Internet Engineering Task Force (IETF)                 M. Petit-Huguenin
Request for Comments: 7160                            Impedance Mismatch
Updates: 3550                                               G. Zorn, Ed.
Category: Standards Track                                    Network Zen
ISSN: 2070-1721                                               April 2014


          Support for Multiple Clock Rates in an RTP Session

Abstract

  This document clarifies the RTP specification regarding the use of
  different clock rates in an RTP session.  It also provides guidance
  on how legacy RTP implementations that use multiple clock rates can
  interoperate with RTP implementations that use the algorithm
  described in this document.  It updates RFC 3550.

Status of This Memo

  This is an Internet Standards Track document.

  This document is a product of the Internet Engineering Task Force
  (IETF).  It represents the consensus of the IETF community.  It has
  received public review and has been approved for publication by the
  Internet Engineering Steering Group (IESG).  Further information on
  Internet Standards is available in Section 2 of RFC 5741.

  Information about the current status of this document, any errata,
  and how to provide feedback on it may be obtained at
  http://www.rfc-editor.org/info/rfc7160.

Copyright Notice

  Copyright (c) 2014 IETF Trust and the persons identified as the
  document authors.  All rights reserved.

  This document is subject to BCP 78 and the IETF Trust's Legal
  Provisions Relating to IETF Documents
  (http://trustee.ietf.org/license-info) in effect on the date of
  publication of this document.  Please review these documents
  carefully, as they describe your rights and restrictions with respect
  to this document.  Code Components extracted from this document must
  include Simplified BSD License text as described in Section 4.e of
  the Trust Legal Provisions and are provided without warranty as
  described in the Simplified BSD License.






Petit-Huguenin & Zorn        Standards Track                    [Page 1]

RFC 7160                  Multiple Clock Rates                April 2014


Table of Contents

  1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
  2.  Terminology . . . . . . . . . . . . . . . . . . . . . . . . .   4
  3.  Legacy RTP  . . . . . . . . . . . . . . . . . . . . . . . . .   4
    3.1.  Different SSRC  . . . . . . . . . . . . . . . . . . . . .   4
    3.2.  Same SSRC . . . . . . . . . . . . . . . . . . . . . . . .   5
      3.2.1.  Monotonic Timestamps  . . . . . . . . . . . . . . . .   5
      3.2.2.  Non-monotonic Timestamps  . . . . . . . . . . . . . .   6
  4.  Recommendations . . . . . . . . . . . . . . . . . . . . . . .   6
    4.1.  RTP Sender (with RTCP)  . . . . . . . . . . . . . . . . .   6
    4.2.  RTP Sender (without RTCP) . . . . . . . . . . . . . . . .   6
    4.3.  RTP Receiver  . . . . . . . . . . . . . . . . . . . . . .   7
  5.  Security Considerations . . . . . . . . . . . . . . . . . . .   7
  6.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .   8
  7.  References  . . . . . . . . . . . . . . . . . . . . . . . . .   8
    7.1.  Normative References  . . . . . . . . . . . . . . . . . .   8
    7.2.  Informative References  . . . . . . . . . . . . . . . . .   8
  Appendix A.  Example Values . . . . . . . . . . . . . . . . . . .  10
  Appendix B.  Using a Fixed Clock Rate . . . . . . . . . . . . . .  12
  Appendix C.  Behavior of Legacy Implementations . . . . . . . . .  12
    C.1.  libccrtp 2.0.2  . . . . . . . . . . . . . . . . . . . . .  12
    C.2.  libmediastreamer0 2.6.0 . . . . . . . . . . . . . . . . .  12
    C.3.  libpjmedia 1.0  . . . . . . . . . . . . . . . . . . . . .  13
    C.4.  Android RTP Stack 4.0.3 . . . . . . . . . . . . . . . . .  13

1.  Introduction

  The clock rate is a parameter of the payload format as identified in
  RTP and RTCP (RTP Control Protocol) by the payload type value.  It is
  often defined as being the same as the sampling rate but that is not
  always the case (see, for example, the G722 and MPA audio codecs
  [RFC3551]).

