Wireshark · Ethereal-dev: RE: [Ethereal-dev] RTP Graph analysis

["Anders Broman (AL/EAB)" <anders.broman@xxxxxxxxxxxx>]

Hi,
I get the following compiler warnings:
ethereal-tap-register.c
rtp_analysis.c
rtp_analysis.c(444) : warning C4101: 'time_delta' : unreferenced local variable
rtp_analysis.c(1242) : warning C4018: '<' : signed/unsigned mismatch

Best regards
Anders

-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx]On Behalf Of Alejandro Vaquero
Sent: den 23 september 2004 04:31
To: 'Ethereal development'
Subject: RE: [Ethereal-dev] RTP Graph analysis


Hi,
	Attached is a new version with this fixed.

	It looks gtk1 is calling on_notebook_switch_page (the function
for the "switch_page" signal) when the windows is destroy. At that point
user_data was also free that caused the exception. Not sure why the
signal is activated when the windows is destroy (this doesn't happened
in gtk2). Anyway, I make a workaround to disconnect this signal before
the windows is destroyed to fix this issue.

	Please let me know of any other issue.

Thanks and Regards
Alejandro


-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx] On Behalf Of Anders Broman
(AL/EAB)
Sent: Tuesday, September 21, 2004 8:03 AM
To: 'Ethereal development'
Subject: RE: [Ethereal-dev] RTP Graph analysis


Hi,
I get a crash when closing the graph window using gtk1 on windows, MSVC
points to:

static void on_notebook_switch_page(GtkNotebook *notebook _U_,
                                    GtkNotebookPage *page _U_,
                                    gint page_num _U_,
                                    user_data_t *user_data _U_)
{
	user_data->dlg.selected_clist =
=>		(page_num==0) ? user_data->dlg.clist_fwd :
user_data->dlg.clist_rev ;
	user_data->dlg.selected_row = 0;
}

Best regards
Anders
-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx]On Behalf Of Alejandro Vaquero
Sent: den 17 september 2004 16:42
To: 'ronnie sahlberg'; 'Ethereal development'
Subject: RE: [Ethereal-dev] RTP Graph analysis


Hi Ronnie,
	I also tested it using Linux FC2 and with different captures
files.

Thanks
Alejandro 
-----Original Message-----
From: ethereal-dev-bounces@xxxxxxxxxxxx
[mailto:ethereal-dev-bounces@xxxxxxxxxxxx] On Behalf Of ronnie sahlberg
Sent: Friday, September 17, 2004 3:41 AM
To: Ethereal development
Subject: Re: [Ethereal-dev] RTP Graph analysis

Have any of the telco voip people tested this patch yet?
If no one objects to it  I will test it and check it in in a few days.




----- Original Message -----
From: Alejandro Vaquero <alejandrovaquero@xxxxxxxxx>
Date: Mon, 6 Sep 2004 16:20:25 -0600
Subject: [Ethereal-dev] RTP Graph analysis
To: ethereal-dev@xxxxxxxxxxxx




Hi All,

            I made the following enhancements to the RTP analysis tool
(rtp_analysis.c):

 

1) Added a graphic analysis of RTP packets (see
RTP_graph_analysis2.jpg for a screenshot). This graph displays the
forward and reverse RTP packets over the time in the x axis, and the
"jitter" and "difference" in the y axis. Where:

            - jitter: is the mean deviation of the difference in
packet spacing at the receiver compared to the sender (as per RTP
RFC1889). Basically this is the "average jitter".

            - difference: is the time difference in two consecutive
packets at the receiver compared to the sender. This is basically the
"jitter" between two consecutive packets (the time difference when the
packet actually arrived and when it should theory arrive). This is the
value that is more important to understand when there are voice
problems. For example, if a device has a "jitter buffer" of 100ms, and
you see that the "difference" is 150ms, that means there are a couple
of packets that the device will drop because it doesn't feet in his
"jitter buffer".

            

            This graph interface is based on io_stat.c and tested in
Windows 2000 using GTK 1 and 2.

 

2) The graphic also display an "x" at the bottom of the x-axis if
there was a "wrong sequence number" between two consecutives packets.
This usually means a packet lost, or out of order packets. And will
display an "m" for packets with the RTP "Mark set". This means the end
of a silence period.

 

3) Fixed a bug where the reverse RTP packets were not displayed in the
"RTP analysis tool" when using "Stream Analysis." menu option.

 

4) Display the "Delay" and "Jitter" column in "ms" instead of "sec".
And added the percentage of packet lost to the statistic.

 

            Attached is a screenshot and the "diff" of rtp_analysis.c.

 

            Comments/suggestions are very welcome.

 

Thanks and Regards

Alejandro

 

_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev

_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev

_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev

_______________________________________________
Ethereal-dev mailing list
Ethereal-dev@xxxxxxxxxxxx
http://www.ethereal.com/mailman/listinfo/ethereal-dev