Wireshark-users: Re: [Wireshark-users] Jitter wrong in wireshark?
From: "Anders Broman" <a.broman@xxxxxxxxx>
Date: Mon, 26 Feb 2007 21:03:48 +0100
Hi, I've added a note on RTP timestamp, please review. Best regards Anders -----Ursprungligt meddelande----- Från: wireshark-users-bounces@xxxxxxxxxxxxx [mailto:wireshark-users-bounces@xxxxxxxxxxxxx] För Lars Ruoff Skickat: den 26 februari 2007 14:46 Till: Community support list for Wireshark Ämne: Re: [Wireshark-users] Jitter wrong in wireshark? Hi Anders, since this too is a recurring question, perhaps you (or someone else) could add it to the Wiki, just under what i added last week: http://wiki.wireshark.org/RTP_statistics (bottom). (I'm sorry, i don't have the time right now). br, Lars Anders Broman wrote: > Hi, > Looking at the sources there is code there to use different frequencies > But for dynamic payload types there are two prerequisites that must be met: > - The setup signalling must be in the traces for Wireshark to track the > PT to the media type. > - The media type sampling frequency must be in the table of rtp_analysis.c > > This types are currently specified: > static const mimetype_and_clock mimetype_and_clock_map[] = { > {"AMR", 8000}, /* [RFC3267] */ > {"AMR-WB", 16000}, /* [RFC3267] */ > {"EVRC", 8000}, /* [RFC3558] */ > {"EVRC0", 8000}, /* [RFC3558] */ > {"G7221", 16000}, /* [RFC3047] */ > {"G726-16", 8000}, /* [RFC3551] */ > {"G726-24", 8000}, /* [RFC3551] */ > {"G726-32", 8000}, /* [RFC3551] */ > {"G726-40", 8000}, /* [RFC3551] */ > {"G729D", 8000}, /* [RFC3551] */ > {"G729E", 8000}, /* [RFC3551] */ > {"GSM-EFR", 8000}, /* [RFC3551] */ > {"mpa-robust", 90000}, /* [RFC3119] */ > {"SMV", 8000}, /* [RFC3558] */ > {"SMV0", 8000}, /* [RFC3558] */ > {"red", 1000}, /* [RFC4102] */ > {"t140", 1000}, /* [RFC4103] */ > {"BMPEG", 90000}, /* [RFC2343],[RFC3555] */ > {"BT656", 90000}, /* [RFC2431],[RFC3555] */ > {"DV", 90000}, /* [RFC3189] */ > {"H263-1998", 90000}, /* [RFC2429],[RFC3555] */ > {"H263-2000", 90000}, /* [RFC2429],[RFC3555] */ > {"MP1S", 90000}, /* [RFC2250],[RFC3555] */ > {"MP2P", 90000}, /* [RFC2250],[RFC3555] */ > {"MP4V-ES", 90000}, /* [RFC3016] */ > {"pointer", 90000}, /* [RFC2862] */ > {"raw", 90000}, /* [RFC4175] */ > }; > Best regards > Anders > > > > -----Ursprungligt meddelande----- > Från: wireshark-users-bounces@xxxxxxxxxxxxx > [mailto:wireshark-users-bounces@xxxxxxxxxxxxx] För Juan.Wortley@xxxxxxxxx > Skickat: den 24 februari 2007 17:45 > Till: wireshark-users@xxxxxxxxxxxxx > Ämne: Re: [Wireshark-users] Jitter wrong in wireshark? > > Hi Anders, > > Yes, your are right. Codecs affect jitter calculation because of the > sampling frequency. In practice, it affects when calculating the timestamp > to seconds. I.e: when multipling timestamp ticks to convert to seconds one > must use the frequency sampling of the current codec. As wireshark uses > always 0.000125 (it looks like this is not configurable), then if codec is > not G711 calculations will be wrong. > > And that´s my case :( > BR > Juan > > > > > > >> -----Original Message----- >> From: wireshark-users-bounces@xxxxxxxxxxxxx >> [mailto:wireshark-users-bounces@xxxxxxxxxxxxx] On Behalf Of >> ext Anders Broman >> Sent: Sábado, 24 de Febrero de 2007 04:50 a.m. >> To: 'Community support list for Wireshark' >> Subject: Re: [Wireshark-users] Jitter wrong in wireshark? >> >> Hi, >> If I understood the