.  UTF-8/UTF-16 text strings
     have, in general, a variable character length ranging from 1 to 6
     bytes.  Therefore, the TOTAL/THIS solution is preferred.  It could
     also be argued that the LEN and SLEN fields be used for this
     purpose, but while they would provide information about the
     completeness of the text sample, they do not specify the order of
     the fragments.

     In all cases (TYPEs 2, 3 and 4), if the value of THIS is greater
     than TOTAL or if TOTAL equals zero (0x0), the fragment SHALL be
     discarded.

   o Finally, the sample contents following the SLEN field consist of a
     fragment of the UTF-8/UTF-16 character string; no modifiers follow.

4.1.4.  TYPE 3 Header



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE |        LEN( always >6)        |TOTAL  |  THIS |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                      Figure 6.  TYPE 3 Header Format

   This header type is used to transport either the entire modifier
   contents present in a text sample or just the first fragment of them.
   This depends on whether the modifier boxes fit in the current RTP
   payload.

   If a text sample containing modifiers is fragmented, this header MUST
   be used to transport the first fragment or, if possible, the complete
   modifiers.

   In detail:

   o The U, R, and TYPE fields are defined as in Section 4.1.1.



Rey & Matsui                Standards Track                    [Page 23]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   o LEN indicates the length of the modifier contents.  Its value is
     obtained upon fragmentation.  Additionally, the LEN field MUST be
     greater than six (0x0006).  Otherwise, the unit MUST be discarded.

   o The TOTAL/THIS field has the same meaning as for TYPE 2.

     For TYPE 3 units containing the last (trailing) modifier fragment,
     the value of TOTAL MUST be equal to that of THIS (TOTAL=THIS).  In
     addition, TOTAL=THIS MUST be greater than one, because the total
     number of fragments of a text sample is logically always larger
     than one.

     Otherwise, if TOTAL is different from THIS in a TYPE 3 unit, this
     means that the unit contains the first fragment of the modifiers.

   o The SDUR has the same definition for TYPE 1.  Since the fragments
     are always transported in own RTP packets, this field is only
     needed to know how long this fragment is valid.  This may, e.g., be
     used to determine how long it should be kept in the display buffer.

   Note that the SLEN and SIDX fields are not present in TYPE 3 unit
   headers.  This is because a) these fragments do not contain text
   strings and b) these types of fragments are applied over text string
   fragments, which already contain this information.

4.1.5.  TYPE 4 Header



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE |        LEN( always >6)        |TOTAL  |  THIS |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                      Figure 7.  TYPE 4 Header Format

   This header type is placed before modifier fragments, other than the
   first one.

   The U, R, and TYPE fields are used as per Section 4.1.1.

   LEN indicates as for TYPE 3 the length of the modifier contents and
   SHALL also be obtained upon fragmentation.  The LEN field MUST be
   greater than six (0x0006).  Otherwise, the unit MUST be discarded.

   TOTAL/THIS is used as in TYPE 2.




Rey & Matsui                Standards Track                    [Page 24]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   The SDUR field is defined as in TYPE 1.  The reasoning behind the
   absence of SLEN and SIDX is the same as in TYPE 3 units.

4.1.6.  TYPE 5 Header



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE |      LEN( always >3)          |   SIDX        |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

                      Figure 8.  TYPE 5 Header Format

   This header type is used to transport (dynamic) sample descriptions.
   Every sample description MUST have its own TYPE 5 header.

   The U, R, and TYPE fields are used as per Section 4.1.1.

   The LEN field indicates the length of the sample description, plus
   three units accounting for the SIDX and LEN field itself.  Thus, this
   field MUST be greater than three (0x0003).  Otherwise, the unit MUST
   be discarded.

   If the sample is streamed from a 3GP file, the length of the sample
   description contents (i.e., what comes after SIDX in the unit itself)
   is obtained from the file (see Section 4.3).

   The SIDX field contains a dynamic SIDX value assigned to the sample
   description carried as sample content of this unit.  As only dynamic
   sample descriptions are carried using TYPE 5, the possible SIDX
   values are in the (closed) interval [0,127].

   Senders MAY make use of TYPE 5 units.  All receivers MUST implement
   support for TYPE 5 units, since it adds minimum complexity and may
   increase the robustness of the streaming session.

   The next section specifies how SIDX values are calculated.

4.2.  Buffering of Sample Descriptions



   The buffering of sample descriptions is a matter of the client's
   timed text codec implementation.  In order to work properly, this
   payload format requires that:

     o Static sample descriptions MUST be buffered at the client, at
       least, for the duration of the session.





Rey & Matsui                Standards Track                    [Page 25]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     o If dynamic sample descriptions are used, their buffering and
       update of the SIDX values MUST follow the mechanism described in
       the next section.

4.2.1.  Dynamic SIDX Wraparound Mechanism



   The use of dynamic sample descriptions by senders is OPTIONAL.
   However, if they are used, senders MUST implement this mechanism.
   Receivers MUST always implement it.

   Dynamic SIDX values remain active either during the entire duration
   of the session (if used just once) or in different intervals of it
   (if used once or more).

        Note: In the following, SIDX means dynamic SIDX.

   For choosing the wraparound mechanism, the following rationale was
   used: There are 128 dynamic SIDX values possible, [0..127].  If one
   chooses to allow a maximum of 127 to be used as dynamic SIDXs, then
   any reordered packet with a new sample description would make the
   mechanism fail.  For example, if the last packet received is SIDX=5,
   then all 127 values except SIDX=6 would be "active".  Now, if a
   reordered packet arrives with a new description, SIDX=9, it will be
   mistakenly discarded, because the SIDX=9 is, at that moment, marked
   as "active" and active sample descriptions shall not be re-written.
   Therefore, a "guard interval" is introduced.  This guard interval
   reduces the number of active SIDXs at any point in time to 64.
   Although most timed text applications will probably need less than 64
   sample descriptions during a session (in total), a wraparound
   mechanism to handle the need for more is described here.

   Thereby, a sliding window of 64 active SIDX values is used.  Values
   within the window are "active"; all others are marked "inactive".  An
   SIDX value becomes active if at least one sample description
   identified by that SIDX has been received.  Since sample descriptions
   MAY be sent redundantly, it is possible that a client receives a
   given SIDX several times.  However, active sample descriptions SHALL
   NOT
be overwritten: The receiver SHALL ignore redundant sample
   descriptions and it MUST use the already cached copy.  The "guard
   interval" of (64) inactive values ensures that the correct
   association SIDX <-> sample description is always used.

        Informative note: As for the "guard interval" value itself, 64
        as 128/2 was considered simple enough while still meeting the
        expected maximum number of sample descriptions.  Besides that,
        there's no other motivation for choosing 64 or a different
        value.




Rey & Matsui                Standards Track                    [Page 26]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   The following algorithm is used to buffer dynamic sample descriptions
   and to maintain the dynamic SIDX values:

   Let X be the last SIDX received that updated the range of active
   sample descriptions.  Let Y be a value within the allowed range for
   dynamic SIDX: [0,127], and different from X.  Let Z be the SIDX of
   the last received sample description.  Then:

     1. Initialize all dynamic SIDX values as inactive.  For stored
        contents, read the sample description index in the Sample to
        Chunk box ("stsc") for that sample.  For live streaming, the
        first value MAY be zero or any other value in the interval
        above.  Go to step 2.

     2. First, in-band sample description with SIDX=Z is received and
        stored; set X=Z.  Go to step 3.

     3. Any SIDX within the interval [X+1 modulo(128), X+64 modulo(128)]
        is marked as inactive, and any corresponding sample description
        is deleted.  Any SIDX within the interval [X+65 modulo(128), X]
        is set active.  Go to step 4 (wait state).

     4. Wait for next sample description.  Once the client is
        initialized, the interval of active SIDX values MUST change
        whenever a sample description with an SIDX value in the inactive
        set is received.  That is, upon reception of a sample
        description with SIDX=Z, do the following:

        a. If Z is in the (closed) interval [X+1 modulo(128), X+64
           modulo(128)] then set X=Z, store the sample description, and
           go to step 3.

        b. Else, Z must be in the interval [X+65 modulo(128), X], thus:

            i. If SIDX=Z is not stored, then store the sample
               description. Go to beginning of step 4 (wait state).
           ii. Else, go to the beginning of step 4 (wait state).

        Informative note: It is allowed that any value of SIDX=X be sent
        in the interval [0,127].  For example, if [64..127] is the
        current active set and SIDX=0 is sent, a new sample description
        is defined (0) and an old one deleted (64); thus [65..127] and
        [0] are active.  Similarly, one could now send SIDX=64, thus
        inverting the active and inactive sets.

   Example:
        If X=4, any SIDX in the interval [5,68] is inactive.  Active
        SIDX values are in the complementary interval [69,127] plus



Rey & Matsui                Standards Track                    [Page 27]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


        [0,4].  For example, if the client receives a SIDX=6, then the
        active interval is now different: [0,6] plus [71,127].  If the
        received SIDX is in the current active interval, no change SHALL
        be applied.

4.3.  Finding Payload Header Values in 3GP Files



   For the purpose of streaming timed text contents, some values in the
   boxes contained in a 3GP file are mapped to fields of this payload
   header.  This section explains where to find those values.

