Skip to main content

10. - SIP SDP - Session Descripton Protocol

SIP SDP

SIP SDP stands for Session Description Protocol and is a protocol used to describe multimedia sessions for the purposes of session initiation, control, and management. In the context of SIP, SDP is typically used in the body of SIP messages to describe the characteristics of media streams being offered or negotiated in a SIP session. This includes information such as the type of media, codec, bandwidth, and IP addresses and ports for the media streams. SDP is an important part of the SIP protocol because it allows endpoints to negotiate and establish media sessions for audio, video, and other types of multimedia content.

There are a lot of fieldes availabble in SDP and some are mandatory (shown here in blue). New fields can be added as needed. if an agent doesn't understand a field it simply ignores it.

image.png

 

 

 

MIME

MIME stands for Multipurpose Internet Mail Extensions, which is a standard that extends the format of email messages to support text in character sets other than ASCII, as well as attachments of audio, video, images, and application programs. In the context of SIP SDP (Session Description Protocol), MIME is used to specify the media type and format of the SDP body in the SIP message. The MIME type for SDP is "application/sdp". The SDP body contains information about the media streams to be used for the session, such as the codecs, transport protocols, and media formats to be used. The MIME type allows the receiving endpoint to determine how to handle the SDP body in the SIP message.

 

 

INVITE sip:[email protected] SIP/2.0
Via: SIP/2.0/UDP pc33.atlanta.com;branch=z9hG4bK776asdhds
Max-Forwards: 70
To: Bob <sip:[email protected]>
From: Alice <sip:[email protected]>;tag=1928301774
Call-ID: a84b4c76e66710
CSeq: 314159 INVITE
Contact: <sip:[email protected]>
User-Agent: BriaX 3.5.5 build 81887
Allow: INVITE, ACK, CANCEL, OPTIONS, BYE, REFER, NOTIFY, MESSAGE, SUBSCRIBE, INFO
Allow-Events: presence, message-summary, refer, dialog
Session-Expires: 3600;refresher=uas
Content-Type: application/sdp
Content-Length: 283

v=0
o=- 123456 789012 IN IP4 pc33.atlanta.com
s=SIP Call
c=IN IP4 192.0.2.1
t=0 0
m=audio 49170 RTP/AVP 0 8 18 3 101
a=rtpmap:0 PCMU/8000
a=rtpmap:8 PCMA/8000
a=rtpmap:18 G729/8000
a=fmtp:18 annexb=no
a=rtpmap:3 GSM/8000
a=rtpmap:101 telephone-event/8000
a=fmtp:101 0-16
m=video 49172 RTP/AVP 31 34
a=rtpmap:31 H261/90000
a=rtpmap:34 H263/90000

 

In this example, the header includes the following fields:

Expand to view Header description
  • Via: Indicates the transport protocol, IP address, and port number used for the request. In this case, it is a UDP transport protocol and the IP address is pc33.atlanta.com.
  • Max-Forwards: Indicates the maximum number of times the request can be forwarded before it is discarded.
  • To: Contains the display name and SIP URI of the destination user.
  • From: Contains the display name and SIP URI of the source user, along with a unique identifier (tag) for the request.
  • Call-ID: Unique identifier for the call.
  • CSeq: Sequence number for the request.
  • Contact: Indicates the SIP URI that can be used to reach the sender of the request.
  • User-Agent: Indicates the user agent that generated the request.
  • Allow: Lists the SIP methods that the user agent is able to handle.
  • Allow-Events: Lists the types of events that the user agent can generate.
  • Session-Expires: Indicates the duration of the session in seconds and the refresher type (in this case, uas).


The SDP message in the body of the request describes the media capabilities of the sender and includes the following information:

  • v (Protocol version): Indicates the protocol version of the session description. In this case, it is 0.
  • o (Origin): Specifies the originator of the session and a unique identifier for the session. The fields include username, session identifier, session version number, and network type (IN for Internet) and address type (IP4 for IPv6).

