RFC 4749 PROPOSED STANDARD MEDIA · RTP · VOICE

G.729.1 RTP payload: RTP Payload Format for the G.729.1 Audio Codec

How to put G.729.1 compressed speech into RTP packets, including a way to ask the other end to slow down.

On rfc-editor.org ↗
PUBLISHED
October 2006
AUTHORS
A. Sollaud
STREAM
IETF
UPDATED BY
RFC 5459 — RTP G.729.1 DTX
CODEC BIT RATES
8 to 32 kbps in twelve layers
FRAME DURATION
20 ms
RTP CLOCK RATE
16 kHz, timestamp step of 320 per frame
PAYLOAD HEADER
1 octet: 4-bit MBS and 4-bit FT
MEDIA TYPE
audio/G7291
DOI
10.17487/RFC4749

RFC text

Contents
Original text · RFC Editor View on rfc-editor.org ↗
Network Working Group                                         A. Sollaud
Request for Comments: 4749                                France Telecom
Category: Standards Track                                   October 2006


             RTP Payload Format for the G.729.1 Audio Codec

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 a Real-time Transport Protocol (RTP) payload
   format to be used for the International Telecommunication Union
   (ITU-T) G.729.1 audio codec.  A media type registration is included
   for this payload format.

Table of Contents

   1. Introduction ....................................................2
   2. Background ......................................................2
   3. Embedded Bit Rates Considerations ...............................3
   4. RTP Header Usage ................................................3
   5. Payload Format ..................................................4
      5.1. Payload Structure ..........................................4
      5.2. Payload Header: MBS Field ..................................4
      5.3. Payload Header: FT Field ...................................6
      5.4. Audio Data .................................................6
   6. Payload Format Parameters .......................................7
      6.1. Media Type Registration ....................................7
      6.2. Mapping to SDP Parameters ..................................8
           6.2.1. Offer-Answer Model Considerations ...................9
           6.2.2. Declarative SDP Considerations .....................11
   7. Congestion Control .............................................11
   8. Security Considerations ........................................11
   9. IANA Considerations ............................................12
   10. References ....................................................12
      10.1. Normative References .....................................12
      10.2. Informative References ...................................12

1.  Introduction

   The International Telecommunication Union (ITU-T) recommendation
   G.729.1 [1] is a scalable and wideband extension of the
   recommendation G.729 [9] audio codec.  This document specifies the
   payload format for packetization of G.729.1 encoded audio signals
   into the Real-time Transport Protocol (RTP).

   The payload format itself is described in Section 5.  A media type
   registration and the details for the use of G.729.1 with SDP are
   given in Section 6.

   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 [2].

2.  Background

   G.729.1 is an 8-32 kbps scalable wideband (50-7000 Hz) speech and
   audio coding algorithm interoperable with G.729, G.729 Annex A, and
   G.729 Annex B.  It provides a standardized solution for packetized
   voice applications that allows a smooth transition from narrowband to
   wideband telephony.

   The most important services addressed are IP telephony and
   videoconferencing, either for enterprise corporate networks or for
   mass market (like Public Switched Telephone Network (PSTN) emulation
   over DSL or wireless access).  Target devices can be IP phones or
   other VoIP handsets, home gateways, media gateways, IP Private Branch
   Exchange (IPBX), trunking equipment, voice messaging servers, etc.

   For all those applications, the scalability feature allows tuning the
   bit rate versus quality trade-off, possibly in a dynamic way during a
   session, taking into account service requirements and network
   transport constraints.

   The G.729.1 coder produces an embedded bitstream structured in 12
   layers corresponding to 12 available bit rates between 8 and 32 kbps.
   The first layer, at 8 kbps, is called the core layer and is bitstream
   compatible with the ITU-T G.729/G.729A coder.  At 12 kbps, a second
   layer improves the narrowband quality.  Upper layers provide wideband
   audio (50-7000 Hz) between 14 and 32 kbps, with a 2 kbps granularity
   allowing graceful quality improvements.  Only the core layer is
   mandatory to decode understandable speech; upper layers provide
   quality enhancement and wideband enlargement.

