I kinda assumed it was using VPN tunneling through Starlink whenever it showed the UDP connection. The UDP connection itself shouldn't indicate that, it's just a matter of timing when it happened. I assumed Starlink would only work during the bellyflop, and the udp connection messages only happened during the belly flop. The only evidence is the slightly different behavior that happened after the flip along with the video working a bit better after than before. It also failed in a different mode: normally the failure mode was a freeze, but post flip there were MPEG encoding errors. Was also wondering if they were communicating to the ground in the microwave spectrum since they seem to have a rattier signal when it is overcast.
SN11 and SN15 were both in to the clouds, without visual tracking. Maybe the video feed reception requires a directional antenna that is pointed towards the transmitter manually?