I concur. Note the racket that starts up as the transonic condensation clouds form, and then cuts out very suddenly.
Jepp, right. I think (educated guess) that this is more or less logical.
First the microphone picks up the noise via the air. This is what most microphones are designed for
After the air density gets thinner and thinner[1], the microphone starts to pick up most of audio wave signals via direct vibrations through the mounting.
I would compare it with e.g. scratching on your notebook while using the build-in microphone. It [the mic.] will still 'hear' the scratches although the sound is very different (kind of low pass filtered).
See the note in the lower left corner at 0:07.
Didn't notice that the first time, thanks for the hint:tiphat:
[1] ...and/or the noise-source gets trans/super-sonic...