Yes. The air density changes rapidly depending on whether or not that slice of air was in the shadow of a building lengthening the path and attenuating the pressure change and it's rate of change.

That feels plausible. I think the effect of shock on the camera sensor also plays a part.

This looks like an early frame, a couple of seconds before the shock wave reaches the camera. That is why there are so many clear videos, as it takes a while for the shock wave to reach the observers - if you are 3 kilometers away, you have 10 seconds to watch the explosion, before it hits you.

Camera sensors should also not generate artifacts due to shock, unless they fail. But what is possible, that very fast moving subjects create artifacts in the image due to the limited shutter speed.

Seems like a combination of that and what seems like some algorithm performing temporal denoising on the video going haywire from the quick moving picture.