RE: [AVT] questions about jitter

["Alan Clark" <alan.d.clark at telchemy.com>]

Hi

It depends on the delay variation introduced by your model.  In many cases
high levels of packet delay variation introduced on access links or low
bandwidth LANs tend to be non-Gaussian in nature .. using a smoothed PPDV
model may not be representative.  The VoIP Troubleshooter site discusses
this in some detail and also has a downloadable IP network simulator that
uses a time series model to generate delay traces.

Regards

Alan Clark
Telchemy

PS: In your equation - presumably d is an absolute value.

-----Original Message-----
From: avt-bounces at ietf.org [mailto:avt-bounces at ietf.org]On Behalf Of
Fangweiwei
Sent: Friday, October 22, 2004 9:19 PM
To: avt
Subject: [AVT] questions about jitter


avt，hi！

	As I know, Jitter is standard deviation of inter-packet time ,but As in
the RFC3350:

      fJitter += (1.0/16.0) * ((double)d - fJitter);

    I don't know why RFC 3550 computes it as that.

    Now I do some work as Network Emulation. And when I changed the virtual
network, from good to bad (eg.packet loss), I calculate the packet jitter
using standard deviation fomula.and It reflects that the network quality. If
packet loss too much, the packet jitter will become big...

    but unfortunately, I find the fJitter changes slightly,and it's value
has little relation with network quality.I don't know why?

    Thank you very much for all your answer!


               Yours Sincerely


　　　　　　　　Fangweiwei
　　　　　　　　fangweiwei at tom.com
　　　　　　　　　　2004-10-23

  ====================================================
  Weiwei Fang
  Network Emulation Group
  Lab for Computer Architecture and Network Technology
  Bei Hang University
  P.B.: 37 Xueyuan Road  Beijing 100083  P.R.China
  ====================================================


_______________________________________________
Audio/Video Transport Working Group
avt at ietf.org
https://www1.ietf.org/mailman/listinfo/avt