   The codec operates on 20-ms frames, and the default sampling rate is
   16 kHz.  Input and output at 8 kHz are also supported, at all bit
   rates.

3.  Embedded Bit Rates Considerations

   The embedded property of G.729.1 streams provides a mechanism to
   adjust the bandwidth demand.  At any time, a sender can change its
   sending bit rate without external signalling, and the receiver will
   be able to properly decode the frames.  It may help to control
   congestion, since the bandwidth can be adjusted by selecting another
   bit rate.

   The ability to adjust the bandwidth may also help when having a fixed
   bandwidth link dedicated to voice calls, for example in a residential
   or trunking gateway.  In that case, the system can change the bit
   rates depending on the number of simultaneous calls.  This will only
   impact the sending bandwidth.  In order to adjust the receiving
   bandwidth as well, we introduce an in-band signalling to request the
   other party to change its own sending bit rate.  This in-band request
   is called MBS, for Maximum Bit rate Supported.  It is described in
   Section 5.2.  Note that it is only useful for two-way unicast G.729.1
   traffic, because when A sends an in-band MBS to B in order to request
   that B modify its sending bit rate, it concerns the stream from B to
   A.  If there is no G.729.1 stream in the reverse direction, the MBS
   will have no effect.

4.  RTP Header Usage

   The format of the RTP header is specified in RFC 3550 [3].  This
   payload format uses the fields of the header in a manner consistent
   with that specification.

   The RTP timestamp clock frequency is the same as the default sampling
   frequency: 16 kHz.

   G.729.1 has also the capability to operate with 8 kHz sampled input/
   output signals at all bit rates.  It does not affect the bitstream,
   and the decoder does not require a priori knowledge about the
   sampling rate of the original signal at the input of the encoder.
   Therefore, depending on the implementation and the audio acoustic
   capabilities of the devices, the input of the encoder and/or the
   output of the decoder can be configured at 8 kHz; however, a 16 kHz
   RTP clock rate MUST always be used.

   The duration of one frame is 20 ms, corresponding to 320 samples at
   16 kHz.  Thus the timestamp is increased by 320 for each consecutive
   frame.

   The M bit MUST be set to zero in all packets.

   The assignment of an RTP payload type for this packet format is
   outside the scope of the document, and will not be specified here.
   It is expected that the RTP profile under which this payload format
   is being used will assign a payload type for this codec or specify
   that the payload type is to be bound dynamically (see Section 6.2).

5.  Payload Format

5.1.  Payload Structure

   The complete payload consists of a payload header of 1 octet,
   followed by zero or more consecutive audio frames at the same bit
   rate.

   The payload header consists of two fields: MBS (see Section 5.2) and
   FT (see Section 5.3).

      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
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
     |  MBS  |   FT  |                                               |
     +-+-+-+-+-+-+-+-+                                               +
     :                zero or more frames at the same bit rate       :
     :                                                               :
     +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

5.2.  Payload Header: MBS Field

   MBS (4 bits): maximum bit rate supported.  Indicates a maximum bit
   rate to the encoder at the site of the receiver of this payload.  The
   value of the MBS field is set according to the following table:

                         +-------+--------------+
                         |  MBS  | max bit rate |
                         +-------+--------------+
                         |   0   |    8 kbps    |
                         |   1   |    12 kbps   |
                         |   2   |    14 kbps   |
                         |   3   |    16 kbps   |
                         |   4   |    18 kbps   |
                         |   5   |    20 kbps   |
                         |   6   |    22 kbps   |
                         |   7   |    24 kbps   |
                         |   8   |    26 kbps   |
                         |   9   |    28 kbps   |
                         |   10  |    30 kbps   |
                         |   11  |    32 kbps   |
                         | 12-14 |  (reserved)  |
                         |   15  |    NO_MBS    |
                         +-------+--------------+

   The MBS is used to tell the other party the maximum bit rate one can
   receive.  The encoder MUST NOT exceed the sending rate indicated by
   the received MBS.  Note that, due to the embedded property of the
   coding scheme, the encoder can send frames at the MBS rate or any
   lower rate.  As long as it does not exceed the MBS, the encoder can
   change its bit rate at any time without previous notice.