  An RTP sender can switch between different payloads during the
  lifetime of an RTP session and because clock rates are defined by
  payload format, it is possible that the clock rate will also vary
  during an RTP session.  Schulzrinne, et al. [RFC3550] lists using
  multiple clock rates as one of the reasons to not use different
  payloads on the same Synchronization Source (SSRC).  Unfortunately,
  this advice has not always been followed and some RTP implementations
  change the payload in the same SSRC, even if the different payloads
  use different clock rates.








Petit-Huguenin & Zorn        Standards Track                    [Page 2]

RFC 7160                  Multiple Clock Rates                April 2014


  This creates three problems:

  o  The method used to calculate the RTP timestamp field in an RTP
     packet is underspecified.

  o  When the same SSRC is used for different clock rates, it is
     difficult to know what clock rate was used for the RTP timestamp
     field in an RTCP Sender Report (SR) packet.

  o  When the same SSRC is used for different clock rates, it is
     difficult to know what clock rate was used for the interarrival
     jitter field in an RTCP Receiver Report (RR) packet.

  Table 1 contains a non-exhaustive list of fields in RTCP packets that
  uses a clock rate as a unit:

         +---------------------+------------------+------------+
         | Field name          | RTCP packet type | Reference  |
         +---------------------+------------------+------------+
         | RTP timestamp       | SR               | [RFC3550]  |
         |                     |                  |            |
         | Interarrival jitter | RR               | [RFC3550]  |
         |                     |                  |            |
         | min_jitter          | XR Summary Block | [RFC3611]  |
         |                     |                  |            |
         | max_jitter          | XR Summary Block | [RFC3611]  |
         |                     |                  |            |
         | mean_jitter         | XR Summary Block | [RFC3611]  |
         |                     |                  |            |
         | dev_jitter          | XR Summary Block | [RFC3611]  |
         |                     |                  |            |
         | Interarrival jitter | IJ               | [RFC5450]  |
         |                     |                  |            |
         | RTP timestamp       | SMPTETC          | [RFC5484]  |
         |                     |                  |            |
         | Jitter              | RSI Jitter Block | [RFC5760]  |
         |                     |                  |            |
         | Median jitter       | RSI Stats Block  | [RFC5760]  |
         +---------------------+------------------+------------+

                                 Table 1

  Section 3 and its subsections try to list all of the algorithms known
  to be used in existing RTP implementations at the time of writing.
  These sections are not normative.

  Section 4 and its subsections recommend a unique algorithm that
  modifies RFC 3550.  These sections are normative.



Petit-Huguenin & Zorn        Standards Track                    [Page 3]

RFC 7160                  Multiple Clock Rates                April 2014


2.  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 [RFC2119].

  In addition, this document uses the following terms:

  Clock rate     The multiplier used to convert from a wallclock value
                 in seconds to an equivalent RTP timestamp value
                 (without the fixed random offset).  Note that RFC 3550
                 uses various terms like "clock frequency", "media
                 clock rate", "timestamp unit", "timestamp frequency",
                 and "RTP timestamp clock rate" as synonymous to clock
                 rate.

  RTP Sender     A logical network element that sends RTP packets,
                 sends RTCP SR packets, and receives RTCP reception
                 report blocks.

  RTP Receiver   A logical network element that receives RTP packets,
                 receives RTCP SR packets, and sends RTCP reception
                 report blocks.

3.  Legacy RTP

  The following sections describe the various ways in which legacy RTP
  implementations behave when multiple clock rates are used.  "Legacy
  RTP" refers to RFC 3550 without the modifications introduced by this
  document.

3.1.  Different SSRC

  One way of managing multiple clock rates is to use a different SSRC
  for each different clock rate, as in this case there is no ambiguity
  on the clock rate used by fields in the RTCP packets.  This method
  also seems to be the original intent of RTP as can be deduced from
  points 2 and 3 of Section 5.2 of RFC 3550.

  On the other hand, changing the SSRC can be a problem for some
  implementations designed to work only with unicast IP addresses,
  where having multiple SSRCs is considered a corner case.  Lip
  synchronization can also be a problem in the interval between the
  beginning of the new stream and the first RTCP SR packet.







Petit-Huguenin & Zorn        Standards Track                    [Page 4]

RFC 7160                  Multiple Clock Rates                April 2014


3.2.  Same SSRC

  The simplest way to manage multiple clock rates is to use the same
  SSRC for all of the payload types regardless of the clock rates.