previous discussion correctly the RTP >> timestamp which is based on the sampling frequency is used in >> the calculations. If the wrong sampling frequency is used the >> calculations will be off. >> >> If a dynamic PT is used (>95) it probably not G711. >> Best regards >> Anders >> >> ________________________________________ >> Från: wireshark-users-bounces@xxxxxxxxxxxxx >> [mailto:wireshark-users-bounces@xxxxxxxxxxxxx] För >> Juan.Wortley@xxxxxxxxx >> Skickat: den 23 februari 2007 18:29 >> Till: wireshark-users@xxxxxxxxxxxxx >> Ämne: Re: [Wireshark-users] Jitter wrong in wireshark? >> >> Hi Anders, >> good question. I´m not sure but I guess is codec is PCMA G711 >> 8KHz (will take a look during call setup in a couple of hours). >> >> However "Payload type" is 97 (0x61), and wireshark shows it as >> "Payload type=Unknown" >> >> According to the formula, jitter calculation depends on >> timestamps and arrival time differences. I don´t know how >> codec could affect. >> >> BR >> Juan >> >> >> ________________________________________ >> From: wireshark-users-bounces@xxxxxxxxxxxxx >> [mailto:wireshark-users-bounces@xxxxxxxxxxxxx] On Behalf Of >> ext Anders Broman (AL/EAB) >> Sent: Viernes, 23 de Febrero de 2007 01:05 p.m. >> To: Community support list for Wireshark >> Subject: SV: [Wireshark-users] Jitter wrong in wireshark? >> Hi, >> Which codec is used? >> Best regards >> Anders >> >> ________________________________________ >> Från: wireshark-users-bounces@xxxxxxxxxxxxx genom >> Juan.Wortley@xxxxxxxxx >> Skickat: fr 2007-02-23 16:53 >> Till: wireshark-users@xxxxxxxxxxxxx >> Ämne: [Wireshark-users] Jitter wrong in wireshark? >> Hi All, >> Below is a rtp analysis from a wireshark 0.99.5 capture in a >> live network. >> Doing by hand jitter calculation it doesn´t match. And of >> course, if below data is real it means we are having jitter of >> more than 1 minute !! >> Exampe: first jitter should be 138.90/16 = 8.68125 [ms] and is showing >> 8803,82 [ms] !!! >> Do someone know what could be wrong or what I´m missing? >> BR >> Juan >> >> Forward Packet Sequence Delta (ms) Jitter (ms) >> IP BW (kbps) Marker Status Date Length >> 1 0 0.00 0.00 1.45 [ Ok ] 02/19/2007 >> 18:44:25.915 231 >> 2 1 138.90 8803.82 2.90 [ Ok ] 02/19/2007 >> 18:44:26.053 231 >> 3 2 119.59 17058.61 4.34 [ Ok ] >> 02/19/2007 >> 18:44:26.173 231 >> 4 3 139.67 24796.21 5.79 [ Ok ] >> 02/19/2007 >> 18:44:26.313 231 >> 5 4 180.22 32047.69 7.24 [ Ok ] >> 02/19/2007 >> 18:44:26.493 231 >> 6 5 241.67 38842.10 8.69 [ Ok ] >> 02/19/2007 >> 18:44:26.735 231 >> 7 6 96.71 45220.93 10.14 [ Ok ] >> 02/19/2007 >> 18:44:26.831 231 >> 8 7 198.41 51194.72 10.14 [ Ok ] >> 02/19/2007 18:44:27.030 231 >> 9 8 205.78 56794.69 8.69 [ Ok ] >> 02/19/2007 >> 18:44:27.235 231 >> 10 9 193.62 62045.42 8.69 [ Ok ] >> 02/19/2007 >> 18:44:27.429 231 >> 11 10 197.77 66967.72 8.69 [ Ok ] >> 02/19/2007 >> 18:44:27.627 231 >> 12 11 208.31 71581.72 7.24 [ Ok ] >> 02/19/2007 >> 18:44:27.835 231 >> 13 12 190.92 75908.43 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.026 231 >> 14 13 199.23 79964.20 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.225 231 >> 15 14 183.31 83767.48 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.409 231 >> 16 15 200.50 87331.98 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.609 231 >> 17 16 216.07 90672.73 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.825 