Correct. IR video tends to bloom. If the object is hot enough, (for instance, a jet engine exhaust, a burning vehicle, a refinery flare...) you'll have a blob

Correct. IR video tends to bloom. If the object is hot enough, (for instance, a jet engine exhaust, a burning vehicle, a refinery flare...) you'll have a blob of smudge around/behind it on the video. The video is contending that the rotation of what is seen on the video isn't the object itself rotating. Instead, the blob is elongated as a result of glare/lens flare/diffraction gradient, and this elongation is aligned with the rotational axis of the gimbal. As the gimbal nears gimbal lock, it is forced to rapidly rotate to continue tracking the object. This rapid rotation of the gimbal causes the glare to rapidly rotate. I do have a background in IR video, gimbals, and translating between gimbal telemetry and real life coordinates and the explanation in the video looks good to me.