  Unfortunately, there is no clear definition on how the RTP timestamp
  should be calculated in this case.  The following subsections present
  the algorithms currently in use.

3.2.1.  Monotonic Timestamps

  This method of calculating the RTP timestamp ensures that the value
  increases monotonically.  The formula used by this method is as
  follows:

  timestamp = previous_timestamp
              + (current_capture_time - previous_capture_time)
              * current_clock_rate

  The problem with this method is that the jitter calculation on the
  receiving side gives an invalid result during the transition between
  two clock rates, as shown in Table 2 (Appendix A).  The capture and
  arrival time are measured in seconds, starting at the beginning of
  the capture of the first packet; clock rate is measured in Hz; the
  RTP timestamp does not include the random offset; and the transit,
  jitter, and average jitter use the clock rate as a unit.

  Calculating the correct transit time on the receiving side can be
  done by using the following formulas:

  1.  current_capture_time = (current_timestamp - previous_timestamp) /
      current_clock_rate + previous_capture_time

  2.  transit = current_clock_rate * (arrival_time -
      current_capture_time)

  3.  previous_capture_time = current_capture_time

  The main problem with this method, in addition to the fact that the
  jitter calculation described in RFC 3550 cannot be used, is that it
  is dependent on the previous RTP packets, which can be reordered or
  lost in the network.









Petit-Huguenin & Zorn        Standards Track                    [Page 5]

RFC 7160                  Multiple Clock Rates                April 2014


3.2.2.  Non-monotonic Timestamps

  An alternate way of generating the RTP timestamps is to use the
  following formula:

                  timestamp = capture_time * clock_rate

  With this formula, the jitter calculation is correct but the RTP
  timestamp values are no longer increasing monotonically as shown in
  Table 3 (Appendix A).  RFC 3550 states that "[t]he sampling instant
  MUST be derived from a clock that increments monotonically . . .",
  but it does not say that the RTP timestamp must increment
  monotonically.

  The advantage with this method is that it works with the jitter
  calculation described in RFC 3550, as long as the correct clock rates
  are used.  It seems that this is what most implementations are using
  (based on a survey done at SIPit26 and on a survey of open source
  implementations, see Appendix C).

4.  Recommendations

  The following subsections describe behavioral recommendations for RTP
  senders (with and without RTCP) and RTP receivers.

4.1.  RTP Sender (with RTCP)

  An RTP Sender with RTCP turned on MUST use a different SSRC for each
  different clock rate.  An RTCP BYE MUST be sent and a new SSRC MUST
  be used if the clock rate switches back to a value already seen in
  the RTP stream.

  To accelerate lip synchronization, the next compound RTCP packet sent
  by the RTP sender MUST contain multiple SR packets, the first one
  containing the mapping for the current clock rate and the subsequent
  SR packet(s) containing the mapping for the other clock rates seen
  during the last period.

  The RTP extension defined by Perkins & Schierl [RFC6051] MAY be used
  to accelerate the synchronization.

4.2.  RTP Sender (without RTCP)

  An RTP Sender with RTCP turned off (i.e., having set the RTP Sender
  and RTP Receiver bandwidth modifiers [RFC3556] to 0) SHOULD use a
  different SSRC for each different clock rate but MAY use different
  clock rates on the same SSRC as long as the RTP timestamp is
  calculated as explained below:



Petit-Huguenin & Zorn        Standards Track                    [Page 6]

RFC 7160                  Multiple Clock Rates                April 2014


  Each time the clock rate changes, the start_offset and capture_start
  values are calculated with the following formulas:

  start_offset += (capture_time - capture_start) * previous_clock_rate
  capture_start = capture_time

  For the first RTP packet, the values are initialized with the
  following values:

                  start_offset = random_initial_offset
                  capture_start = capture_time

  After eventually updating these values, the RTP timestamp is
  calculated with the following formula:

         timestamp = (capture_time - capture_start) * clock_rate
                     + start_offset

  Note that in all the formulas, capture_start is the first instant
  that the new timestamp rate is used.  The output of the above method
  is exemplified in Table 4 (Appendix A).

4.3.  RTP Receiver

  An RTP Receiver MUST calculate the jitter using the following
  formula:

        D(i,j) = (arrival_time_j * clock_rate_i - timestamp_j)
                 - (arrival_time_i * clock_rate_i - timestamp_i)

  An RTP Receiver MUST be able to handle a compound RTCP packet with
  multiple SR packets.