   Additionally, for the duration and sample description indexes,
   extension mechanisms are provided.  All senders MUST implement the
   extension mechanisms described herein.

   If the file is streamed out of a 3GP file, the following guidelines
   SHALL be followed.

        Note: All fields in the objects (boxes) of a 3GP file are found
        in network byte order.

   Information obtained from the Sample Table Box (stbl):

        o Sample Descriptions and Sample Description length: The Sample
          Description box (stsd, inside the stbl) contains the sample
          descriptions.  For timed text media, each element of stsd is a
          timed text sample entry (type "tx3g").

          The (unsigned) 32 bits of the "size" field in the stsd box
          represent the length (in bytes) of the sample description, as
          carried in TYPE 5 units.  On the other hand, the LEN field of
          TYPE 5 units is restricted to 16 bits.  Therefore, if the
          value of "size" is greater than (2^16-1-3)[bytes], then the
          sample description SHALL NOT be streamed with this payload
          format.  There is no extension mechanism defined in this case,
          since fragmentation of sample descriptions is not defined
          (sample descriptions are typically up to some 200 bytes in
          size).  Note: The three (3) accounts for the TYPE 5 header
          fields included in the LEN value.

        o SDUR from the Decoding Time to Sample Box (stts).  The
          (unsigned) 32 bits of the "sample delta" field are used for
          calculating SDUR.  However, since the SDUR field is only 3
          bytes long, text samples with duration values larger than
          (2^24-1)/(timestamp clockrate)[seconds] cannot be streamed
          directly.  The solution is simple: Copies of the corresponding
          text sample SHALL be sent.  Thereby, the timestamp and
          duration values SHALL be adjusted so that a continuous display



Rey & Matsui                Standards Track                    [Page 28]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


          is guaranteed as if just one sample would have been sent.
          That is, a sample with timestamp TS and duration SDUR can be
          sent as two samples having timestamps TS1 and TS2 and
          durations SDUR1 and SDUR2, such that TS1=TS, TS2=TS1+SDUR1,
          and SDUR=SDUR1+SDUR2.

        o Text sample length from the Sample Size Box (stsz).  The
          (unsigned) 32 bits of the "sample size" or "entry size" (one
          of them, depending on whether the sample size is fixed or
          variable) indicate the length (in bytes) of the 3GP text
          sample.  For obtaining the length of the (actual) streamed
          text sample, the lengths of the text string byte count (2
          bytes) and, in case of UTF-16 strings, the length the BOM
          (also 2 bytes) SHALL be deducted.  This is illustrated in
          Figure 9.

          Text Sample according to 3GPP TS 26.245

                               TEXT SAMPLE (length=stsz)
                 .--------------------------------------------------.
                /                                                    \
                               TEXT STRING  (length=TBC)
                    .------------------------------------.
                   /                                      \
                TBC BOM                                     MODIFIERS
               +---+---+----------------------------------+-----------+
                                     ||
                                     ||    TBC BOM  -> TLEN  field
                                     ||   +---+---+    U bit
                                     ||
                                     \/

          Text Sample according to this Payload Format

                                 TEXT SAMPLE (length=SLEN w/o TBC,BOM)
                        .--------------------------------------------.
                       /                                              \
                                     TEXT STRING (length=TLEN)
                        .--------------------------------.
                       /                                  \
                                    TEXT STRING             MODIFIERS
                       +----------------------------------+-----------+

              KEY:
              TBC = Text string Byte Count
              BOM = Byte Order Mark

                    Figure 9.  Text sample composition



Rey & Matsui                Standards Track                    [Page 29]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


          Moreover, since the LEN field in TYPE 1 unit header is 16 bits
          long, larger text sample sizes than (2^16-1-8) [bytes] SHALL
          NOT
be streamed.  Also, in this case, no extension mechanism
          is defined.  This is because this maximum is considered enough
          for the targeted streaming applications. (Note: The eight (8)
          accounts for the TYPE 1 header fields included in the LEN
          value).

        o SIDX from the Sample to Chunk Box (stsc): The stsc Box is used
          to find samples and their corresponding sample descriptions.
          These are referenced by the "sample description index", a
          32-bit (unsigned) integer.  If possible, these indices may be
          directly mapped to the SIDX field.  However, there are several
          cases where this may not be possible:

                  a) The total number of indices used is greater than
               the number of indices available, i.e., if the static
               sample descriptions are more than 127 or the dynamic ones
               are more than 64.

                  b) The original SIDX value ranges do not fit in the
               allowed ranges for static (129-254) or dynamic (0-127)
               values.

          Therefore, when assigning SIDX values to the sample
          descriptions, the following guidelines are provided:

          o    Static sample descriptions can simply be assigned
               consecutive values within the range 129-254 (closed
               interval).  This range should be well enough for static
               sample descriptions.

          o    As for dynamic sample descriptions:

                  a) Streams that use less than 64 dynamic sample
               descriptions SHOULD use consecutive values for SIDX
               anywhere in the range 0-127 (closed interval).

                  b) For streams with more than 64 sample descriptions,
               the SIDX values MUST be assigned in usage order, and if
               any sample description shall be used after it has been
               set inactive, it will need to be re-sent and assigned a
               new SIDX value (according to the algorithm in Section
               4.2.1).







Rey & Matsui                Standards Track                    [Page 30]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   Information obtained from the Media Data Box:

        o Text strings, TLEN, U bit, and modifiers from the Media Data
          Box (mdat).  Text strings, 16-bit text string byte count, Byte
          Order Mark (BOM, indicating UTF encoding), and modifier boxes
          can be found here.

          For TYPE 1 units, the value of TLEN is extracted from the text
          string byte count that precedes the text string in the text
          sample, as stored in the 3GP file.  If UTF-16 encoding is
          used, two (2) more bytes have to be deducted from this byte
          count beforehand, in order to exclude the BOM.  See Figure 9.

4.4.  Fragmentation of Timed Text Samples



   This section explains why text samples may have to be fragmented and
   discusses some of the possible approaches to doing it.  A solution is
   proposed together with rules and recommendations for fragmenting and
   transporting text samples.

   3GPP Timed Text applications are expected to operate at low bitrates.
   This fact, added to the small size of timed text samples (typically
   one or two hundred bytes) makes fragmentation of text samples a rare
   event.  Samples should usually fit into the MTU size of the used
   network path.

   Nevertheless, some text strings (e.g., ending roll in a movie) and
   some modifier boxes (i.e., for hyperlinks, for karaoke, or for
   styles) may become large.  This may also apply for future modifier
   boxes.  In such cases, the first option to consider is whether it is
   possible to adjust the encoding (e.g., the size of sample) in such a
   way that fragmentation is avoided.  If it is, this is preferred to
   fragmentation and SHOULD be done.

   Otherwise, if this is not possible or other constraints prevent it,
   fragmentation MAY be used, and the basic guidelines given in this
   document MUST be followed:

   o It is RECOMMENDED that text samples be fragmented as seldom as
     possible, i.e., the least possible number of fragments is created
     out of a text sample.

   o If there is some bitrate and free space in the payload available,
     sample descriptions (if at hand) SHOULD be aggregated.

   o Text strings MUST split at character boundaries; see TYPE 2 header.
     Otherwise, it is not possible to display the text contents of a
     fragment if a previous fragment was lost.  As a consequence, text



Rey & Matsui                Standards Track                    [Page 31]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     string fragmentation requires knowledge of the UTF-8/UTF-16
     encoding formats to determine character boundaries.

   o Unlike text strings, the modifier boxes are NOT REQUIRED to be
     split at meaningful boundaries.  However, it is RECOMMENDED that
     this be done whenever possible.  This decreases the effects of
     packet loss.  This payload format does not ensure that partially
     received modifiers are applied to text strings.  If only part of
     the modifiers is received, it is an application issue how to deal
     with these, i.e., whether or not to use them.

        Informative note: Ensuring that partially received modifiers can
        be applied to text strings in all cases (for all modifier types
        and for all fragment loss constellations) would place additional
        requirements on the payload format.  In particular, this would
        require that: a) senders understand the semantics of the
        modifier boxes and b) specific fragment headers for each of the
        modifier boxes are defined, in addition to the payload formats
        defined below.  Understanding the modifiers semantics means
        knowing, e.g., where each modifier starts and ends, which text
        fragments are affected, which modifiers may or may not be split,
        or what the fields indicate.  This is necessary to be able to
        split the modifiers in such a way that each fragment can be
        applied independently of previous packet losses.  This would
        require a more intelligent fragmentation entity and more complex
        headers.  Given the low probability of fragmentation and the
        desire to keep the requirements low, it does not seem reasonable
        to specify such modifier box specific headers.

   o Modifier and text string fragments SHOULD be protected against
     packet losses, i.e., using FEC [7], retransmission [11], repetition
     (Section 5), or an equivalent technique.  This minimizes the
     effects of packet loss.

   o An additional requirement when fragmenting text samples is that the
     start of the modifiers MUST be indicated using the payload header
     defined for that purpose, i.e., a TYPE 3 unit MUST be used (see
     Section 4.1.4).  This enables a receiver to detect the start of the
     modifiers as long as there are not two or more consecutive packet
     losses.

   o Finally, sample descriptions SHALL NOT be fragmented because they
     contain important information that may affect several text samples.








Rey & Matsui                Standards Track                    [Page 32]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.5.  Reassembling Text Samples at the Receiver



   The payload headers defined in this document allow reassembling
   fragmented text samples.  For this purpose, the standard RTP
   timestamp, the duration field (SDUR), and the fields TOTAL/THIS in
   the payload headers are used.

   Units that belong to the same text sample MUST have the same
   timestamp.  TYPE 5 units do not comply with this rule since they are
   not part of any particular text sample.

   The process for collecting the different fragments (units) of a text
   sample is as follows:

     1. Search for units having the same timestamp value, i.e., units
        that belong to the same text sample or sample descriptions that
        shall become available at that time instant.  If several units
        of the same sample are repeated, only one of them SHALL be used.
        Repeated units are those that have the same timestamp and the
        same values for TOTAL/THIS.

                Note that, as mentioned in Section 4.1.1, the receiver
                SHALL ignore units with unrecognized TYPE value.
                However, the RTP header fields and the rest of the units
                (if any) in the payload are still useful.

     2. Check within this set whether any of the units from the text
        sample is missing.  This is done using the TOTAL and THIS
        fields; the TOTAL field indicates how many fragments were
        created out of the text sample, and the THIS field indicates the
        position of this fragment in the text sample.  As result of this
        operation, two outcomes are possible:

          a. No fragment is missing.  Then, the THIS field SHALL be used
             to order the fragments and reassemble the text sample
             before forwarding it to the decoding application.  Special
             care SHALL be taken when reassembling the text string as
             indicated in bullet 4 below.

          b. One or more fragments are missing: Check whether this
             fragment belongs to the text string or to the modifiers.
             TYPE 2 units identify text string fragments, and TYPE 3 and
             4 identify modifier fragments:

              i. If the fragment or fragments missing belong to the text
                 string and the modifiers were received complete, then
                 the received text characters may, at least, be
                 displayed as plain text.  Some modifiers may only be



Rey & Matsui                Standards Track                    [Page 33]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


                 applied as long as it is possible to identify the
                 character numbers, e.g., if only the last text string
                 fragment is lost.  This is the case for modifiers
                 defining specific font styles ('styl'), highlighted
                 characters ('hlit'), karaoke feature ('krok'), and
                 blinking characters ('blnk').  Other modifiers such as
                 'dlay' or 'tbox' can be applied without the knowledge
                 of the character number.  It is an application issue to
                 decide whether or not to apply the modifiers.

             ii. If the fragment missing belongs to the modifiers and
                 the text strings were received complete, then the
                 incomplete modifiers may be used.  The text string
                 SHOULD at least be displayed as plain text.  As
                 mentioned in Section 4.4, modifiers may split without
                 observing meaningful boundaries.  Hence, it may not
                 always be possible to make use of partially received
                 modifiers.  However, to avoid this, it is RECOMMENDED
                 that the modifiers do split at meaningful boundaries.

            iii. A third possibility is that it is not possible to
                 discern whether modifiers or text strings were received
                 complete.  For example, if the TYPE 3 unit of a sample
                 plus the following or preceding packet is lost, there
                 is no way for the RTP receiver to know if one or both
                 packets lost belong to the modifiers or if there are
                 also some missing text strings.  Repetition, FEC,
                 retransmission, or other protection mechanisms as per
                 section 4.6 are RECOMMENDED to avoid this situation.

             iv. Finally, if it is sure that neither text strings nor
                 modifiers were received complete, then the text strings
                 and the modifiers may be rendered partially or may be
                 discarded.  This is an application choice.

     3. Sample descriptions can be directly associated with the
        reassembled text samples, via the sample description index
        (SIDX).

     4. Reassembling of text strings: Since the text strings transported
        in RTP packets MUST NOT include any byte order mark (BOM), the
        receiver MUST prepend it to the reassembled UTF-16 string before
        handling it to the timed text decoder (see Figure 9).  The value
        of the BOM is 0xFEFF because only big endian serialization of
        UTF-16 strings is supported by this payload format.






Rey & Matsui                Standards Track                    [Page 34]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.6.  On Aggregate Payloads



   Units SHOULD be aggregated to avoid overhead, whenever possible.  The
   aggregate payloads MUST comply with one of the following ordered
   configurations:

   1. Zero or more sample descriptions (TYPE 5) followed by zero or more
      whole text samples (TYPE 1 units).  At least one unit of either
      type MUST be present.

   2. Zero or more sample descriptions followed by zero or one modifier
      fragment, either TYPE 3 or TYPE 4.  At least one unit MUST be
      present.

   3. Zero or more sample descriptions, followed by zero or one text
      string fragment (TYPE 2), followed by zero or one TYPE 3 unit.  If
      a TYPE 2 unit and a TYPE 3 unit are present, then they MUST belong
      to the same text sample.  At least one unit MUST be present.

   Some observations:

   o Different aggregates than the ones listed above SHALL NOT be used.

   o Sample descriptions MUST be placed in the aggregate payload before
     the occurrence of any non-TYPE 5 units.

   o Correct reception of TYPE 5 units is important since their contents
     may be referenced by several other units in the stream.

     Receivers are unable to use text samples until their corresponding
     sample descriptions are received.  Accordingly, a sender SHOULD
     send multiple copies of a sample description to ensure reliability
     (see Section 5).  Receivers MAY use payload-specific feedback
     messages [21] to tell a sender that they have received a particular
     sample description.

   o Regarding timestamp calculation: In general, the rules for
     calculating the timestamp of units in an aggregate payload depend
     on the type of unit.  Based on the possible constellations for
     aggregate payloads, as above, we have:

           o Sample descriptions MUST receive the RTP timestamp of the
             packet in which they are included.

             Note that for TYPE 5 units, the timestamp actually does not
             represent the instant when they are played out, but instead
             the instant at which they become available for use.




Rey & Matsui                Standards Track                    [Page 35]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


           o For the first configuration: The first TYPE 1 unit receives
             the RTP timestamp.  The timestamp of any subsequent TYPE 1
             unit MUST be obtained by adding sample duration and
             timestamp, both of the preceding TYPE 1 unit.

           o For the second and third configuration, all units, TYPE 2,
             3, and 4, MUST receive the RTP timestamp.

           Refer to detailed examples on the timestamp calculation
           below.

   o As per configuration 3 above, a payload MAY contain several
     fragments of one (and only one) text sample.  If it does, then
     exactly one TYPE 2 unit followed by exactly one TYPE 3 unit is
     allowed in the same payload.  This is in line with RFC 3640 [12],
     Section 2.4, which explicitly disallows combining fragments of
     different samples in the same RTP payload.  Note that, in this
     special case, no timestamp calculation is needed.  That is, the RTP
     timestamp of both units is equal to the timestamp in the packet's
     RTP header.

   o Finally, note that the use of empty text samples allows for
     aggregating non-consecutive TYPE 1 units in the same payload.  Two
     text samples, with timestamps TS1 and TS3 and durations SDUR1 and
     SDUR3, are not consecutive if it holds TS1+SDUR1 < TS3.  A solution
     for this is to include an empty TYPE 1 unit with duration SDUR2
     between them, such that TS2+SDUR2 = TS1+SDUR1+SDUR2 = TS3.

   Some examples of aggregate payloads are illustrated in Figure 10.
   (Note: The figure is not scaled.)





















Rey & Matsui                Standards Track                    [Page 36]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


      N/A    TS1   TS2     TS3
    +------+-----+------+-----+
    |TYPE5 |TYPE1|TYPE1 |TYPE1|
    +------+-----+------+-----+
      N/A   sdur1  sdur2  sdur3

                                   N/A    TS4
                                 +-----+-------+
                                 |TYPE5| TYPE 1|                   a)
                                 +-----+-------+
                                   N/A   sdur4

                                        TS4         TS4    TS4
                                 +--------------+ +--------------+
                                 |    TYPE2     | |TYPE2 |TYPE 3 | b)
                                 +--------------+ +--------------+
                                       sdur4       sdur4   sdur4

                                        TS4             TS4
                                 +--------------+ +--------------+
                                 | TYPE2| TYPE 3| |     TYPE4    | c)
                                 +--------------+ +--------------+
                                   sdur4  sdur4        sdur4

    |----------PAYLOAD 1------|  |--PAYLOAD 2---| |--PAYLOAD 3---|
               rtpts1               rtpts2           rtpts3

        KEY:
        TSx    = Text Sample x
        rtptsy = the standard RTP timestamp for PAYLOAD y
        sdurx  = the duration of Text Sample x
        N/A    =  not applicable

                  Figure 10.  Example aggregate payloads

   In Figure 10, four text samples (TS1 through TS4) are sent using
   three RTP packets.  These configurations have been chosen to show how
   the 5 TYPE headers are used.  Additionally, three different
   possibilities for the last text sample, TS4, are depicted: a), b),
   and c).

   In Figure 11, option b) from Figure 10 is chosen to illustrate how
   the timestamp for each unit is found.