231 >> 18 17 137.81 93809.57 8.69 [ Ok ] >> 02/19/2007 >> 18:44:28.963 231 >> 19 18 139.08 96750.28 8.69 [ Ok ] >> 02/19/2007 >> 18:44:29.102 231 >> 20 19 119.60 99508.41 10.14 [ Ok ] >> 02/19/2007 >> 18:44:29.222 231 >> 21 20 119.59 102094.16 10.14 [ Ok ] >> 02/19/2007 >> 18:44:29.341 231 >> 22 21 119.18 104518.33 10.14 [ Ok ] >> 02/19/2007 18:44:29.460 231 >> 23 22 126.80 106790.51 11.58 [ Ok ] >> 02/19/2007 >> 18:44:29.587 231 >> 24 23 112.05 108921.60 11.58 [ Ok ] >> 02/19/2007 >> 18:44:29.699 231 >> 25 24 145.14 110917.43 11.58 [ Ok ] >> 02/19/2007 >> 18:44:29.844 231 >> 26 25 115.57 112790.36 13.03 [ Ok ] >> 02/19/2007 18:44:29.960 231 >> 27 26 118.57 114546.05 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.079 231 >> 28 27 139.44 116190.71 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.218 231 >> 29 28 120.40 117733.77 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.338 231 >> 30 29 118.59 119180.49 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.457 231 >> 31 30 120.00 120536.71 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.577 231 >> 32 31 120.23 121808.15 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.697 231 >> 33 32 119.60 123000.17 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.817 231 >> 34 33 118.72 124117.74 13.03 [ Ok ] >> 02/19/2007 >> 18:44:30.936 231 >> 35 34 119.88 125165.39 13.03 [ Ok ] >> 02/19/2007 >> 18:44:31.055 231 >> 36 35 146.42 126145.90 13.03 [ Ok ] >> 02/19/2007 >> 18:44:31.202 231 >> 37 36 57.34 118265.37 11.90 SET >> Incorrect timestamp 02/19/2007 18:44:31.259 90 >> >> Reverse Packet Sequence Delta (ms) Jitter (ms) >> IP BW (kbps) Marker Status Date Length >> 44 1 0.00 0.00 1.45 [ Ok ] 02/19/2007 >> 18:44:36.119 231 >> 46 2 126.78 99992.08 2.90 [ Ok ] >> 02/19/2007 >> 18:44:36.246 231 >> 47 3 293.16 193724.25 4.34 [ Ok ] >> 02/19/2007 >> 18:44:36.539 231 >> 48 4 159.97 281606.49 5.79 [ Ok ] >> 02/19/2007 >> 18:44:36.699 231 >> 49 5 119.98 363998.58 7.24 [ Ok ] >> 02/19/2007 >> 18:44:36.819 231 >> 50 6 160.07 441238.67 8.69 [ Ok ] >> 02/19/2007 >> 18:44:36.979 231 >> 51 7 261.00 513644.94 8.69 [ Ok ] >> 02/19/2007 18:44:37.240 231 >> 52 8 43.09 481544.82 7.56 SET >> Incorrect timestamp 02/19/2007 18:44:37.283 90 >> _______________________________________________ >> Wireshark-users mailing list >> Wireshark-users@xxxxxxxxxxxxx >> http://www.wireshark.org/mailman/listinfo/wireshark-users >> > _______________________________________________ > Wireshark-users mailing list > Wireshark-users@xxxxxxxxxxxxx > http://www.wireshark.org/mailman/listinfo/wireshark-users > > _______________________________________________ > Wireshark-users mailing list > Wireshark-users@xxxxxxxxxxxxx > http://www.wireshark.org/mailman/listinfo/wireshark-users _______________________________________________ Wireshark-users mailing list Wireshark-users@xxxxxxxxxxxxx http://www.wireshark.org/mailman/listinfo/wireshark-users
- References:
- Re: [Wireshark-users] Jitter wrong in wireshark?
- From: Lars Ruoff
- Re: [Wireshark-users] Jitter wrong in wireshark?
- Prev by Date: Re: [Wireshark-users] Diameter unknown AVPs
- Next by Date: [Wireshark-users] communication between wireshark and my process
- Previous by thread: Re: [Wireshark-users] Jitter wrong in wireshark?
- Next by thread: [Wireshark-users] I've "officially" dropped support for Windows NT 4.0 ...
- Index(es):