5.  Security Considerations

  When the algorithm described in Section 4.1 is used, the security
  considerations described in RFC 3550 apply.

  The algorithm described in Section 4.2 is new and so its security
  properties were not considered in RFC 3550.  Although the RTP
  timestamp is initialized with a random value like before, the
  timestamp value depends on the current and previous clock rates; this
  may or may not introduce a security vulnerability in the protocol.








Petit-Huguenin & Zorn        Standards Track                    [Page 7]

RFC 7160                  Multiple Clock Rates                April 2014


6.  Acknowledgements

  Thanks to Colin Perkins, Ali C. Begen, Harald Alvestrand, Qin Wu,
  Jonathan Lennox, Barry Leiba, David Harrington, Stephen Farrell,
  Spencer Dawkins, Wassim Haddad, and Magnus Westerlund for comments,
  suggestions, and questions that helped to improve this document.

  Thanks to Bo Burman, who provided the values in Table 4 of
  Appendix A.

  Thanks to Robert Sparks and the attendees of SIPit 26 for the survey
  on multiple clock rates interoperability.

7.  References

7.1.  Normative References

  [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
             Requirement Levels", BCP 14, RFC 2119, March 1997.

  [RFC3550]  Schulzrinne, H., Casner, S., Frederick, R., and V.
             Jacobson, "RTP: A Transport Protocol for Real-Time
             Applications", STD 64, RFC 3550, July 2003.

7.2.  Informative References

  [AVT-VAR-RATE]
             Wenger, S. and C. Perkins, "RTP Timestamp Frequency for
             Variable Rate Audio Codecs", Work in Progress, October
             2004.

  [RFC3551]  Schulzrinne, H. and S. Casner, "RTP Profile for Audio and
             Video Conferences with Minimal Control", STD 65, RFC 3551,
             July 2003.

  [RFC3556]  Casner, S., "Session Description Protocol (SDP) Bandwidth
             Modifiers for RTP Control Protocol (RTCP) Bandwidth", RFC
             3556, July 2003.

  [RFC3611]  Friedman, T., Caceres, R., and A. Clark, "RTP Control
             Protocol Extended Reports (RTCP XR)", RFC 3611, November
             2003.

  [RFC5450]  Singer, D. and H. Desineni, "Transmission Time Offsets in
             RTP Streams", RFC 5450, March 2009.

  [RFC5484]  Singer, D., "Associating Time-Codes with RTP Streams", RFC
             5484, March 2009.



Petit-Huguenin & Zorn        Standards Track                    [Page 8]

RFC 7160                  Multiple Clock Rates                April 2014


  [RFC5760]  Ott, J., Chesterfield, J., and E. Schooler, "RTP Control
             Protocol (RTCP) Extensions for Single-Source Multicast
             Sessions with Unicast Feedback", RFC 5760, February 2010.

  [RFC6051]  Perkins, C. and T. Schierl, "Rapid Synchronisation of RTP
             Flows", RFC 6051, November 2010.













































Petit-Huguenin & Zorn        Standards Track                    [Page 9]

RFC 7160                  Multiple Clock Rates                April 2014


Appendix A.  Example Values

  The following tables illustrate the timestamp and jitter values
  produced when the various methods discussed in the text are used.

  The values shown are purely exemplary, illustrative, and non-
  normative.

  +-------+-------+-----------+---------+---------+--------+----------+
  | Capt. | Clock | RTP       | Arrival | Transit | Jitter | Average  |
  | time  | rate  | timestamp | time    |         |        | jitter   |
  +-------+-------+-----------+---------+---------+--------+----------+
  | 0     | 8000  | 0         | 0.1     | 800     |        |          |
  |       |       |           |         |         |        |          |
  | 0.02  | 8000  | 160       | 0.12    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.04  | 8000  | 320       | 0.14    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.06  | 8000  | 480       | 0.16    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.08  | 16000 | 800       | 0.18    | 2080    | 480    | 30       |
  |       |       |           |         |         |        |          |
  | 0.1   | 16000 | 1120      | 0.2     | 2080    | 0      | 28       |
  |       |       |           |         |         |        |          |
  | 0.12  | 16000 | 1440      | 0.22    | 2080    | 0      | 26       |
  |       |       |           |         |         |        |          |
  | 0.14  | 8000  | 1600      | 0.24    | 320     | 720    | 70       |
  |       |       |           |         |         |        |          |
  | 0.16  | 8000  | 1760      | 0.26    | 320     | 0      | 65       |
  +-------+-------+-----------+---------+---------+--------+----------+

