Reverse Engineer RTP Parameters

[Sebastian de Castelberg]

Hi list,

I'm trying to reverse engineer a RTP audio stream from and to a
proprietary device. The data is sent in UDP packets to a multicast
address (see the enclosed wireshark dump [1])
With the following gst pipeline I was able to hear the audio coming
from the device. (The device sends data to 232.0.2.1 and receives data
from 232.0.2.2)

  # 'client' pipeline
  gst-launch-0.10 udpsrc multicast-group=232.0.2.1 auto-multicast=true
port=49253 \
  multicast-iface="eth0" caps="application/x-rtp,
media=(string)audio,clock-rate=(int)8000,encoding-name=(string)L16" \
  ! rtpL16depay ! alsasink

I haven't yet found a gstreamer setup for the opposite way, streaming
to the device. The following pipeline just produced a crackling noise
on the device:

  # 'server' pipeline
  gst-launch-0.10 audiotestsrc ! audioconvert ! \
  audio/x-raw-int,depth=16,rate=8000 ! rtpL16pay ! udpsink
host=232.0.2.2 port=49253

The output changes to a constant noise when adding channels=8 to the
caps string.

Based on the attached dumps [1], [2], do these pipelines make sense?
What combination would you suggest for the server pipeline? How can I
modify the RTP payload type and timing?

Thanks for your help
Sebastian

#########################################################
[1]: Wireshark dump proprietary device (2 packets)
#########################################################

No.     Time        Source                Destination           Protocol Info
      8 0.026451    <localip>          <multicastip>             RTP
   PT=DynamicRTP-Type-98, SSRC=0x7D9AFD54, Seq=3, Time=240

Real-Time Transport Protocol
    10.. .... = Version: RFC 1889 Version (2)
    ..0. .... = Padding: False
    ...0 .... = Extension: False
    .... 0000 = Contributing source identifiers count: 0
    0... .... = Marker: False
    Payload type: DynamicRTP-Type-98 (98)
    Sequence number: 3
    Timestamp: 240
    Synchronization Source identifier: 0x7d9afd54 (2107309396)
    Payload: e7ff1900f8ffdeff19004400d6ff3b003b00bcff19000000...

No.     Time        Source                Destination           Protocol Info
     10 0.036594    <localip>          <multicastip>             RTP
   PT=DynamicRTP-Type-98, SSRC=0x7D9AFD54, Seq=4, Time=320

Real-Time Transport Protocol
    10.. .... = Version: RFC 1889 Version (2)
    ..0. .... = Padding: False
    ...0 .... = Extension: False
    .... 0000 = Contributing source identifiers count: 0
    0... .... = Marker: False
    Payload type: DynamicRTP-Type-98 (98)
    Sequence number: 4
    Timestamp: 320
    Synchronization Source identifier: 0x7d9afd54 (2107309396)
    Payload: 88000000a9001100effff8ffd6ffbcff35ffa3ff78ff67ff...

################################################################
[2]: Wireshark dump reverse engineered gstreamer 'server' pipeline (2 packets)
###############################################################

No.     Time        Source                Destination           Protocol Info
   2498 58.790631   192.168.0.30          232.0.2.1             RTP
  PT=DynamicRTP-Type-96, SSRC=0x37E90F89, Seq=34530, Time=2083047297

Real-Time Transport Protocol
    10.. .... = Version: RFC 1889 Version (2)
    ..0. .... = Padding: False
    ...0 .... = Extension: False
    .... 0000 = Contributing source identifiers count: 0
    0... .... = Marker: False
    Payload type: DynamicRTP-Type-96 (96)
    Sequence number: 34530
    Timestamp: 2083047297
    Synchronization Source identifier: 0x37e90f89 (938020745)
    Payload: dd50000022b0414658246496652459bc43b825b20337e05c...