   Note that the MBS is a codec bit rate; the actual network bit rate is
   higher and depends on the overhead of the underlying protocols.

   The MBS received is valid until the next MBS is received, i.e., a
   newly received MBS value overrides the previous one.

   If a payload with a reserved MBS value is received, the MBS MUST be
   ignored.

   The MBS field MUST be set to 15 for packets sent to a multicast group
   and MUST be ignored on packets received from a multicast group.

   The MBS field MUST be set to 15 in all packets when the actual MBS
   value is sent through non-RTP means.  This is out of the scope of
   this specification.

   See Sections 3 and 7 for more details on the use of MBS for
   congestion control.

5.3.  Payload Header: FT Field

   FT (4 bits): Frame type of the frame(s) in this packet, as per the
   following table:

                  +-------+---------------+------------+
                  |   FT  | encoding rate | frame size |
                  +-------+---------------+------------+
                  |   0   |     8 kbps    |  20 octets |
                  |   1   |    12 kbps    |  30 octets |
                  |   2   |    14 kbps    |  35 octets |
                  |   3   |    16 kbps    |  40 octets |
                  |   4   |    18 kbps    |  45 octets |
                  |   5   |    20 kbps    |  50 octets |
                  |   6   |    22 kbps    |  55 octets |
                  |   7   |    24 kbps    |  60 octets |
                  |   8   |    26 kbps    |  65 octets |
                  |   9   |    28 kbps    |  70 octets |
                  |   10  |    30 kbps    |  75 octets |
                  |   11  |    32 kbps    |  80 octets |
                  | 12-14 |   (reserved)  |            |
                  |   15  |    NO_DATA    |      0     |
                  +-------+---------------+------------+

   The FT value 15 (NO_DATA) indicates that there is no audio data in
   the payload.  This MAY be used to update the MBS value when there is
   no audio frame to transmit.  The payload will then be reduced to the
   payload header.

   If a payload with a reserved FT value is received, the whole payload
   MUST be ignored.

5.4.  Audio Data

   Audio data of a payload contains one or more consecutive audio frames
   at the same bit rate.  The audio frames are packed in order of time,
   that is, oldest first.

   The size of one frame is given by the FT field, as per the table in
   Section 5.3, and the actual number of frames is easy to infer from
   the size of the audio data part:

      nb_frames = (size_of_audio_data) / (size_of_one_frame).

   Only full frames must be considered.  So if there is a remainder to
   the division above, the corresponding remaining bytes in the received
   payload MUST be ignored.

   Note that if FT=15, there will be no audio frame in the payload.

6.  Payload Format Parameters

   This section defines the parameters that may be used to configure
   optional features in the G.729.1 RTP transmission.

   The parameters are defined here as part of the media subtype
   registration for the G.729.1 codec.  A mapping of the parameters into
   the Session Description Protocol (SDP) [5] is also provided for those
   applications that use SDP.  In control protocols that do not use MIME
   or SDP, the media type parameters must be mapped to the appropriate
   format used with that control protocol.

6.1.  Media Type Registration

   This registration is done using the template defined in RFC 4288 [6]
   and following RFC 3555 [7].

   Type name: audio

   Subtype name: G7291

   Required parameters: none

   Optional parameters:

   maxbitrate:  the absolute maximum codec bit rate for the session, in
      bits per second.  Permissible values are 8000, 12000, 14000,
      16000, 18000, 20000, 22000, 24000, 26000, 28000, 30000, and 32000.
      32000 is implied if this parameter is omitted.  The maxbitrate
      restricts the range of bit rates which can be used.  The bit rates
      indicated by FT and MBS fields in the RTP packets MUST NOT exceed
      maxbitrate.

   mbs:  the current maximum codec bit rate supported as a receiver, in
      bits per second.  Permissible values are in the same set as for
      the maxbitrate parameter, with the constraint that mbs MUST be
      lower or equal to maxbitrate.  If the mbs parameter is omitted, it
      is set to the maxbitrate value.  So if both mbs and maxbitrate are
      omitted, they are both set to 32000.  The mbs parameter
      corresponds to a MBS value in the RTP packets as per table in
      Section 5.2 of RFC 4749.  Note that this parameter may be
      dynamically updated by the MBS field of the RTP packets sent; it
      is not an absolute value for the session.