Rey & Matsui                Standards Track                    [Page 37]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


      N/A    TS1   TS2    TS3        TS4            TS4    TS4
    +------+-----+------+-----+  +--------------+ +--------------+
    |TYPE5 |TYPE1|TYPE1 |TYPE1|  |    TYPE2     | |TYPE2 |TYPE 3 |
    +------+-----+------+-----+  +--------------+ +--------------+
      N/A   sdur1 sdur2  sdur3         sdur4       sdur4   sdur4

     (#1)    (#2) (#3)   (#4)           (#5)        (#6)    (#7)

    |----------PAYLOAD 1------|  |--PAYLOAD 2---| |--PAYLOAD 3---|
               rtpts1               rtpts2           rtpts3

               Figure 11.  Selected payloads from Figure 10

   Assuming TSx means Text Sample x, rtptsy represents the standard RTP
   timestamp for PAYLOAD y and sdurx, the duration of Text Sample x, the
   timestamp for unit #z, ts(#z), can be found as the sum of rtptsy and
   the cumulative sum of the durations of preceding units in that
   payload (except in the case of PAYLOAD 3 as per rule 3 above).  Thus,
   we have:

          1. for the units in the first aggregate payload, PAYLOAD 1:

                        ts(#1) = rtpts1
                        ts(#2) = rtpts1
                        ts(#3) = rtpts1 + sdur1
                        ts(#4) = rtpts1 + sdur1 + sdur2

           Note that the TYPE 5 and the first TYPE 1 unit have both the
           RTP timestamp.

          2. for PAYLOAD 2:

                        ts(#5) = rtpts2

          3. for PAYLOAD 3:

                        ts(#6) = ts(#7) = rtpsts2 = rtpts3

           According to configuration 3 above, the TYPE2 and the TYPE 3
           units shall belong to the same sample.  Hence, rtpts3 must be
           equal to rtpts2.  For the same reason, the value of SDUR is
           not be used to calculate the timestamp of the next unit.









Rey & Matsui                Standards Track                    [Page 38]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


4.7.  Payload Examples



   Some examples of payloads using the defined headers are shown 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            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE1|       LEN  (always >=8)       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                     SDUR                      |     TLEN      |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |    TLEN       |                                               |
      +---------------+                                               |
      |                  text string (no.bytes=TLEN)                  |
      |                                                               |
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                   modifiers   (no.bytes=LEN - 8 - TLEN)       |
      |                                                               |
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE1|       LEN  (always >=8)       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                     SDUR                      |     TLEN      |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |    TLEN       |                                               |
      +---------------+                                               |
      |                  text string (no.bytes=TLEN)                  |
      |                                                               |
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                   modifiers   (no.bytes=LEN - 8 - TLEN)       |
      |                                               +-+-+-+-+-+-+-+-+
      |                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

            Figure 12.  A payload carrying two TYPE 1 units

   In Figure 12, an RTP packet carrying two TYPE 1 units is depicted.
   It can be seen how the length fields LEN and TLEN can be used to find
   the start of the next unit (LEN), the start of the modifiers (TLEN),
   and the length of the modifiers (LEN-TLEN).



Rey & Matsui                Standards Track                    [Page 39]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


       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            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE5|      LEN( always >3)          |   SIDX        |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                                                               |
      |                   sample description (no.bytes=LEN - 3)       |
      |                                                               |
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE1|       LEN  (always >=8)       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |     TLEN      |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |      TLEN     |                                               |
      +-+-+-+-+-+-+-+-+                                               |
      |                  text string fragment (no.bytes=TLEN)         |
      |                                                               |
      |                                                               |
      |                                               +-+-+-+-+-+-+-+-+
      |                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

     Figure 13.  An RTP packet carrying a TYPE 5 and a TYPE 1 unit

   In Figure 13, a sample description and a TYPE 1 unit are aggregated.
   The TYPE 1 unit happens to contain only text strings and is small, so
   an additional TYPE 5 unit is included to take advantage of the
   available bits in the packet.
















Rey & Matsui                Standards Track                    [Page 40]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


       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            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE2|          LEN( always >9)      |TOTAL=4|THIS=1 |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                    SDUR                       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |               SLEN            |                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+                               |
      |                  text string fragment (no.bytes=LEN - 9)      |
      |                                                               |
      :                                                               :
      :                                                               :
      |                                               +-+-+-+-+-+-+-+-+
      |                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

    Figure 14.  Payload with first text string fragment of a sample

   In Figures 14, 15, and 16, a text sample is split into three RTP
   packets.  In Figure 14, the text string is big and takes the whole
   packet length.  In Figure 15, the only possibility for carrying two
   fragments of the same text sample is represented (see configuration 3
   in Section 4.6).  The last packet, shown in Figure 16, carries the
   last modifier fragment, a TYPE 4.




















Rey & Matsui                Standards Track                    [Page 41]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


       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            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE2|          LEN( always >9)      |TOTAL=4|THIS=2 |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                    SDUR                       |    SIDX       |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |               SLEN            |                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+                               |
      |                  text string fragment (no.bytes=LEN - 9)      |
      |                                                               |
      |                                                               |
      |                                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE3|        LEN( always >6)        |TOTAL=4|THIS=3 |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+               |
      |                                                               |
      |                    modifiers (no.bytes=LEN - 6)               |
      |                                               +-+-+-+-+-+-+-+-+
      |                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Figure 15.  An RTP packet carrying a TYPE 2 unit and a TYPE 3 unit




















Rey & Matsui                Standards Track                    [Page 42]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


       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            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |U|   R   |TYPE4|        LEN( always >6)        |TOTAL=4|THIS=4 |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                      SDUR                     |               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+               |
      |                                                               |
      |                    modifiers (no.bytes=LEN - 6)               |
      |                                               +-+-+-+-+-+-+-+-+
      |                                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

     Figure 16.  An RTP packet carrying last modifiers fragment (TYPE 4)

4.8.  Relation to RFC 3640



   RFC 3640 [12] defines a payload format for the transport of any non-
   multiplexed MPEG-4 elementary stream.  One of the various MPEG-4
   elementary stream types is MPEG-4 timed text streams, specified in
   MPEG-4 part 17 [26], also known as ISO/IEC 14496-17.  MPEG-4 timed
   text streams are capable of carrying 3GPP timed text data, as
   specified in 3GPP TS 26.245 [1].

   MPEG-4 timed text streams are intentionally constructed so as to
   guarantee interoperability between RFC 3640 and this payload format.
   This means that the construction of the RTP packets carrying timed
   text is the same.  That is, the MPEG-4 timed text elementary stream
   as per ISO/IEC 14496-17 is identical to the (aggregate) payloads
   constructed using this payload format.

   Figure 17 illustrates the process of constructing an RTP packet
   containing timed text.  As can be seen in the partition block, the
   (transport) units used in this payload format are identical to the
   Timed Text Units (TTUs) defined in ISO/IEC 14496-17.  Likewise, the
   rules for payload aggregation as per Section 4.6 are identical to
   those defined in ISO/IEC 14496-17 and are compliant with RFC 3640.
   As a result, an RTP packet that uses this payload format is identical
   to an RTP packet using RFC 3640 conveying TTUs according to ISO/IEC
   14496-17.  In particular, MPEG-4 Part 17 specifies that when using





Rey & Matsui                Standards Track                    [Page 43]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   RFC 3640 for transporting timed text streams, the "streamType"
   parameter value is set to 0x0D, and the value of the
   "objectTypeIndication" in "config" takes the value 0x08.

                +--------------------------------------+
   Text samples | +--------------+   +--------------+  |
   as per 3GPP  | |Text Sample 1 |   |Text Sample N |  |
   TS 26245     | +--------------+   +--------------+  |
                +--------------------------------------+
                                  \/
   +-------------------------------------------------------------------+
   | Partition Text Samples into units.  TTU[i]= TYPE i units.         |
   |                                                                   |
   |[U R TYPE LEN][{TOTAL,THIS}SIDX{SDUR}{TLEN}{SLEN}][SampleContents] |
   |{..} means present if applicable, [..] means always present        |
   +-------------------------------------------------------------------+
                   \/                                \/
   +-------------------------------------------------------------------+
   |                      Aggregation (if possible)                    |
   +-------------------------------------------------------------------+
                   \/                                \/
   +-------------------------------------------------------------------+
   | RTP Entity adds and fills RTP header and Sends RTP packet, where  |
   |  RTP packets according to this Payload Format =                   |
   |  RTP packets carrying MPEG-4 Timed Text ES over RFC 3640          |
   +-------------------------------------------------------------------+

                     Figure 17.  Relation to RFC 3640

   Note: The use of RFC 3640 for transport of ISO/IEC 14496-17 data does
   not require any new SDP parameters or any new mode definition.

4.9.  Relation to RFC 2793



   RFC 2793 [22] and its revision, RFC 4103 [23], specify a protocol for
   enabling text conversation.  Typical applications of this payload
   format are text communication terminals and text conferencing tools.
   Text session contents are specified in ITU-T Recommendation T.140
   [24].  T.140 text is UTF-8 coded as specified in T.140 [24] with no
   extra framing.  The T140block contains one or more T.140 code
   elements as specified in T.140.  Code elements are control sequences
   such as "New Line", "Interrupt", "String Terminator", or "Start of
   String".  Most T.140 code elements are single ISO 10646 [25]
   characters, but some are multiple character sequences.  Each
   character is UTF-8 encoded [18] into one or more octets.






