Re: Decoding G.729 using Ethereal, rtpdump and VoiceAgedecoder

Gabriel,

As mentioned in that doc, the G729 Codec Sample size it is 10bytes
and those 10bytes represent 10ms of voice. Then if you put one, two,
three or more of those 10byte samples in one RTP packet it is a
different story. The default it is to have two samples per RTP packet
(so you have a 20ms packetization period), but you can have three or
more in one RTP. The most common packetization periods for G729 are
20ms, 40ms, 60ms or 80ms. Of course when you add more G729 samples in
one RTP packet, you save bandwidth because there is less IP/UDP/RTP
header involved.

Regards

Alejandro

Gabriel M. wrote:

I came
across this article ( http://www.cisco.com/warp/public/788/pkt-voice-general/bwidth_consume.html )
that states there is a non-default version of g.729 which uses 30ms
packetization period, thus the RTP payload is 30bytes (I have just
noticed that
now, I was under the impression that the payload was 20bytes).

Any idea
what to do in this case?

----------

From:
[email protected]
[ mailto:[email protected] ] On
Behalf Of Alejandro Vaquero

Sent: Tuesday, July
25, 2006 4:41
AM

To: Ethereal
development

Subject: Re:
[Ethereal-dev]
Decoding G.729 using Ethereal, rtpdump and VoiceAgedecoder

Looks you are close because
you get some voice
patterns. If the RTP packetization period is 20ms, then there are two
10byte
samples of G729 encoded data per RTP packet, so make sure when you run
the
"rtpdump" the samples are in the correct order when generating the
bitstream.

Gabriel M. wrote:

Hi,

I am trying to decode a
VoIP conversation using G.729
encoding. I have read older messages on this list (and others) but I am
still
not able to dump a listenable .au file.

The steps that I made so
far:

capture
the traffic using Ethereal and dumping the RTP data (Statistics ->
RTP ->
Show all streams -> SaveAs).

Use
rtpdump ( http://www.cs.columbia.edu/IRT/software/rtptools/ )
to dump only the audio payload (“rtpdump -F payload -f in.rtp -o
out.rtp”)

Use
self-made code (see the code at the end of the e-mail) to convert the
rtpdump
into the bitstream expected by the decoder

Decode
the bitstream using the VoiceAge decoder (decoder out.rtp test2.au)

Play
the test2.au using GoldWave, (with all possible combinations between
16bit
signed/unsigned 8000Hz)

Unfortunately all I can
hear is some weird sound ( I
can see voice patterns in GoldWave, but the sound is metalic, almost
alien-like). Am I doing something wrong ?

References:

http://www.ethereal.com/lists/ethereal-dev/200505/msg00399.html

http://www.dsprelated.com/groups/speechcoding/show/666.php

//rtp_bitstream.c

[code]

#include <stdio.h>

int main(int argc, char
**argv)

{

char c;

FILE *f=fopen("in.rtp", "r");

FILE *fileout = fopen("out3.rtp","w");

long lSize;

char * buffer, * out;

char qbyte, q;

int i, j =0, k;

fseek (f , 0 ,
SEEK_END);

lSize = ftell (f);

rewind (f);

buffer = (char*) malloc (lSize);

out = (char*) malloc (lSize * 16);

fread (buffer,1,lSize,f);

for (i = 0; i<lSize; i++)

{

qbyte = buffer[i];

for (k=0;k<8;k++)

{

q = qbyte & 1;

out[j] = 0x00;

j++;

if (q == 0)

out[j] = 0x71;

else

out[j] = 0x81;

j++;

qbyte >>= 1;

}

}

fwrite (out , 1 , lSize * 16, fileout);

return 0;

}

[/code]

Best regards,

Gabriel.

----------

_______________________________________________

Ethereal-dev mailing list

[email protected]

http://www.ethereal.com/mailman/listinfo/ethereal-dev

----------

_______________________________________________
Ethereal-dev mailing list
[email protected]
http://www.ethereal.com/mailman/listinfo/ethereal-dev

_______________________________________________
Ethereal-dev mailing list
[email protected]
http://www.ethereal.com/mailman/listinfo/ethereal-dev