   ptime:  the recommended length of time (in milliseconds) represented
      by the media in a packet.  See Section 6 of RFC 4566 [5].

   maxptime:  the maximum length of time (in milliseconds) that can be
      encapsulated in a packet.  See Section 6 of RFC 4566 [5]

   Encoding considerations: This media type is framed and contains
      binary data; see Section 4.8 of RFC 4288 [6].

   Security considerations: See Section 8 of RFC 4749

   Interoperability considerations: none

   Published specification: RFC 4749

   Applications which use this media type: Audio and video conferencing
      tools.

   Additional information: none

   Person & email address to contact for further information:
      Aurelien Sollaud, aurelien.sollaud@orange-ftgroup.com

   Intended usage: COMMON

   Restrictions on usage: This media type depends on RTP framing, and
      hence is only defined for transfer via RTP [3].

   Author: Aurelien Sollaud

   Change controller: IETF Audio/Video Transport working group delegated
      from the IESG

6.2.  Mapping to SDP Parameters

   The information carried in the media type specification has a
   specific mapping to fields in the Session Description Protocol (SDP)
   [5], which is commonly used to describe RTP sessions.  When SDP is
   used to specify sessions employing the G.729.1 codec, the mapping is
   as follows:

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

   o  The media subtype ("G7291") goes in SDP "a=rtpmap" as the encoding
      name.  The RTP clock rate in "a=rtpmap" MUST be 16000 for G.729.1.

   o  The parameters "ptime" and "maxptime" go in the SDP "a=ptime" and
      "a=maxptime" attributes, respectively.

   o  Any remaining parameters go in the SDP "a=fmtp" attribute by
      copying them directly from the media type string as a semicolon
      separated list of parameter=value pairs.

   Some example SDP session descriptions utilizing G.729.1 encodings
   follow.

   Example 1: default parameters

      m=audio 53146 RTP/AVP 98
      a=rtpmap:98 G7291/16000

   Example 2: recommended packet duration of 40 ms (=2 frames), maximum
   bit rate is 12 kbps, and initial MBS set to 8 kbps.  It could be a
   loaded PSTN gateway which can operate at 12 kbps but asks to
   initially reduce the bit rate to 8 kbps.

      m=audio 51258 RTP/AVP 99
      a=rtpmap:99 G7291/16000
      a=fmtp:99 maxbitrate=12000; mbs=8000
      a=ptime:40

6.2.1.  Offer-Answer Model Considerations

   The following considerations apply when using SDP offer-answer
   procedures [8] to negotiate the use of G.729.1 payload in RTP:

   o  Since G.729.1 is an extension of G.729, the offerer SHOULD
      announce G.729 support in its "m=audio" line, with G.729.1
      preferred.  This will allow interoperability with both G.729.1 and
      G.729-only capable parties.

      Below is an example of such an offer:

         m=audio 55954 RTP/AVP 98 18
         a=rtpmap:98 G7291/16000
         a=rtpmap:18 G729/8000

      If the answerer supports G.729.1, it will keep the payload type 98
      in its answer, and the conversation will be done using G.729.1.
      Else, if the answerer supports only G.729, it will leave only the
      payload type 18 in its answer, and the conversation will be done
      using G.729 (the payload format for G.729 is defined in Section
      4.5.6 of RFC 3551 [4]).