    O=<username> <sess-id> <sess-version> <nettype> <addrtype> <unicast-address>

    The nettype and addrtype fields can take various values depending on the network protocol being used. Some of the options for the nettype field include:

    • IN: Internet
    • AT: Atmosphere
    • FI: FIP Snooping
    • Token-Ring: IBM Token Ring Network

    Some of the options for the addrtype field include:

    • IP4: IPv4 address
    • IP6: IPv6 address
    • MAC: MAC address
    • NSAP: Network Service Access Point
    • E.164: E.164 telephone number
  • s (Session name): Specifies a human-readable session name.
  • c (Connection information): Specifies the connection information for the session. In this case, it is the IP address 10.10.10.10. - This field specifies the network address type IP4 or IP6 and the connection address IP Address or hostname of the media stream, there are different types of connection information fields: 
    • c=IN IP4 192.168.1.100: This specifies that the media stream will use IPv4 and the IP address is 192.168.1.100.
    • c=IN IP6 ::1: This specifies that the media stream will use IPv6 and the IP address is ::1 (loopback address).
    • c=IN IP4 192.168.1.100/255: This specifies that the media stream will use IPv4 and the IP address is 192.168.1.100, with a subnet mask of 255.255.255.0.

  • t (Timing): Specifies the start and stop times of the session. In this case, it is set to 0, which means the session is not limited by time.
  • m (Media): Specifies the media type (audio), the port number (1234), and the transport protocol (RTP/AVP). There can be multiple media fields in an SDP message.
    • m=audio 49170 RTP/AVP 0 8: This specifies an audio stream on port 49170 using RTP/AVP protocol, with two codecs: PCMU (payload type 0) and PCMA (payload type 8).
    • m=video 51372 RTP/AVP 31 34: This specifies a video stream on port 51372 using RTP/AVP protocol, with two codecs: H.263 (payload type 31) and H.264 (payload type 34).
  • a (Attribute): Specifies additional attributes for the session. In this case, it specifies the RTP payload type for each codec (0 for PCMU, 8 for PCMA, and 101 for telephone-event), the codec names (PCMU, PCMA, and telephone-event), and the sampling rate (8000). The last line specifies the allowable range of event codes for the telephone-event codec. there can be multiple attribute fields in an SDP message
  • a=rtpmap:0 PCMU/8000: This specifies the codec for payload type 0 as PCMU with a sampling rate of 8000 Hz.
  • a=rtpmap:31 H263/90000: This specifies the codec for payload type 31 as H.263 with a clock rate of 90000 Hz.
  • a=sendrecv: This specifies that the media stream is bi-directional (send and receive).
  • a=rtcp-mux: This specifies that the RTP and RTCP packets are multiplexed on the same port.

 

RTPMAP

In the context of SDP, RTPMAP (RTP Mapping) is an attribute used to map a particular codec to an RTP payload type number. RTP is the Real-time Transport Protocol, which is used to transmit audio and video over IP networks. RTP uses payload types to identify the format of the data being transmitted.

The RTPMAP attribute provides a way for the sender to signal the receiver about the payload type numbers and the corresponding codecs being used. The attribute specifies the encoding name, the clock rate of the codec, and the number of audio channels being transmitted.

The syntax for the RTPMAP attribute is as follows:

a=rtpmap:<payload type> <encoding name>/<clock rate>[/<encoding parameters>]

where:

  • <payload type> is an integer value representing the payload type number.
  • <encoding name> is a string that identifies the codec being used, such as "PCMU" for G.711 mu-law audio or "H264" for H.264 video.
  • <clock rate> is an integer value indicating the clock rate of the codec, in Hz.
  • <encoding parameters> is an optional string that specifies additional parameters for the codec, such as packetization mode or frame size.

Here is an example of an RTPMAP attribute for G.711 mu-law audio:

a=rtpmap:0 PCMU/8000

This specifies that payload type 0 is used for G.711 mu-law audio, with a clock rate of 8000 Hz.

Overall, RTPMAP is an important attribute in SDP as it allows the sender and receiver to negotiate and agree upon the codec to be used for transmitting the media.

All RTPMAP codecs


Payload type (PT) Name Type No. of channels Clock rate (Hz)[note 1] Frame size (byte) Default packet interval (ms) Description References
0 PCMU audio 1 8000 any 20 ITU-T G.711 PCM μ-Law audio 64 kbit/s RFC 3551
1 reserved (previously FS-1016 CELP) audio 1 8000     reserved, previously FS-1016 CELP audio 4.8 kbit/s RFC 3551, previously RFC 1890
2 reserved (previously G721 or G726-32) audio 1 8000     reserved, previously ITU-T G.721 ADPCM audio 32 kbit/s or ITU-T G.726 audio 32 kbit/s RFC 3551, previously RFC 1890
3 GSM audio 1 8000 20 20 European GSM Full Rate audio 13 kbit/s (GSM 06.10) RFC 3551
4 G723 audio 1 8000 30 30 ITU-T G.723.1 audio RFC 3551
5 DVI4 audio 1 8000 any 20 IMA ADPCM audio 32 kbit/s RFC 3551
6 DVI4 audio 1 16000 any 20 IMA ADPCM audio 64 kbit/s RFC 3551
7 LPC audio 1 8000 any 20 Experimental Linear Predictive Coding audio 5.6 kbit/s RFC 3551
8 PCMA audio 1 8000 any 20 ITU-T G.711 PCM A-Law audio 64 kbit/s RFC 3551
9 G722 audio 1 8000[note 2] any 20 ITU-T G.722 audio 64 kbit/s RFC 3551 - Page 14
10 L16 audio 2 44100 any 20 Linear PCM 16-bit Stereo audio 1411.2 kbit/s,[2][3][4] uncompressed RFC 3551, Page 27
11 L16 audio 1 44100 any 20 Linear PCM 16-bit audio 705.6 kbit/s, uncompressed RFC 3551, Page 27
12 QCELP audio 1 8000 20 20 Qualcomm Code Excited Linear Prediction RFC 2658, RFC 3551
13 CN audio 1 8000     Comfort noise. Payload type used with audio codecs that do not support comfort noise as part of the codec itself such as G.711, G.722.1, G.722, G.726, G.727, G.728, GSM 06.10, Siren, and RTAudio. RFC 3389
14 MPA audio 1, 2 90000 8–72   MPEG-1 or MPEG-2 audio only RFC 3551, RFC 2250
15 G728 audio 1 8000 2.5 20 ITU-T G.728 audio 16 kbit/s RFC 3551
16 DVI4 audio 1 11025 any 20 IMA ADPCM audio 44.1 kbit/s RFC 3551
17 DVI4 audio 1 22050 any 20 IMA ADPCM audio 88.2 kbit/s RFC 3551
18 G729 audio 1 8000 10 20 ITU-T G.729 and G.729a audio 8 kbit/s; Annex B is implied unless the annexb=no parameter is used RFC 3551, Page 20, RFC 3555, Page 15
19 reserved (previously CN) audio         reserved, previously comfort noise RFC 3551
25 CELLB video   90000     Sun CellB video[5] RFC 2029
26 JPEG video   90000     JPEG video RFC 2435
28 nv video   90000     Xerox PARC's Network Video (nv)[6][7] RFC 3551, Page 32
31 H261 video   90000     ITU-T H.261 video RFC 4587
32 MPV video   90000     MPEG-1 and MPEG-2 video RFC 2250
33 MP2T audio/video   90000     MPEG-2 transport stream RFC 2250
34 H263 video   90000     H.263 video, first version (1996) RFC 3551, RFC 2190
72–76 reserved           reserved because RTCP packet types 200–204 would otherwise be indistinguishable from RTP payload types 72–76 with the marker bit set RFC 3550, RFC 3551
77–95 unassigned           note that RTCP packet type 207 (XR, Extended Reports) would be indistinguishable from RTP payload types 79 with the marker bit set RFC 3551, RFC 3611
dynamic H263-1998 video   90000     H.263 video, second version (1998) RFC 3551, RFC 4629, RFC 2190
dynamic H263-2000 video   90000     H.263 video, third version (2000) RFC 4629
dynamic (or profile) H264 AVC video   90000     H.264 video (MPEG-4 Part 10) RFC 6184, previously RFC 3984
dynamic (or profile) H264 SVC video   90000     H.264 video RFC 6190
dynamic (or profile) H265 video   90000     H.265 video (HEVC) RFC 7798
dynamic (or profile) theora video   90000     Theora video draft-barbato-avt-rtp-theora
dynamic iLBC audio 1 8000 20, 30 20, 30 Internet low Bitrate Codec 13.33 or 15.2 kbit/s RFC 3952
dynamic PCMA-WB audio 1 16000 5   ITU-T G.711.1 A-law RFC 5391
dynamic PCMU-WB audio 1 16000 5   ITU-T G.711.1 μ-law RFC 5391
dynamic G718 audio   32000 (placeholder) 20   ITU-T G.718 draft-ietf-payload-rtp-g718
dynamic G719 audio (various) 48000 20   ITU-T G.719 RFC 5404