                      Table 2: Monotonic Timestamps



















Petit-Huguenin & Zorn        Standards Track                   [Page 10]

RFC 7160                  Multiple Clock Rates                April 2014


  +-------+-------+-----------+---------+---------+--------+----------+
  | Capt. | Clock | RTP       | Arrival | Transit | Jitter | Average  |
  | time  | rate  | timestamp | time    |         |        | jitter   |
  +-------+-------+-----------+---------+---------+--------+----------+
  | 0     | 8000  | 0         | 0.1     | 800     |        |          |
  |       |       |           |         |         |        |          |
  | 0.02  | 8000  | 160       | 0.12    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.04  | 8000  | 320       | 0.14    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.06  | 8000  | 480       | 0.16    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.08  | 16000 | 1280      | 0.18    | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.1   | 16000 | 1600      | 0.2     | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.12  | 16000 | 1920      | 0.22    | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.14  | 8000  | 1120      | 0.24    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.16  | 8000  | 1280      | 0.26    | 800     | 0      | 0        |
  +-------+-------+-----------+---------+---------+--------+----------+

                    Table 3: Non-monotonic Timestamps



























Petit-Huguenin & Zorn        Standards Track                   [Page 11]

RFC 7160                  Multiple Clock Rates                April 2014


  +-------+-------+-----------+---------+---------+--------+----------+
  | Capt. | Clock | RTP       | Arrival | Transit | Jitter | Average  |
  | time  | rate  | timestamp | time    |         |        | jitter   |
  +-------+-------+-----------+---------+---------+--------+----------+
  | 0     | 8000  | 0         | 0.1     | 800     |        |          |
  |       |       |           |         |         |        |          |
  | 0.02  | 8000  | 160       | 0.12    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.04  | 8000  | 320       | 0.14    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.06  | 8000  | 480       | 0.16    | 800     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.08  | 16000 | 640       | 0.18    | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.1   | 16000 | 960       | 0.2     | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.12  | 16000 | 1280      | 0.22    | 1600    | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.14  | 8000  | 1600      | 0.24    | 320     | 0      | 0        |
  |       |       |           |         |         |        |          |
  | 0.16  | 8000  | 1760      | 0.26    | 320     | 0      | 0        |
  +-------+-------+-----------+---------+---------+--------+----------+

        Table 4: Recommended Method for RTP Sender (without RTCP)

Appendix B.  Using a Fixed Clock Rate

  An alternate way of fixing the issue with using multiple clock rates
  was proposed by Wenger and Perkins [AVT-VAR-RATE].  This document
  proposed to define a unified clock rate, but the proposal was
  rejected at IETF 61.

Appendix C.  Behavior of Legacy Implementations

C.1.  libccrtp 2.0.2

  This library uses the formula described in Section 3.2.2.

  Note that this library uses gettimeofday(2) which is not guaranteed
  to increment monotonically (e.g., when the clock is adjusted by NTP).

C.2.  libmediastreamer0 2.6.0

  This library (which uses the oRTP library) uses the formula described
  in Section 3.2.2.

  Note that in some environments this library uses gettimeofday(2),
  which is not guaranteed to increment monotonically.



Petit-Huguenin & Zorn        Standards Track                   [Page 12]

RFC 7160                  Multiple Clock Rates                April 2014


C.3.  libpjmedia 1.0

  This library uses the formula described in Section 3.2.2.

C.4.  Android RTP Stack 4.0.3

  This library changes the SSRC each time the format changes, as
  described in Section 3.1.

Authors' Addresses

  Marc Petit-Huguenin
  Impedance Mismatch

  EMail: [email protected]


  Glen Zorn (editor)
  Network Zen
  227/358 Thanon Sanphawut
  Bang Na, Bangkok  10260
  Thailand

  Phone: +66 (0) 8-1000-4155
  EMail: [email protected]


























Petit-Huguenin & Zorn        Standards Track                   [Page 13]