      Note that when used at 8 kbps in G.729-compatible mode, the
      G.729.1 decoder supports G.729 Annex B.  Therefore, Annex B can be
      advertised (by default, annexb=yes for G729 media type; see
      Section 4.1.9 of RFC 3555 [7]).

   o  The "maxbitrate" parameter is bi-directional.  If the offerer sets
      a maxbitrate value, the answerer MUST reply with a smaller or
      equal value.  The actual maximum bit rate for the session will be
      the minimum.

   o  If the received value for "maxbitrate" is between 8000 and 32000
      but not in the permissible values set, it SHOULD be read as the
      closest lower valid value.  If the received value is lower than
      8000 or greater than 32000, the session MUST be rejected.

   o  The "mbs" parameter is not symmetric.  Values in the offer and the
      answer are independent and take into account local constraints.
      One party MUST NOT start sending frames at a bit rate higher than
      the "mbs" of the other party.  The parameter allows announcing
      this value, prior to the sending of any packet, to prevent the
      remote sender from exceeding the MBS at the beginning of the
      session.

   o  If the received value for "mbs" is greater or equal to 8000 but
      not in the permissible values set, it SHOULD be read as the
      closest lower valid value.  If the received value is lower than
      8000, the session MUST be rejected.

   o  The parameters "ptime" and "maxptime" will in most cases not
      affect interoperability.  The SDP offer-answer handling of the
      "ptime" parameter is described in RFC 3264 [8].  The "maxptime"
      parameter MUST be handled in the same way.

   o  Any unknown parameter in an offer MUST be ignored by the receiver
      and MUST NOT be included in the answer.

   Some special rules apply for mono-directional traffic:

   o  For sendonly streams, the "mbs" parameter is useless and SHOULD
      NOT be used.

   o  For recvonly streams, the "mbs" parameter is the only way to
      communicate the MBS to the sender, since there is no RTP stream
      towards it.  So to request a bit rate change, the receiver will
      need to use an out-of-band mechanism, like a SIP RE-INVITE.

   Some special rules apply for multicast:

   o  The "mbs" parameter MUST NOT be used.

   o  The "maxbitrate" parameter becomes declarative and MUST NOT be
      negotiated.  This parameter is fixed, and a participant MUST use
      the configuration that is provided for the session.


6.2.2.  Declarative SDP Considerations

   For declarative use of SDP such as in SAP [10] and RTSP [11], the
   following considerations apply:

   o  The "mbs" parameter MUST NOT be used.

   o  The "maxbitrate" parameter is declarative and provides the
      parameter that SHALL be used when receiving and/or sending the
      configured stream.


7.  Congestion Control

   Congestion control for RTP SHALL be used in accordance with RFC 3550
   [3] and any appropriate profile (for example, RFC 3551 [4]).  The
   embedded and variable bit rates capability of G.729.1 provides a
   mechanism that may help to control congestion; see Section 3 for more
   details.

   The number of frames encapsulated in each RTP payload influences the
   overall bandwidth of the RTP stream, due to the header overhead.
   Packing more frames in each RTP payload can reduce the number of
   packets sent and hence the header overhead, at the expense of
   increased delay and reduced error robustness.

8.  Security Considerations

   RTP packets using the payload format defined in this specification
   are subject to the general security considerations discussed in the
   RTP specification [3] and any appropriate profile (for example, RFC
   3551 [4]).

   As this format transports encoded speech/audio, the main security
   issues include confidentiality, integrity protection, and
   authentication of the speech/audio itself.  The payload format itself
   does not have any built-in security mechanisms.  Any suitable
   external mechanisms, such as SRTP [12], MAY be used.

   This payload format and the G.729.1 encoding do not exhibit any
   significant non-uniformity in the receiver-end computational load and
   thus are unlikely to pose a denial-of-service threat due to the
   receipt of pathological datagrams.

9.  IANA Considerations

   IANA has registered audio/G7291 as a media subtype; see Section 6.1.

10.  References

10.1.  Normative References

   [1]  International Telecommunications Union, "G.729 based Embedded
        Variable bit-rate coder: An 8-32 kbit/s scalable wideband coder
        bitstream interoperable with G.729", ITU-T Recommendation
        G.729.1, May 2006.

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

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

   [5]  Handley, M., Jacobson, V., and C. Perkins, "SDP: Session
        Description Protocol", RFC 4566, July 2006.

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

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

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

10.2.  Informative References

   [9]  International Telecommunications Union, "Coding of speech at 8
        kbit/s using conjugate-structure algebraic-code-excited linear-
        prediction (CS-ACELP)", ITU-T Recommendation G.729, March 1996.

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

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

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

Author's Address

   Aurelien Sollaud
   France Telecom
   2 avenue Pierre Marzin
   Lannion Cedex  22307
   France

   Phone: +33 2 96 05 15 06
   EMail: aurelien.sollaud@orange-ftgroup.com

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).

ISO Plain Language

On this page
PROBLEMG.729.1 is a scalable wideband speech codec whose bit rate can change on the fly, but no payload format existed for carrying it over RTP, and no way existed for a receiver to tell the sender to use less bandwidth.
MECHANISMEach RTP payload carries a one-octet header with a 4-bit MBS field advertising the highest bit rate the receiver can take and a 4-bit FT field giving the rate and frame size of the audio frames that follow.
WHERE IT STANDSThe payload format lets a sender drop its bit rate without any signalling and lets a receiver ask the sender to do so with the MBS field. The RFC was later updated by RFC 5459.

§1What This RFC Specifies

The ITU-T recommendation G.729.1 is a scalable wideband extension of the recommendation G.729 audio codec. This document specifies the payload format for packetising G.729.1 encoded audio signals into the Real-time Transport Protocol. The payload format itself is described in Section 5, and a media type registration together with the details for using G.729.1 with SDP are given in Section 6.

The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY and OPTIONAL are to be interpreted as described in RFC 2119.

The RFC defines three things: how compressed audio is laid out inside an RTP packet, how the media type for it is registered, and how a session description announces it.

§2The G.729.1 Codec

G.729.1 is an 8 to 32 kbps scalable wideband, 50 to 7000 Hz, speech and audio coding algorithm that is interoperable with G.729, G.729 Annex A and G.729 Annex B. It provides a standardised solution for packetised voice applications that allows a smooth transition from narrowband to wideband telephony.

The services addressed are IP telephony and videoconferencing, either in enterprise corporate networks or in the mass market, such as Public Switched Telephone Network emulation over DSL or wireless access. Target devices can be IP phones or other VoIP handsets, home gateways, media gateways, IP Private Branch Exchange equipment, trunking equipment and voice messaging servers.

For all of those applications the scalability feature allows tuning the bit rate against the quality trade-off, possibly in a dynamic way during a session, taking into account service requirements and network transport constraints.

The G.729.1 coder produces an embedded bitstream structured in 12 layers, corresponding to 12 available bit rates between 8 and 32 kbps. The first layer, at 8 kbps, is the core layer and is bitstream compatible with the G.729 and G.729A coders. At 12 kbps a second layer improves the narrowband quality. Upper layers provide wideband audio between 14 and 32 kbps with a 2 kbps granularity, allowing graceful quality improvements. Only the core layer is mandatory for decoding understandable speech; the upper layers provide quality enhancement and wideband enlargement. The codec operates on 20 ms frames, and the default sampling rate is 16 kHz. Input and output at 8 kHz are also supported, at all bit rates.

The codec builds its output in layers. The bottom layer alone gives understandable narrowband speech; each layer above adds quality or bandwidth. A sender can add or drop layers at any moment and the receiver decodes what it gets.

§3Embedded Bit Rates and the MBS Request

The embedded property of G.729.1 streams provides a mechanism to adjust bandwidth demand. At any time a sender can change its sending bit rate without external signalling, and the receiver will still be able to decode the frames properly. That may help to control congestion, since the bandwidth can be adjusted by selecting another bit rate.

The ability to adjust bandwidth may also help on a fixed bandwidth link dedicated to voice calls, for example in a residential or trunking gateway, where the system can change bit rates depending on the number of simultaneous calls. This only affects the sending bandwidth. To adjust the receiving bandwidth as well, the document introduces an in-band signalling request that asks the other party to change its own sending bit rate. That request is called MBS, for Maximum Bitrate Supported, and is described in Section 5.2.

MBS is only useful for two-way unicast G.729.1 traffic, because when A sends an in-band MBS to B to ask B to modify its sending bit rate, it concerns the stream from B to A. If there is no G.729.1 stream in the reverse direction, the MBS has no effect.

A sender can lower its own bit rate at any time without asking. Lowering the other end's bit rate takes a request, and that request travels inside the audio packets as a field called MBS.

