Revision history [back]

I think you've made an invalid assumption here, in that the rtp clock drift calculation can actually be used to measure clock drift between machines in general.

RTP clock drift is between the machine sending RTP traffic and the receiver, so unless your hosts are actually sending RTP traffic, which they probably aren't, you won't be able to capture any RTP traffic, no matter how much you try to force Wireshark to decode UDP packets as RTP.

There are protocols and tools out there that actually determine clock drift as their purpose, e.g. ntp. Attempting to infer clock drift by capturing network packets is probably an exercise in frustration.