Rey & Matsui                Standards Track                    [Page 44]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   This payload format may also be used for conversational applications
   (even for instant messaging).  However, this is not its main target.
   The differentiating feature of 3GPP Timed Text media format is that
   it allows text decoration.  This is especially useful in multimedia
   presentations, karaoke, commercial banners, news tickers, clickable
   text strings, and captions.  T.140 text contents used in RFC 2793 do
   not allow the use of text decoration.

   Furthermore, the conversational text RTP payload format recommends a
   method to include redundant text from already transmitted packets in
   order to reduce the risk of text loss caused by packet loss.  Thereby
   payloads would include a redundant copy of the last payload sent.
   This payload format does not describe such a method, but this is also
   applicable here.  As explained in Section 5, packet redundancy SHOULD
   be used, whenever possible.  The aggregation guidelines in Section
   4.6 allow redundant payloads.

5.  Resilient Transport



   Apart from the basic fragmentation guidelines described in the
   section above, the simplest option for packet-loss-resilient
   transport is packet repetition.  This mechanism may consist of a
   strict window-based repetition mechanism or, simply, a repetition
   mechanism in a wider sense, where new and old packets are mixed, for
   example.

   A server MAY decide to use repetition as a measure for packet loss
   resilience.  Thereby, a server MAY send the same RTP payloads or just
   some of the units from the payloads.

   As for the case of complete payloads, single repeated units MUST
   exactly match the same units sent in the first transmission; i.e., if
   fragmentation is needed, it SHALL be performed only once for each
   text sample.  Only then, a receiver can use the already received and
   the repeated units to reconstruct the original text samples.  Since
   the RTP timestamp is used to group together the fragments of a
   sample, care must taken to preserve the timing of units when
   constructing new RTP packets.

        For example, if a text sample was originally sent as a single
        non-fragmented text sample (one TYPE 1 unit), a repetition of
        that sample MUST be sent also as a single non-fragmented text
        sample in one unit.  Likewise, if the original text sample was
        fragmented and spread over several RTP packets (say, a total of
        3 units), then the repeated fragments SHALL also have the same
        byte boundaries and use the same unit headers and bytes per
        fragment.




Rey & Matsui                Standards Track                    [Page 45]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   With repetition, repeated units resolve to the same timestamp as
   their originals.  Where redundant units are available, only one of
   them SHALL be used.

   Regarding the RTP header fields:

   o If the whole RTP payload is repeated, all payload-specific fields
     in the RTP header (the M, TS and PT fields) MUST keep their
     original values except the sequence number, which MUST be
     incremented to comply with RTP (the fields TOTAL/THIS enable to
     re-assemble fragments with different sequence numbers).

   o In packets containing single repeated units, the general rules in
     Section 3 for assigning values to the RTP header fields apply.
     Keeping the value of the RTP timestamp to preserve the timing of
     the units is particularly relevant here.

   Apart from repetition, other mechanisms such as FEC [7],
   retransmission [11], or similar techniques could be used to cope with
   packet losses.

6.  Congestion Control



   Congestion control for RTP SHALL be implemented in accordance with
   RTP [3] and the applicable RTP profile, e.g., RTP/AVP [17].

   When using this payload format, mainly two factors may affect the
   congestion control:

   o The use of (unit) aggregation may make the payload format more
     bandwidth efficient, by avoiding header overhead and thus reducing
     the used bitrate.

   o The use of resilient transport mechanisms: Although timed text
     applications typically operate at low bitrates, the increase due to
     resilient transport shall be considered for congestion control
     mechanisms.  This applies to all mechanisms but especially to less
     efficient ones like repetition.













Rey & Matsui                Standards Track                    [Page 46]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


7.  Scene Description



7.1.  Text Rendering Position and Composition



   In order to set up a timed text session, regardless of the stream
   being stored in a 3GP file or streamed live, some initial layout
   information is needed by the communicating peers.

      +-------------------------------------------+
      |      <-> tx                               |    +-------------+
      |     +-------------------------------+     |<---|Display Area |
      |  ^  |                               |     |    +-------------+
      |  :  |                               |     |
      |  :ty|                               |     |    +-------------+
      |  :  |                               |<---------|Video track  |
      |  :  |                               |     |    +-------------+
      |  :  |                               |     |
      |  :  |                               |     |
      |  :  |                               |     |
      |  v  |                               |     |
      |  -  |   x-------------------------+ |     |    +-------------+
      |h ^  |   |                         |<-----------|Text Track   |
      |e :  +---|-------------------------|-+     |    +-------------+
      |i :      | +---------------------+ |       |
      |g :      | |                     | |       |    +-------------+
      |h :      | |                     |<------------ |Text Box     |
      |t v      | +---------------------+ |       |    +-------------+
      |  -      +-------------------------+       |
      +-------------------------------------------+
                <........................>
                        w i d t h

   Figure 18.  Illustration of text rendering position and composition

   The parameters used for negotiating the position and size of the text
   track in the display area are shown in Figure 18.  These are the
   "width" and "height" of the text track, its translation values, "tx"
   and "ty", and its "layer" or proximity to the user.

   At the same time, the sender of the stream needs to know the
   receiver's capabilities.  In this case, the maximum allowable values
   for the text track height and width: "max-h" and "max-w", for the
   stream the receiver shall display.

   This layout information MUST be conveyed in a reliable form before
   the start of the session, e.g., during session announcement or in an
   Offer/Answer (O/A) exchange.  An example of a reliable transport may
   be the out-of-band channel used for SDP.  Sections 8 and 9 provide



Rey & Matsui                Standards Track                    [Page 47]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   details on the mapping of these parameters to SDP descriptions and
   their usage in O/A.

   For stored content, the layout values expressing stream properties
   MUST be obtained from the Track Header Box.  See Section 7.3.

   For live streaming, appropriate values as negotiated during session
   setup shall be used.

7.2.  SMIL Usage



   The attributes contained in the Track Header Boxes of a 3GP file only
   specify the spatial relationship of the tracks within the given 3GP
   file.

   If multiple 3GP files are sent, they require spatial synchronization.
   For example, for a text and video stream, the positions of the text
   and video tracks in Figure 18 shall be determined.  For this purpose,
   SMIL [9] MAY be used.

   SMIL assigns regions in the display to each of those files and places
   the tracks within those regions.  Generally, in SMIL, the position of
   one track (or stream) is expressed relative to another track.  This
   is different from the 3GP file, where the upper left corner is the
   reference for all translation offsets.  Hence, only if the position
   in SMIL is relative to the video track origin, then this translation
   offset has the same value as (tx, ty) in the 3GP file.

   Note also that the original track header information is used for each
   track only within its region, as assigned by SMIL.  Therefore, even
   if SMIL scene description is used, the track header information
   pieces SHOULD be sent anyway, as they represent the intrinsic media
   properties.  See 3GPP SMIL Language Profile in [27] for details.

7.3.  Finding Layout Values in a 3GP File



   In a 3GP file, within the Track Header Box (tkhd):

        o tx, ty: These values specify the translation offset of the
          (text) track relative to the upper left corner of the video
          track, if present.  They are the second but last and third but
          last values in the unity matrix; values are fixed-point 16.16
          values, restricted to be (signed) integers (i.e., the lower 16
          bits of each value shall be all zeros).  Therefore, only the
          first 16 bits are used for obtaining the value of the media
          type parameters.





Rey & Matsui                Standards Track                    [Page 48]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


        o width, height: They have the same name in the tkhd box.  All
          (unsigned) 32 bits are meaningful.

        o layer: All (signed) 16 bits are used.

8.  3GPP Timed Text Media Type



   The media subtype for the 3GPP Timed Text codec is allocated from the
   standards tree.  The top-level media type under which this payload
   format is registered is 'video'.  This registration is done using the
   template defined in [29] and following RFC 3555 [28].

   The receiver MUST ignore any unrecognized parameter.

   Media type: video

   Media subtype: 3gpp-tt

   Required parameters

        rate:
                Refer to Section 3 in RFC 4396.

        sver:
                The parameter "sver" contains a list of supported
                backwards-compatible versions of the timed text format
                specification (3GPP TS 26.245) that the sender accepts
                to receive (and that are the same that it would be
                willing to send).  The first value is the value
                preferred to receive (or preferred to send).  The first
                value MAY be followed by a comma-separated list of
                versions that SHOULD be used as alternatives.  The order
                is meaningful, being first the most preferred and last
                the least preferred.  Each entry has the format
                Zi(xi*256+yi), where "Zi" is the number of the Release
                and "xi" and "yi" are taken from the 3GPP specification
                version (i.e., vZi.xi.yi).  For example, for 3GPP TS
                26.245 v6.0.0, Zi(xi*256+yi)=6(0), the version value is
                "60".  (Note that "60" is the concatenation of the
                values Zi=6 and (xi*256+yi)=0 and not their product.)

                If no "sver" value is available, for example, when
                streaming out of a 3GP file, the default value "60",
                corresponding to the 3GPP Release 6 version of 3GPP TS
                26.245, SHALL be used.