HOW THE MBS FIELD BEHAVES
  • A ceiling, not a settingThe sender may use the MBS rate or any lower rate, and may change rate at any time without notice.
  • Last one winsA received MBS stays valid until the next MBS arrives, and a new value overrides the old one.
  • Reserved valuesIf a payload arrives with a reserved MBS value, that MBS is ignored.
  • No value in multicastThe MBS field is set to 15 in packets sent to a multicast group, and ignored on packets received from one.

§4RTP Header Usage

The format of the RTP header is specified in RFC 3550, and this payload format uses the fields of that header consistently with that specification.

The RTP timestamp clock frequency is the same as the default sampling frequency, 16 kHz. G.729.1 can also operate with 8 kHz sampled input and output signals at all bit rates. That does not affect the bitstream, and the decoder does not need to know in advance the sampling rate of the original signal at the input of the encoder. Depending on the implementation and the acoustic capabilities of the devices, the encoder input and the decoder output can be configured at 8 kHz, but a 16 kHz RTP clock rate MUST always be used.

The duration of one frame is 20 ms, corresponding to 320 samples at 16 kHz, so the timestamp increases by 320 for each consecutive frame. The M bit MUST be set to zero in all packets.

The assignment of an RTP payload type for this packet format is outside the scope of the document. It is expected that the RTP profile under which this payload format is used will assign a payload type for this codec, or specify that the payload type is bound dynamically.

Each packet carries a timestamp that counts 16,000 units per second. One audio frame lasts 20 ms, so the timestamp moves by 320 each frame. The marker bit is always zero, because every packet carries whole frames.

§5The Payload Format

The complete payload consists of a payload header of 1 octet followed by zero or more consecutive audio frames at the same bit rate. The payload header has two fields: MBS, which is 4 bits, and FT, which is 4 bits.

The MBS field carries the maximum bit rate supported and indicates a maximum bit rate to the encoder at the site of the receiver of this payload. Its values run from 0 for 8 kbps through 11 for 32 kbps, with 12 to 14 reserved and the value 15 meaning NO_MBS. The MBS tells the other party the maximum bit rate one can receive, so the encoder MUST NOT exceed the sending rate indicated by the received MBS. Because of the embedded property of the coding scheme, the encoder can send frames at the MBS rate or any lower rate, and can change its bit rate at any time without previous notice as long as it does not exceed the MBS. The MBS is a codec bit rate; the actual network bit rate is higher and depends on the overhead of the underlying protocols. A received MBS value stays valid until the next MBS is received. If a payload with a reserved MBS value is received, the MBS MUST be ignored. The MBS field MUST be set to 15 for packets sent to a multicast group, and MUST be ignored on packets received from a multicast group. It MUST also be set to 15 in all packets when the actual MBS value is sent through non-RTP means.

The FT field gives the frame type of the frames in the packet. Its values run from 0 for 8 kbps and 20 octets through 11 for 32 kbps and 80 octets, with 12 to 14 reserved and the value 15 meaning NO_DATA. The value 15 indicates that there is no audio data in the payload. This MAY be used to update the MBS value when there is no audio frame to transmit, in which case the payload is reduced to the payload header. If a payload with a reserved FT value is received, the whole payload MUST be ignored.

Audio data in a payload contains one or more consecutive audio frames at the same bit rate, packed in order of time, oldest first. The size of one frame comes from the FT field, and the number of frames is inferred from the size of the audio data part: nb_frames = size_of_audio_data divided by size_of_one_frame. Only full frames are considered, so any remainder in the received payload MUST be ignored.

A packet holds one byte of header followed by audio frames. The header says how fast the sender is going and how big each frame is. The receiver works out how many frames arrived by dividing the data length by the frame size, and drops any leftover bytes.

