On Tue, May 27, 2008 at 4:30 PM, Andy Spitzer <[EMAIL PROTECTED]> wrote: > Woof! > > On Tue, 27 May 2008 11:58:39 -0400, M. Ranganathan <[EMAIL PROTECTED]> wrote: > >> I started sipxbridge and continuously pumped packets over 50 bridges at >> the rate of 1 256 byte packet every 100 ms per bridge bidirectional. > > The worst case real-world audio packet rate of interest is 20 mS G.711 RTP > (172 bytes/packet).
After I posted the results, I looked around on google and realized that 100 ms was too much and hence I did the test again for 20 ms between packets with the parameters you suggest above. There was no perceptible degradation in voice quality. I will redo the CPU load test again and publish results. >All other audio will be way less than that. If you really want to punish >yourself, try 10 mS G.711 (92 bytes/packet)...but no one uses it. OK I will try that one as well. > > So 256 bytes per 100 ms is not a great real world test. G.711 RTP @ 20 mS is > what all hard phones produce by default. Seems to have become a defacto > standard when no one was looking. > > Video traffic is totally different and really depends on the frame rate, > picture size, and encoding. Let's cross that bridge later. Thanks for your suggestions. Ranga. > > --Woof! > -- M. Ranganathan _______________________________________________ sipx-dev mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-dev Unsubscribe: http://list.sipfoundry.org/mailman/listinfo/sipx-dev