Rey & Matsui                Standards Track                    [Page 49]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   Optional parameters:

        tx:
                This parameter indicates the horizontal translation
                offset in pixels of the text track with respect to the
                origin of the video track.  This value is the decimal
                representation of a 16-bit signed integer.  Refer to TS
                3GPP 26.245 for an illustration of this parameter.

        ty:
                This parameter indicates the vertical translation offset
                in pixels of the text track with respect to the origin
                of the video track.  This value is the decimal
                representation of a 16-bit signed integer.  Refer to TS
                3GPP 26.245 for an illustration of this parameter.

        layer:
                This parameter indicates the proximity of the text track
                to the viewer.  More negative values mean closer to the
                viewer.  This parameter has no units.  This value is the
                decimal representation of a 16-bit signed integer.

        tx3g:
                This parameter MUST be used for conveying sample
                descriptions out-of-band.  It contains a comma-separated
                list of base64-encoded entries.  The entries of this
                list MAY follow any particular order and the list SHALL
                NOT
be empty.  Each entry is the result of running
                base64 encoding over the concatenation of the (static)
                SIDX value as an 8-bit unsigned integer and the (static)
                sample description for that SIDX, in that order.  The
                format of a sample description entry can be found in
                3GPP TS 26.245 Release 6 and later releases.  All
                servers and clients MUST understand this parameter and
                MUST be capable of using the sample description(s)
                contained in it.  Please refer to RFC 3548 [6] for
                details on the base64 encoding.

        width:
                This parameter indicates the width in pixels of the text
                track or area of the text being sent.  This value is the
                decimal representation of a 32-bit unsigned integer.
                Refer to TS 3GPP 26.245 for an illustration of this
                parameter.







Rey & Matsui                Standards Track                    [Page 50]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


        height:
                This parameter indicates the height in pixels of the
                text track being sent.  This value is the decimal
                representation of a 32-bit unsigned integer.  Refer to
                TS 3GPP 26.245 for an illustration of this parameter.

        max-w:
                This parameter indicates display capabilities.  This is
                the maximum "width" value that the sender of this
                parameter supports.  This value is the decimal
                representation of a 32-bit unsigned integer.

        max-h:
                This parameter indicates display capabilities.  This is
                the maximum "height" value that the sender of this
                parameter supports.  This value is the decimal
                representation of a 32-bit unsigned integer.

   Encoding considerations:

        This media type is framed (see Section 4.8 in [29]) and
        partially contains binary data.

   Restrictions on usage:

        This media type depends on RTP framing, and hence is only
        defined for transfer via RTP [3].  Transport within other
        framing protocols is not defined at this time.

   Security considerations:

        Please refer to Section 11 of RFC 4396.

   Interoperability considerations:

        The 3GPP Timed Text media format and its file storage is
        specified in Release 6 of 3GPP TS 26.245, "Transparent end-to-
        end packet switched streaming service (PSS); Timed Text Format
        (Release 6)".  Note also that 3GPP may in future releases
        specify extensions or updates to the timed text media format in
        a backwards-compatible way, e.g., new modifier boxes or
        extensions to the sample descriptions.  The payload format
        defined in RFC 4396 allows for such extensions.  For future 3GPP
        Releases of the Timed Text Format, the parameter "sver" is used
        to identify the exact specification used.






Rey & Matsui                Standards Track                    [Page 51]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


        The defined storage format for 3GPP Timed Text format is the
        3GPP File Format (3GP) [30]. 3GP files may be transferred using
        the media type video/3gpp as registered by RFC 3839 [31].  The
        3GPP File Format is a container file that may contain, e.g.,
        audio and video that may be synchronized with the 3GPP Timed
        Text.

   Published specification: RFC 4396

   Applications which use this media type:

        Multimedia streaming applications.

   Additional information:

        The 3GPP Timed Text media format is specified in 3GPP TS 26.245,
        "Transparent end-to-end packet switched streaming service (PSS);
        Timed Text Format (Release 6)".  This document and future
        extensions to the 3GPP Timed Text format are publicly available
        at http://www.3gpp.org.

        Magic number(s): None.

        File extension(s): None.

        Macintosh File Type Code(s): None.

   Person & email address to contact for further information:

        Jose Rey, jose.rey@eu.panasonic.com
        Yoshinori Matsui, matsui.yoshinori@jp.panasonic.com
        Audio/Video Transport Working Group.

   Intended usage: COMMON

   Authors:
        Jose Rey
        Yoshinori Matsui

   Change controller: IETF Audio/Video Transport Working Group delegated
        from the IESG.










Rey & Matsui                Standards Track                    [Page 52]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


9.  SDP Usage



9.1.  Mapping to SDP



   The information carried in the media type specification has a
   specific mapping to fields in SDP [4].  If SDP is used to specify
   sessions using this payload format, the mapping is done as follows:

   o The media type ("video") goes in the SDP "m=" as the media name.

       m=video <port number> RTP/<RTP profile> <dynamic payload type>

   o The media subtype ("3gpp-tt") and the timestamp clockrate "rate"
     (the RECOMMENDED 1000 Hz or other value) go in SDP "a=rtpmap" line
     as the encoding name and rate, respectively:

       a=rtpmap:<payload type> 3gpp-tt/1000

   o The REQUIRED parameter "sver" goes in the SDP "a=fmtp" attribute by
     copying it directly from the media type string as a semicolon-
     separated parameter=value pair.

   o The OPTIONAL parameters "tx", "ty", "layer", "tx3g", "width",
     "height", "max-w" and "max-h" go in the SDP "a=fmtp" attribute by
     copying them directly from the media type string as a semicolon
     separated list of parameter=value(s) pairs:

       a=fmtp:<dynamic payload type> <parameter
       name>=<value>[,<value>][; <parameter name>=<value>]

   o   Any parameter unknown to the device that uses the SDP SHALL be
       ignored.  For example, parameters added to the media format in
       later specifications MAY be copied into the SDP and SHALL be
       ignored by receivers that do not understand them.

9.2.  Parameter Usage in the SDP Offer/Answer Model



   In this section, the meaning of the SDP parameters defined in this
   document within the Offer/Answer [13] context is explained.

   In unicast, sender and receiver typically negotiate the streams,
   i.e., which codecs and parameter values are used in the session.
   This is also possible in multicast to a lesser extent.

   Additionally, the meaning of the parameters MAY vary depending on
   which direction is used.  In the following sections, a
   "<directionality> offer" means an offer that contains a stream set to
   <directionality>.  <directionality> may take the values sendrecv,



Rey & Matsui                Standards Track                    [Page 53]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   sendonly, and recvonly.  Similar considerations apply for answers.
   For example, an answer to a sendonly offer is a recvonly answer.

9.2.1. Unicast Usage



   The following types of parameters are used in this payload format:

     1. Declarative parameters: Offerer and answerer declare the values
        they will use for the incoming (sendrecv/recvonly) or outgoing
        (sendonly) stream.  Offerer and answerer MAY use different
        values.

          a. "tx", "ty", and "layer": These are parameters describing
             where the received text track is placed.  Depending on the
             directionality:

              i. They MUST appear in all sendrecv offers and answers and
                 in all recvonly offers and answers (thus applying to
                 the incoming stream).  In the case of sendrecv offers
                 and answers and in recvonly offers, these values SHOULD
                 be used by the sender of the stream unless it has a
                 particular preference, in which case, it MUST make sure
                 that these different values do not corrupt the
                 presentation.  For recvonly answers, the answerer MAY
                 accept the proposed values for the incoming stream (in
                 a sendonly offer; see ii. below) or respond with
                 different ones.  The offerer MUST use the returned
                 values.

             ii. They MAY appear in sendonly offers and MUST appear in
                 sendonly answers.  In sendonly offers, they specify the
                 values that the offerer proposes for sending (see
                 example in Section 9.3).  In sendonly answers, these
                 values SHOULD be copied from the corresponding recvonly
                 offer upon accepting the stream, unless a particular
                 preference by the receiver of the stream exists, as
                 explained in the previous point.

     2. Parameters describing the display capabilities, "max-h" and
        "max-w", which indicate the maximum dimensions of the text track
        (text display area) for the incoming stream "tx" and "ty" values
        (see Figure 18).  "max-h" and "max-w" MUST be included in all
        offers and answers where "tx" and "ty" refer to the incoming
        stream, thus excluding sendonly offers and answers (see example
        in Section 9.3), where they SHALL NOT be present.






Rey & Matsui                Standards Track                    [Page 54]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     3. Parameters describing the sent stream properties, i.e., the
        sender of the stream decides upon the values of these:

          a. "width" and "height" specify the text track dimensions.
             They SHALL ALWAYS be present in sendrecv and sendonly
             offers and answers.  For recvonly answers, the answerer
             MUST include the offered parameter values (if any) verbatim
             in the answer upon accepting the stream.

          b. "tx3g" contains static sample descriptions.  It MAY only be
             present in sendrecv and sendonly offers and answers.  This
             parameter applies to the stream that offerers or answerers
             send.

     4. Negotiable parameters, which MUST be agreed on.  This is the
        case of "sver".  This parameter MUST be present in every offer
        and answer.  The answerer SHALL choose one supported value from
        the offerer's list, or else it MUST remove the stream or reject
        the session.

     5. Symmetric parameters: "rate", timestamp clockrate, belongs to
        this class.  Symmetric parameters MUST be echoed verbatim in the
        answer.  Otherwise, the stream MUST be removed or the session
        rejected.

   The following table summarizes all options:

























Rey & Matsui                Standards Track                    [Page 55]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     +..---------------------------+----------+----------+----------+
     |   ``--..__  Directionality/ | sendrecv | recvonly | sendonly |
     + Type of   ``--..__   O or A +----------+----------+----------+
     |    Parameter      ``--..__  |   O/A    |   O/A    |   O/A    |
     +--------------+------------``+----------+----------+----------+
     | Declarative  |tx, ty, layer |   M/M    |   M/M    |   m/M    |
     |              |              |          |          |          |
     +--------------+--------------+----------+----------+----------+
     | Display      |max-h, max-w  |   M/M    |   M/M    |   -/-    |
     | Capabilities |              |          |          |          |
     +--------------+--------------+----------+----------+----------+
     | Stream       |height, width |   M/M    |   -/(M)  |   M/M    |
     | properties   |tx3g          |   m/m    |   -/-    |   m/m    |
     |              |              |          |          |          |
     +--------------+--------------+----------+----------+----------+
     |  Negotiable  |sver          |   M/M    |   M/M    |   M/M    |
     |              |              |          |          |          |
     +--------------+--------------+----------+----------+----------+
     |  Symmetric   |rate          |   M/M    |   M/M    |   M/M    |
     +--------------+--------------+----------+----------+----------+

          Table 1.  Parameter usage in Unicast Offer / Answer.

   KEY:
        o M means MUST be present.
        o m means MAY be present (such as proposed values).
        o (M) or (m) means MUST or MAY, if applicable.
        o a hyphen ("-") means the parameter MUST NOT be present.

   Other observations regarding parameter usage:

     o Translation and transparency values: In sendonly offers, "tx",
       "ty", and "layer" indicate proposed values.  This is useful for
       visually composed sessions where the different streams occupy
       different parts of the display, e.g., a video stream and the
       captions.  These are just suggested values; the peer rendering
       the text ultimately decides where to place the text track.

     o Text track (area) dimensions, "height" and "width": In the case
       of sendonly offers, an answerer accepting the offer MUST be
       prepared to render the stream using these values.  If any of
       these conditions are not met, the stream MUST be removed or the
       session rejected.

     o Display capabilities, "max-h" and "max-w": An answerer sending a
       stream SHALL ensure that the "height" and "width" values in the
       answer are compatible with the offerer's signaled capabilities.




Rey & Matsui                Standards Track                    [Page 56]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     o Version handling via "sver": The idea is that offerer and
       answerer communicate using the same version.  This is achieved by
       letting the answerer choose from a list of supported versions,
       "sver".  For recvonly streams, the first value in the list is the
       preferred version to receive.  Consequently, for sendonly (and
       sendrecv) streams, the first value is the one preferred for
       sending (and receiving).  The answerer MUST choose one value and
       return it in the answer.  Upon receiving the answer, the offerer
       SHALL be prepared to send (sendonly and sendrecv) and receive
       (recvonly and sendrecv) a stream using that version.  If none of
       the versions in the list is supported, the stream MUST be removed
       or the session rejected.  Note that, if alternative non-
       compatible versions are offered, then this SHALL be done using
       different payload types.

9.2.2.  Multicast Usage



   In multicast, the parameter usage is similar to the unicast case,
   except as follows:

   o the parameters "tx", "ty", and "layer" in multicast offers only
     have meaning for sendrecv and recvonly streams.  In order for all
     clients to have the same vision of the session, they MUST be used
     symmetrically.

   o for "height", "width", and "tx3g" (for sendrecv and sendonly),
     multicast offers specify which values of these parameters the
     participants MUST use for sending.  Thus, if the stream is
     accepted, the answerer MUST also include them verbatim in the
     answer (also "tx3g", if present).

   o The capability parameters, "max-h" and "max-w", SHALL NOT be used
     in multicast.  If the offered text track should change in size, a
     new offer SHALL be used instead.

   o Regarding version handling:

     In the case of multicast offers, an answerer MAY accept a multicast
     offer as long as one of the versions listed in the "sver" is
     supported.  Therefore, if the stream is accepted, the answerer MUST
     choose its preferred version, but, unlike in unicast, the offerer
     SHALL NOT change the offered stream to this chosen version because
     there may be other session participants that do support the newer
     extensions.  Consequently, different session participants may end
     up using different backwards-compatible media format versions.  It
     is RECOMMENDED that the multicast offer contains a limited number
     of versions, in order for all participants to have the same view of
     the session.  This is a responsibility of the session creator.  If



Rey & Matsui                Standards Track                    [Page 57]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


     none of the offered versions is supported, the stream SHALL be
     removed or the session rejected.  Also in this case, if alternative
     non-compatible versions are offered, then this SHALL be done using
     different payload types.

9.3.  Offer/Answer Examples



   In these unicast O/A examples, the long lines are wrapped around.
   Static sample descriptions are shortened for clarity.

   For sendrecv:

   O -> A

   m=video <port> RTP/AVP 98
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=100; layer=0; height=80; width=100; max-h=120;
   max-w=160; sver=6256,60; tx3g=81...
   a=sendrecv

   A -> O

   m=video <port> RTP/AVP 98..
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=95; layer=0; height=90; width=100; max-h=100;
   max-w=160; sver=60; tx3g=82...
   a=sendrecv

   In this example, the offerer is telling the answerer where it will
   place the received stream and what is the maximum height and width
   allowable for the stream that it will receive.  Also, it tells the
   answerer the dimensions of the text track for the stream sent and
   which sample description it shall use.  It offers two versions, 6256
   and 60.  The answerer responds with an equivalent set of parameters
   for the stream it receives.  In this case, the answerer's "max-h" and
   "max-w" are compatible with the offerer's "height" and "width".
   Otherwise, the answerer would have to remove this stream, and the
   offerer would have to issue a new offer taking the answerer's
   capabilities into account.  This is possible only if multiple payload
   types are present in the initial offer so that at least one of them
   matches the answerer's capabilities as expressed by "max-h" and
   "max-w" in the negative answer.  Note also that the answerer's text
   box dimensions fit within the maximum values signaled in the offer.
   Finally, the answerer chooses to use version 60 of the timed text
   format.






Rey & Matsui                Standards Track                    [Page 58]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   For recvonly:

   Offerer -> Answerer

   m=video <port> RTP/AVP 98
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=100; layer=0; max-h=120; max-w=160; sver=6256,60
   a=recvonly

   A -> O

   m=video <port> RTP/AVP 98..
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=100; layer=0; height=90; width=100; sver=60;
   tx3g=82...
   a=sendonly

   In this case, the offer is different from the previous case: It does
   not include the stream properties "height", "width", and "tx3g".  The
   answerer copies the "tx", "ty", and "layer" values, thus
   acknowledging these.  "max-h" and "max-w" are not present in the
   answer because the "tx" and "ty" (and "layer") in this special case
   do not apply to the received stream, but to the sent stream.  Also,
   if offerer and answerer had very different display sizes, it would
   not be possible to express the answerer's capabilities.  In the
   example above and for an answerer with a 50x50 display, the
   translation values are already out of range.

   For sendonly:

   O -> A

   m=video <port> RTP/AVP 98
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=100; layer=0; height=80; width=100;
   sver=6256,60; tx3g=81...
   a=sendonly

   A -> O

   m=video <port> RTP/AVP 98..
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=100; ty=100; layer=0; height=80; width=100; max-h=100;
   max-w=160; sver=60
   a=recvonly






Rey & Matsui                Standards Track                    [Page 59]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   Note that "max-h" and "max-w" are not present in the offer.  Also,
   with this answer, the answerer would accept the offer as is (thus
   echoing "tx", "ty", "height", "width", and "layer") and additionally
   inform the offerer about its capabilities: "max-h" and "max-w".

   Another possible answer for this case would be:

   A -> O

   m=video <port> RTP/AVP 98..
   a=rtpmap:98 3gpp-tt/1000
   a=fmtp:98 tx=120; ty=105; layer=0; max-h=95; max-w=150; sver=60
   a=recvonly

   In this case, the answerer does not accept the values offered.  The
   offerer MUST use these values or else remove the stream.

9.4.  Parameter Usage outside of Offer/Answer



   SDP may also be employed outside of the Offer/Answer context, for
   instance for multimedia sessions that are announced through the
   Session Announcement Protocol (SAP) [14] or streamed through the Real
   Time Streaming Protocol (RTSP) [15].

   In this case, the receiver of a session description is required to
   support the parameters and given values for the streams, or else it
   MUST reject the session.  It is the responsibility of the sender (or
   creator) of the session descriptions to define the session parameters
   so that the probability of unsuccessful session setup is minimized.
   This is out of the scope of this document.

10.  IANA Considerations



   IANA has registered the media subtype name "3gpp-tt" for the media
   type "video" as specified in Section 8 of this document.

11.  Security Considerations



   RTP packets using the payload format defined in this specification
   are subject to the security considerations discussed in the RTP
   specification [3] and any applicable RTP profile, e.g., AVP [17].

   In particular, an attacker may invalidate the current set of active
   sample descriptions at the client by means of repeating a packet with
   an old sample description, i.e., replay attack.  This would mean that
   the display of the text would be corrupted, if displayed at all.
   Another form of attack may consist of sending redundant fragments,
   whose boundaries do not match the exact boundaries of the originals



Rey & Matsui                Standards Track                    [Page 60]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   (as indicated by LEN) or fragments that carry different sample
   lengths (SLEN).  This may cause a decoder to crash.

   These types of attack may easily be avoided by using source
   authentication and integrity protection.

   Additionally, peers in a timed text session may desire to retain
   privacy in their communication, i.e., confidentiality.

   This payload format does not provide any mechanisms for achieving
   these.  Confidentiality, integrity protection, and authentication
   have to be solved by a mechanism external to this payload format,
   e.g., SRTP [10].

12.  References



12.1.  Normative References



   [1]  Transparent end-to-end packet switched streaming service (PSS);
        Timed Text Format (Release 6), TS 26.245 v 6.0.0, June 2004.

   [2]  ISO/IEC 14496-12:2004 Information technology - Coding of audio-
        visual objects - Part 12: ISO base media file format.

   [3]  Schulzrinne, H.,  Casner, S., Frederick, R., and V. Jacobson,
        "RTP: A Transport Protocol for Real-Time Applications", STD 64,
        RFC 3550, July 2003.

   [4]  Handley, M. and V. Jacobson, "SDP: Session Description
        Protocol", RFC 2327, April 1998.

   [5]  Bradner, S., "Key words for use in RFCs to Indicate Requirement
        Levels", BCP 14, RFC 2119, March 1997.

   [6]  Josefsson, S., "The Base16, Base32, and Base64 Data Encodings",
        RFC 3548, July 2003.

12.2.  Informative References



   [7]  Rosenberg, J. and H. Schulzrinne, "An RTP Payload Format for
        Generic Forward Error Correction", RFC 2733, December 1999.

   [8]  Perkins, C. and O. Hodson, "Options for Repair of Streaming
        Media", RFC 2354, June 1998.

   [9]  W3C, "Synchronised Multimedia Integration Language (SMIL 2.0)",
        August, 2001.




Rey & Matsui                Standards Track                    [Page 61]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   [10] Baugher, M., McGrew, D., Naslund, M., Carrara, E., and K.
        Norrman, "The Secure Real-time Transport Protocol (SRTP)", RFC
        3711, March 2004.

   [11] Rey, J., Leon, D., Miyazaki, A., Varsa, V., and R. Hakenberg,
        "RTP Retransmission Payload Format", Work in Progress, September
        2005.

   [12] van der Meer, J., Mackie, D., Swaminathan, V., Singer, D., and
        P. Gentric, "RTP Payload Format for Transport of MPEG-4
        Elementary Streams", RFC 3640, November 2003.

   [13] Rosenberg, J. and H. Schulzrinne, "An Offer/Answer Model with
        Session Description Protocol (SDP)", RFC 3264, June 2002.

   [14] Handley, M., Perkins, C., and E. Whelan, "Session Announcement
        Protocol", RFC 2974, October 2000.

   [15] Schulzrinne, H., Rao, A., and R. Lanphier, "Real Time Streaming
        Protocol (RTSP)", RFC 2326, April 1998.

   [16] Transparent end-to-end packet switched streaming service (PSS);
        Protocols and codecs (Release 6), TS 26.234 v 6.1.0, September
        2004.

   [17] Schulzrinne, H. and S. Casner, "RTP Profile for Audio and Video
        Conferences with Minimal Control", STD 65, RFC 3551, July 2003.

   [18] Yergeau, F., "UTF-8, a transformation format of ISO 10646", STD
        63, RFC 3629, November 2003.

   [19] Hoffman, P. and F. Yergeau, "UTF-16, an encoding of ISO 10646",
        RFC 2781, February 2000.

   [20] Friedman, T., Caceres, R., and A. Clark, "RTP Control Protocol
        Extended Reports (RTCP XR)", RFC 3611, November 2003.

   [21] Ott, J., Wenger, S., Sato, N., Burmeister, C., and J. Rey,
        "Extended RTP Profile for RTCP-based Feedback (RTP/AVPF)", Work
        in Progress, August 2004.

   [22] Hellstrom, G., "RTP Payload for Text Conversation", RFC 2793,
        May 2000.

   [23] Hellstrom, G. and P. Jones, "RTP Payload for Text Conversation",
        RFC 4103, June 2005.





Rey & Matsui                Standards Track                    [Page 62]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   [24] ITU-T Recommendation T.140 (1998) - Text conversation protocol
        for multimedia application, with amendment 1, (2000).

   [25] ISO/IEC 10646-1: (1993), Universal Multiple Octet Coded
        Character Set.

   [26] ISO/IEC FCD 14496-17 Information technology - Coding of audio-
        visual objects - Part 17: Streaming text format, Work in
        progress, June 2004.

   [27] Transparent end-to-end Packet-switched Streaming Service (PSS);
        3GPP SMIL language profile, (Release 6), TS 26.246 v 6.0.0, June
        2004.

   [28] Casner, S. and P. Hoschka, "MIME Type Registration of RTP
        Payload Formats", RFC 3555, July 2003.

   [29] Freed, N. and J. Klensin, "Media Type Specifications and
        Registration Procedures", BCP 13, RFC 4288, December 2005.

   [30] Transparent end-to-end packet switched streaming service (PSS);
        3GPP file format (3GP) (Release 6), TS 26.244 V6.3. March 2005.

   [31] Castagno, R. and D. Singer, "MIME Type Registrations for 3rd
        Generation Partnership Project (3GPP) Multimedia files", RFC
        3839, July 2004.

























Rey & Matsui                Standards Track                    [Page 63]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


13.  Basics of the 3GP File Structure



   This section provides a coarse overview of the 3GP file structure,
   which follows the ISO Base Media file Format [2].

   Each 3GP file consists of "Boxes".  In general, a 3GP file contains
   the File Type Box (ftyp), the Movie Box (moov), and the Media Data
   Box (mdat).  The File Type Box identifies the type and properties of
   the 3GP file itself.  The Movie Box and the Media Data Box, serving
   as containers, include their own boxes for each media.  Boxes start
   with a header, which indicates both size and type (these fields are
   called, namely, "size" and "type").  Additionally, each box type may
   include a number of boxes.

   In the following, only those boxes are mentioned that are useful for
   the purposes of this payload format.

   The Movie Box (moov) contains one or more Track Boxes (trak), which
   include information about each track.  A Track Box contains, among
   others, the Track Header Box (tkhd), the Media Header Box (mdhd), and
   the Media Information Box (minf).

   The Track Header Box specifies the characteristics of a single track,
   where a track is, in this case, the streamed text during a session.
   Exactly one Track Header Box is present for a track.  It contains
   information about the track, such as the spatial layout (width and
   height), the video transformation matrix, and the layer number.
   Since these pieces of information are essential and static (i.e.,
   constant) for the duration of the session, they must be sent prior to
   the transmission of any text samples.

   The Media Header Box contains the "timescale" or number of time units
   that pass in one second, i.e., cycles per second or Hertz.  The Media
   Information Box includes the Sample Table Box (stbl), which contains
   all the time and data indexing of the media samples in a track. Using
   this box, it is possible to locate samples in time and to determine
   their type, size, container, and offset into that container. Inside
   the Sample Table Box, we can find the Sample Description Box (stsd,
   for finding sample descriptions), the Decoding Time to Sample Box
   (stts, for finding sample duration), the Sample Size Box (stsz), and
   the Sample to Chunk Box (stsc, for finding the sample description
   index).

   Finally, the Media Data Box contains the media data itself.  In timed
   text tracks, this box contains text samples.  Its equivalent to audio
   and video is audio and video frames, respectively.  The text sample
   consists of the text length, the text string, and one or several
   Modifier Boxes.  The text length is the size of the text in bytes.



Rey & Matsui                Standards Track                    [Page 64]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


   The text string is plain text to render.  The Modifier Box is
   information to render in addition to the text, such as color, font,
   etc.

14.  Acknowledgements



   The authors would like to thank Dave Singer, Jan van der Meer, Magnus
   Westerlund, and Colin Perkins for their comments and suggestions
   about this document.

   The authors would also like to thank Markus Gebhard for the free and
   publicly available JavE ASCII Editor (used for the ASCII drawings in
   this document) and Henrik Levkowetz for the Idnits web service.

Authors' Addresses



   Jose Rey
   Panasonic R&D Center Germany GmbH
   Monzastr. 4c
   D-63225 Langen, Germany

   EMail: jose.rey@eu.panasonic.com
   Phone: +49-6103-766-134
   Fax:   +49-6103-766-166


   Yoshinori Matsui
   Matsushita Electric Industrial Co., LTD.
   1006 Kadoma
   Kadoma-shi, Osaka, Japan

   EMail: matsui.yoshinori@jp.panasonic.com
   Phone: +81 6 6900 9689
   Fax:   +81 6 6900 9699

















Rey & Matsui                Standards Track                    [Page 65]

RFC 4396          Payload Format for 3GPP Timed Text       February 2006


Full Copyright Statement



   Copyright (C) The Internet Society (2006).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET
   ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED,
   INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE
   INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
   WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Intellectual Property



   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.  The IETF invites any interested party to
   bring to its attention any copyrights, patents or patent
   applications, or other proprietary rights that may cover technology
   that may be required to implement this standard.  Please address the
   information to the IETF at ietf-ipr@ietf.org.

Acknowledgement



   Funding for the RFC Editor function is provided by the IETF
   Administrative Support Activity (IASA).









Rey & Matsui                Standards Track                    [Page 66]