THE ONE-OCTET PAYLOAD HEADER (RFC 4749)
G.729.1 RTP payload
FieldNameSizeValueDescription
MBSMBS4 bitsMaximum bit rate supported by the receiver of this payload. 0 to 11 map to 8 to 32 kbps, 15 means NO_MBS.
FTFT4 bitsFrame type of the frames in this packet. 0 to 11 give the rate and frame size, 15 means NO_DATA.
FT VALUES: ENCODING RATE AND FRAME SIZE (RFC 4749)
FTEncoding rateFrame size
08 kbps20 octets
112 kbps30 octets
214 kbps35 octets
316 kbps40 octets
418 kbps45 octets
520 kbps50 octets
622 kbps55 octets
724 kbps60 octets
826 kbps65 octets
928 kbps70 octets
1030 kbps75 octets
1132 kbps80 octets
12-14(reserved)
15NO_DATA0

§6Media Type and SDP Parameters

The media type is registered from the template in RFC 4288, following RFC 3555. The type name is audio, the subtype name is G7291, and there are no required parameters. The optional parameters are maxbitrate, the absolute maximum codec bit rate for the session in bits per second, whose permissible values are 8000, 12000, 14000, 16000, 18000, 20000, 22000, 24000, 26000, 28000, 30000 and 32000, with 32000 implied if the parameter is omitted; mbs, the current maximum codec bit rate supported as a receiver, with the same set of permissible values but constrained to be lower than or equal to maxbitrate, and set to the maxbitrate value if omitted; and ptime and maxptime, the recommended and maximum length of time in milliseconds that the media in a packet represents. The bit rates indicated by the FT and MBS fields in the RTP packets MUST NOT exceed maxbitrate.

For SDP, the media type goes in the m= line as the media name, the subtype goes in a=rtpmap as the encoding name with an RTP clock rate that MUST be 16000 for G.729.1, ptime and maxptime go in the matching attributes, and any remaining parameters go in a=fmtp as a semicolon separated list of parameter and value pairs.

Under the SDP offer-answer model, the offerer SHOULD announce G.729 support in its m=audio line with G.729.1 preferred, so that a G.729-only party can still take part. The maxbitrate parameter is bi-directional: if the offerer sets a value, the answerer MUST reply with a smaller or equal value, and the actual maximum for the session is the minimum. The mbs parameter is not symmetric; values in the offer and the answer are independent and reflect local constraints, and neither party MUST start sending frames at a bit rate higher than the other party's mbs. For sendonly streams the mbs parameter is useless and SHOULD NOT be used, while for recvonly streams it is the only way to communicate the MBS to the sender, since there is no RTP stream towards it. For multicast the mbs parameter MUST NOT be used, and maxbitrate becomes declarative and MUST NOT be negotiated. In declarative SDP, such as SAP and RTSP, the mbs parameter MUST NOT be used and maxbitrate is declarative.

The media type is audio/G7291, and a session description carries the rate limits as parameters. The session maximum is whichever side asks for less. The MBS value is per direction, so each side states its own limit, and it means nothing in multicast.

§7Congestion Control

Congestion control for RTP SHALL be used in accordance with RFC 3550 and any appropriate profile, for example RFC 3551. The embedded and variable bit rate capability of G.729.1 provides a mechanism that may help to control congestion.

The number of frames encapsulated in each RTP payload influences the overall bandwidth of the RTP stream, because of the header overhead. Packing more frames in each payload can reduce the number of packets sent and hence the header overhead, at the expense of increased delay and reduced error robustness.

Putting several frames in one packet saves the cost of repeated headers, but adds delay and makes a single lost packet cost more audio. The codec's variable rate gives another lever, since the sender can drop to a lower rate when the network is busy.

§8Security Considerations

RTP packets using this payload format are subject to the general security considerations discussed in the RTP specification and in any appropriate profile.

Because the format transports encoded speech and audio, the main security issues are confidentiality, integrity protection, and authentication of the speech and audio itself. The payload format itself has no built-in security mechanisms. Any suitable external mechanisms, such as SRTP, MAY be used.

The document also notes that this payload format and the G.729.1 encoding do not exhibit any significant non-uniformity in the receiver-end computational load, and so are unlikely to pose a denial-of-service threat due to the receipt of pathological datagrams.

The format carries no protection of its own. A separate mechanism such as SRTP is needed to keep the call contents private and to stop them being altered in transit.

§9References