Home

Awesome

rtpdump

Extract media files from RTP streams in pcap format

codec support

This program is intended to support usual audio/video codecs used on IMS networks (VoLTE/VoWiFi).
Therefore, some codecs might be limited to usual scenarios on these networks.

Payload TypeSupport
1-23 NAL UnitYes
24 STAP-ANo - planned
25 STAP-BNo
26 MTAP16No
27 MTAP24Yes
28 FU-AYes
29 FU-BNo

ipsec support

In order to support dumping VoWiFi media some support for ESP (Encapsulating Security Payload) decryption is present.

Encryption AlgorithmSupport
3DES CBCYes
DES CBCNo - Planned
AES CBCNo - Planned

Keys are read from file 'esp-keys.txt' on the current directory by default. One key per file, for example:

[SPI] [Encryption Algorithm] [Key]
0x00d40016 des3_cbc 0x091199869ec18afd8e38f77eb1252685924937d3921a178e
0xcb97da43 des3_cbc 0xaaa316cd3fa41daa9afe6e8f42a9ae0ce2bd5128cef5a60f

Global flag -k can be used to indicate another key file path. Check -help.

replaying

Its possible to replay a RTP stream, specifying the destination host and port. The stream consumer can be a actual mobile handset or any application that can interpret RTP streams (e.g VLC).

The stream is replayed as is, taking into account the original timestamps in the pcap file and mantaining the original RTP payload type. It's up to the receiver to interpret the appropriate stream codec.

For example, VLC accepts a SDP input file:

v=0
c=IN IP4 127.0.0.1
m=audio 1234 RTP/AVP 99
a=rtpmap:99 AMR/8000

rtpdump play --host localhost --port 1234 [pcap containing amr-nb payload type 99]

usage

compiling

Checkout gopacket. Linux should be straightforward.
For Windows, make sure mingw(32/64) toolchain is on PATH for gopacket WinPcap dependency. Install WinPcap on standard location C:\WpdPack

planned features

  1. Better automation and scription support
  2. Include stream analisys, packets lost, jitter, etc
  3. Media player directly from pcap. ffmpeg support.
  4. Jitter buffer to simulate original condition, i.e. packet loss due to jitter
  5. Support multiple speach frames in audio packet

contributions

Are always appreciated.