Ask Your Question

Revision history [back]

click to hide/show revision 1
initial version

How to remove SID frame from RTP Jitter calculation ?

Hi everyone,

I'm using wireshark to analyse a VOLTE call. SIP frames are decoded, RTP is properly decoded as AMR (using AMR dynamic payload type = 104 and AMR coding = BW-efficient). Wireshark is able to find all the RTP streams and show the different statistics for each packet (jitter, delay, etc).

My issue is that AMR packets are sent with two possible intervals within the same Volte call, as displayed in the "FT bits":

  • every 40ms (*) (voice : AMR-WB 12.65 kbit/s)
  • or every 160ms (SID : AMR-WB SID (Comfort Noise Frame) image description

(*) : this interval can also be only 20ms depending on phone and radio network settings.

Unfortunately that's polluting the way the jitter is calculated. It would be great if the jitter could be computed twice : one jitter for voice frames , and one jitter for SID frames, depending on the FT bits. Each time there's a SID after a voice, or a voice after SID, the jitter should not be reset.

I am going to do it manually, but that's quite painful... Anyone knows of a better way , please ?

Thanks a lot, PX

How to remove SID frame from RTP Jitter calculation ?

Hi everyone,

I'm using wireshark to analyse a VOLTE call. SIP frames are decoded, RTP is properly decoded as AMR (using AMR dynamic payload type = 104 and AMR coding = BW-efficient). Wireshark is able to find all the RTP streams and show the different statistics for each packet (jitter, delay, etc).

My issue is that AMR packets are sent with two possible intervals within the same Volte call, as displayed in the "FT bits":

  • every 40ms (*) (voice : AMR-WB 12.65 kbit/s)
  • or every 160ms (SID : AMR-WB SID (Comfort Noise Frame) image descriptionFrame)

(*) : this interval can also be only 20ms depending on phone and radio network settings.

Unfortunately that's polluting the way the jitter is calculated. It would be great if the jitter could be computed twice : one jitter for voice frames , and one jitter for SID frames, depending on the FT bits. Each time there's a SID after a voice, or a voice after SID, the jitter should not be reset.bits.

I am going to do it manually, but that's quite painful... Anyone knows of a better way , please ?

Thanks a lot, PX