dynamic G7221 audio   16000, 32000 20   ITU-T G.722.1 and G.722.1 Annex C RFC 5577
dynamic G726-16 audio 1 8000 any 20 ITU-T G.726 audio 16 kbit/s RFC 3551
dynamic G726-24 audio 1 8000 any 20 ITU-T G.726 audio 24 kbit/s RFC 3551
dynamic G726-32 audio 1 8000 any 20 ITU-T G.726 audio 32 kbit/s RFC 3551
dynamic G726-40 audio 1 8000 any 20 ITU-T G.726 audio 40 kbit/s RFC 3551
dynamic G729D audio 1 8000 10 20 ITU-T G.729 Annex D RFC 3551
dynamic G729E audio 1 8000 10 20 ITU-T G.729 Annex E RFC 3551
dynamic G7291 audio   16000 20   ITU-T G.729.1 RFC 4749
dynamic GSM-EFR audio 1 8000 20 20 ITU-T GSM-EFR (GSM 06.60) RFC 3551
dynamic GSM-HR-08 audio 1 8000 20   ITU-T GSM-HR (GSM 06.20) RFC 5993
dynamic (or profile) AMR audio (various) 8000 20   Adaptive Multi-Rate audio RFC 4867
dynamic (or profile) AMR-WB audio (various) 16000 20   Adaptive Multi-Rate Wideband audio (ITU-T G.722.2) RFC 4867
dynamic (or profile) AMR-WB+ audio 1, 2 or omit 72000 13.3–40   Extended Adaptive Multi Rate – WideBand audio RFC 4352
dynamic (or profile) vorbis audio (various) (various)     Vorbis audio RFC 5215
dynamic (or profile) opus audio 1, 2 48000[note 3] 2.5–60 20 Opus audio RFC 7587
dynamic (or profile) speex audio 1 8000, 16000, 32000 20   Speex audio RFC 5574
dynamic mpa-robust audio 1, 2 90000 24–72   Loss-Tolerant MP3 audio RFC 5219 (previously RFC 3119)
dynamic (or profile) MP4A-LATM audio   90000 or others     MPEG-4 Audio (includes AAC) RFC 6416 (previously RFC 3016)
dynamic (or profile) MP4V-ES video   90000 or others     MPEG-4 Visual RFC 6416 (previously RFC 3016)
dynamic (or profile) mpeg4-generic audio/video   90000 or other     MPEG-4 Elementary Streams RFC 3640
dynamic VP8 video   90000     VP8 video RFC 7741
dynamic VP9 video   90000     VP9 video draft-ietf-payload-vp9
dynamic L8 audio (various) (various) any 20 Linear PCM 8-bit audio with 128 offset RFC 3551 Section 4.5.10 and Table 5
dynamic DAT12 audio (various) (various) any 20 (by analogy with L16) IEC 61119 12-bit nonlinear audio RFC 3190 Section 3
dynamic L16 audio (various) (various) any 20 Linear PCM 16-bit audio RFC 3551 Section 4.5.11, RFC 2586
dynamic L20 audio (various) (various) any 20 (by analogy with L16) Linear PCM 20-bit audio RFC 3190 Section 4
dynamic L24 audio (various) (various) any 20 (by analogy with L16) Linear PCM 24-bit audio RFC 3190 Section 4
dynamic raw video   90000     Uncompressed Video RFC 4175
dynamic ac3 audio (various) 32000, 44100, 48000     Dolby AC-3 audio RFC 4184
dynamic eac3 audio (various) 32000, 44100, 48000     Enhanced AC-3 audio RFC 4598
dynamic t140 text   1000     Text over IP RFC 4103
dynamic EVRC
EVRC0
EVRC1		 audio   8000     EVRC audio RFC 4788
dynamic EVRCB
EVRCB0
EVRCB1		 audio   8000     EVRC-B audio RFC 4788
dynamic EVRCWB
EVRCWB0
EVRCWB1		 audio   16000     EVRC-WB audio RFC 5188
dynamic jpeg2000 video   90000     JPEG 2000 video RFC 5371
dynamic UEMCLIP audio   8000, 16000     UEMCLIP audio RFC 5686
dynamic ATRAC3 audio   44100     ATRAC3 audio RFC 5584
dynamic ATRAC-X audio   44100, 48000     ATRAC3+ audio RFC 5584
dynamic ATRAC-ADVANCED-LOSSLESS audio   (various)     ATRAC Advanced Lossless audio RFC 5584
dynamic DV video   90000     DV video RFC 6469 (previously RFC 3189)
dynamic BT656 video         ITU-R BT.656 video RFC 3555
dynamic BMPEG video         Bundled MPEG-2 video RFC 2343
dynamic SMPTE292M video         SMPTE 292M video RFC 3497
dynamic RED audio         Redundant Audio Data RFC 2198
dynamic VDVI audio         Variable-rate DVI4 audio RFC 3551
dynamic MP1S video         MPEG-1 Systems Streams video RFC 2250
dynamic MP2P video         MPEG-2 Program Streams video RFC 2250
dynamic tone audio   8000 (default)     tone RFC 4733
dynamic telephone-event audio   8000 (default)     DTMF tone RFC 4733
dynamic aptx audio 2 – 6 (equal to sampling rate) 4000 ÷ sample rate 4[note 4] aptX audio RFC 7310
dynamic jxsv video   90000     JPEG XS video RFC 9134

 

FMTP

The "fmtp" (format parameters) attribute in SDP (Session Description Protocol) is used to describe the format parameters of a media stream, specifically for codecs that support dynamic negotiation of parameters. It specifies the media format parameters as a set of parameter name-value pairs, separated by semicolons.

The format of the "fmtp" attribute is as follows:

a=fmtp:<format> <parameter_name>=<value>;<parameter_name>=<value>;...
  • <format>: specifies the media format to which the parameters apply.
  • <parameter_name>: specifies the name of the parameter.
  • <value>: specifies the value of the parameter.

For example, the "fmtp" attribute for the H.264 video codec may look like:

a=fmtp:120 profile-level-id=42801E; packetization-mode=1; sprop-parameter-sets=Z0IACpZTBYmI,aMljiA==

In this example, the <format> value is 120 (which is the format number for H.264). The <parameter_name> values include "profile-level-id", "packetization-mode", and "sprop-parameter-sets", and their corresponding <value>s are "42801E", "1", and "Z0IACpZTBYmI,aMljiA==", respectively. These parameters provide information such as the profile and level of the codec, the packetization mode, and the parameter sets required for decoding the video stream.

The "fmtp" attribute is typically used in conjunction with the "rtpmap" attribute to describe media formats that use RTP (Real-time Transport Protocol) for transmission.

 

This is a SIP response to an INVITE request shown above. It has a status code of 200 OK, indicating that the request was successful. The Via header indicates the path that the request followed. The To header contains the tag that identifies the current dialog. The From header contains the tag that identifies the previous dialog. The Call-ID header is used to identify the call. The CSeq header contains the sequence number and method of the request. The Contact header specifies the address where the response should be sent. The Content-Type header indicates that the body of the message is in SDP format. The SDP message body includes connection information, timing information, and media information for both audio and video. In this response, some codecs have been removed from both audio and video streams.

SIP/2.0 200 OK
Via: SIP/2.0/UDP pc33.atlanta.com;branch=z9hG4bK776asdhds
To: Bob <sip:[email protected]>;tag=2482893830n
From: Alice <sip:[email protected]>;tag=1928301774
Call-ID: a84b4c76e66710
CSeq: 314159 INVITE
Contact: <sip:[email protected]:5060>
Content-Type: application/sdp
Content-Length: 204

v=0
o=- 123456 789012 IN IP4 192.0.2.4
s=SIP Call
c=IN IP4 192.0.2.1
t=0 0
m=audio 49170 RTP/AVP 0 18
a=rtpmap:0 PCMU/8000
a=rtpmap:18 G729/8000
a=fmtp:18 annexb=no
m=video 49172 RTP/AVP 31
a=rtpmap:31 H261/90000

  • v=0: The protocol version being used is 0.

  • o=- 123456 789012 IN IP4 pc33.atlanta.com: This is the origin field. The - indicates that the session is not valid for further communication. 123456 is the session ID, which is chosen arbitrarily. 789012 is the session version number. IN stands for "Internet", and IP4 pc33.atlanta.com is the address of the host that originated the session.

  • s=SIP Call: The session name, which is an optional field. In this case, it is "SIP Call".

  • c=IN IP4 192.0.2.1: The connection information for the session. This field specifies the network type, address type, and address information. In this case, it indicates that the session is using the Internet network, IPv4 address, and the address is 192.0.2.1.

  • t=0 0: The timing information for the session. The first 0 indicates the start time of the session, which is "now". The second 0 indicates the session will not terminate automatically.

  • m=audio 49170 RTP/AVP 0 18: This is the media description for the audio stream. audio indicates that this is an audio stream. 49170 is the port number being used for the RTP stream. RTP/AVP specifies the protocol being used (RTP) and the payload format for the audio data (AVP). 0 and 18 are the RTP payload types being used for this stream.

  • a=rtpmap:0 PCMU/8000: This attribute maps the payload type 0 to the audio codec PCMU and indicates that the audio is sampled at a rate of 8000 Hz.

  • a=rtpmap:18 G729/8000: This attribute maps the payload type 18 to the audio codec G.729 and indicates that the audio is sampled at a rate of 8000 Hz.

  • a=fmtp:18 annexb=no: This is an optional attribute that provides additional parameters for the G.729 codec. In this case, it specifies that the codec is not using the Annex B format.

  • m=video 49172 RTP/AVP 31: This is the media description for the video stream. video indicates that this is a video stream. 49172 is the port number being used for the RTP stream. RTP/AVP specifies the protocol being used (RTP) and the payload format for the video data (AVP). 31 is the RTP payload type being used for this stream.

  • a=rtpmap:31 H261/90000: This attribute maps the payload type 31 to the video codec H.261 and indicates that the video is sampled at a rate of 90000 Hz.

Multiple M Lines

 

In SIP SDP, multiple "m=" lines can be included in a single SDP message to provide information about multiple media streams within a session. Each "m=" line corresponds to a different type of media, such as audio, video, or data. The "m=" line specifies the transport protocol and port number to be used for that particular media stream, as well as the format of the media data being transmitted. Each "m=" line can also have its own set of media-level attributes, such as the RTP payload types being used for different codecs, or the clock rate of the media data. By including multiple "m=" lines in a single SDP message, a SIP session can support multiple media streams simultaneously.

image.png

Back to top