No.     Time        Source                Destination           Protocol Info
   2501 58.831841   192.168.0.30          232.0.2.1             RTP
  PT=DynamicRTP-Type-96, SSRC=0x37E90F89, Seq=34531, Time=2083047627

Real-Time Transport Protocol
    10.. .... = Version: RFC 1889 Version (2)
    ..0. .... = Padding: False
    ...0 .... = Extension: False
    .... 0000 = Contributing source identifiers count: 0
    0... .... = Marker: False
    Payload type: DynamicRTP-Type-96 (96)
    Sequence number: 34531
    Timestamp: 2083047627
    Synchronization Source identifier: 0x37e90f89 (938020745)
    Payload: 398e52d8625666335dfa4aa52e7d0cd6e9a9c922af169e9d...
_______________________________________________
gstreamer-devel mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel

[Ole André Vadla Ravnås]

On Wednesday, September 28, 2011, Sebastian de Castelberg <[hidden email]> wrote:
(snip)
> I haven't yet found a gstreamer setup for the opposite way, streaming
> to the device. The following pipeline just produced a crackling noise
> on the device:
>
>  # 'server' pipeline
>  gst-launch-0.10 audiotestsrc ! audioconvert ! \
>  audio/x-raw-int,depth=16,rate=8000 ! rtpL16pay ! udpsink
> host=232.0.2.2 port=49253

Add "is-live=TRUE" right after audiotestsrc, otherwise you'll be flooding your device with packets.

Cheers,
Ole André

_______________________________________________
gstreamer-devel mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel

[Sebastian de Castelberg]

hi Ole,

thanks for your reply. I added the is-live parameter to the
audiotestsrc. The audio output on the device is still chopped and a
network trace does not show any change in the rtp stream.
In the meantime I've found out how to set a more specific timing on
the RTP stream by passing the following parameters to rtpL16pay:

  gst-launch-0.10 audiotestsrc is-live=TRUE ! audioconvert \
  ! audio/x-raw-int,depth=16,rate=8000 ! rtpL16pay pt=98 mtu=172
seqnum-offset=0 timestamp-offset=0 timestamp-offset=80 \
  ! udpsink host=232.0.2.2 port=49253

this results in the following RTP dump:

23:22:29.874071 IP 192.168.0.30.56674 > 232.0.2.2.49253: UDP, length 172
23:22:29.883607 IP 192.168.0.30.56674 > 232.0.2.2.49253: UDP, length 172
23:22:29.893871 IP 192.168.0.30.56674 > 232.0.2.2.49253: UDP, length 172
...

The timing and packet size seem to match the expectations from the
original dump. The Sequence number is being incremented, the Timestamp
increased by 80 and the time difference between two packets is ~0.01s.
Nonetheless, the sound is still choppy (peak, silence, ... , silence,
peak, ...).

To me, it seems that the payload is somehow broken. (encoding, timing,
...). Any advices on how to analyze the whole pipeline?
Here's a gst-launch -v output for the two pipes:

### working: device -> desktop ###
gst-launch-0.10 -v udpsrc multicast-group=232.0.2.1
auto-multicast=true port=49253 multicast-iface="eth0" ! capsfilter
caps="application/x-rtp,
media=(string)audio,clock-rate=(int)8000,encoding-name=(string)L16,payload=(int)98"
! rtpL16depay ! audioconvert ! audioresample ! audio/x-raw-int,
rate=8000 ! alsasink
Setting pipeline to PAUSED ...
Pipeline is live and does not need PREROLL ...
Setting pipeline to PLAYING ...
New clock: GstSystemClock
/GstPipeline:pipeline0/GstCapsFilter:capsfilter0.GstPad:src: caps =
application/x-rtp, media=(string)audio, clock-rate=(int)8000,
encoding-name=(string)L16, payload=(int)98
/GstPipeline:pipeline0/GstRtpL16Depay:rtpl16depay0.GstPad:src: caps =
audio/x-raw-int, endianness=(int)4321, signed=(boolean)true,
width=(int)16, depth=(int)16, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstRtpL16Depay:rtpl16depay0.GstPad:sink: caps =
application/x-rtp, media=(string)audio, clock-rate=(int)8000,
encoding-name=(string)L16, payload=(int)98
/GstPipeline:pipeline0/GstAudioConvert:audioconvert0.GstPad:src: caps
= audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstAudioConvert:audioconvert0.GstPad:sink: caps
= audio/x-raw-int, endianness=(int)4321, signed=(boolean)true,
width=(int)16, depth=(int)16, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstAudioResample:audioresample0.GstPad:src:
caps = audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstAudioResample:audioresample0.GstPad:sink:
caps = audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstCapsFilter:capsfilter1.GstPad:src: caps =
audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstCapsFilter:capsfilter1.GstPad:sink: caps =
audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >
/GstPipeline:pipeline0/GstAlsaSink:alsasink0.GstPad:sink: caps =
audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1,
channel-positions=(GstAudioChannelPosition)<
GST_AUDIO_CHANNEL_POSITION_FRONT_MONO >

### not working: desktop -> device (noisy/tinny chunks) ###
akureyri :: ~ > gst-launch-0.10 -v audiotestsrc is-live=TRUE !
audio/x-raw-int,rate=8000,channels=1 ! audioresample ! audioconvert !
rtpL16pay pt=98 mtu=172 seqnum-offset=0 timestamp-offset=0
timestamp-offset=80 ! udpsink host=232.0.2.2 port=49253
Setting pipeline to PAUSED ...
/GstPipeline:pipeline0/GstAudioTestSrc:audiotestsrc0.GstPad:src: caps
= audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
Pipeline is live and does not need PREROLL ...
Setting pipeline to PLAYING ...
New clock: GstSystemClock
/GstPipeline:pipeline0/GstCapsFilter:capsfilter0.GstPad:src: caps =
audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstCapsFilter:capsfilter0.GstPad:sink: caps =
audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstAudioResample:audioresample0.GstPad:src:
caps = audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstAudioResample:audioresample0.GstPad:sink:
caps = audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstAudioConvert:audioconvert0.GstPad:src: caps
= audio/x-raw-int, endianness=(int)4321, signed=(boolean)true,
width=(int)16, depth=(int)16, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstAudioConvert:audioconvert0.GstPad:sink: caps
= audio/x-raw-int, endianness=(int)1234, signed=(boolean)true,
width=(int)32, depth=(int)32, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstRtpL16Pay:rtpl16pay0.GstPad:src: caps =
application/x-rtp, media=(string)audio, clock-rate=(int)8000,
encoding-name=(string)L16, encoding-params=(string)1, channels=(int)1,
payload=(int)98, ssrc=(uint)3298703621, clock-base=(uint)80,
seqnum-base=(uint)0
/GstPipeline:pipeline0/GstRtpL16Pay:rtpl16pay0.GstPad:sink: caps =
audio/x-raw-int, endianness=(int)4321, signed=(boolean)true,
width=(int)16, depth=(int)16, rate=(int)8000, channels=(int)1
/GstPipeline:pipeline0/GstRtpL16Pay:rtpl16pay0: timestamp = 80
/GstPipeline:pipeline0/GstRtpL16Pay:rtpl16pay0: seqnum = 0
/GstPipeline:pipeline0/GstUDPSink:udpsink0.GstPad:sink: caps =
application/x-rtp, media=(string)audio, clock-rate=(int)8000,
encoding-name=(string)L16, encoding-params=(string)1, channels=(int)1,
payload=(int)98, ssrc=(uint)3298703621, clock-base=(uint)80,
seqnum-base=(uint)0

Thanks
Sebastian

On Wed, Sep 28, 2011 at 10:12 PM, Ole André Vadla Ravnås
<[hidden email]> wrote:

> On Wednesday, September 28, 2011, Sebastian de Castelberg
> <[hidden email]> wrote:
> (snip)
>> I haven't yet found a gstreamer setup for the opposite way, streaming
>> to the device. The following pipeline just produced a crackling noise
>> on the device:
>>
>>  # 'server' pipeline
>>  gst-launch-0.10 audiotestsrc ! audioconvert ! \
>>  audio/x-raw-int,depth=16,rate=8000 ! rtpL16pay ! udpsink
>> host=232.0.2.2 port=49253
>
> Add "is-live=TRUE" right after audiotestsrc, otherwise you'll be flooding
> your device with packets.
>
> Cheers,
> Ole André
>
> _______________________________________________
> gstreamer-devel mailing list
> [hidden email]
> http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel
>
>

_______________________________________________
gstreamer-devel mailing list
[hidden email]
http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel