It's different because it coincides with the jet banking angle going clockwise. Background rotates when the jet rolls (i.e. a change in bank angle), the glare rotates when the pod head rolls relative to the pod body.
There's a long smooth rotation here partly because it's at the steepest part of the curve, but also because of the jet rotation clockwise which has the effect of making the curve steeper (try toggling "Use real bank angle" to see the effect) and also rotating the background in the same direction.
Please let me know if I get that right :
- the fact that the glare does not rotate with the horizon (jet banking) is inconsistent with a physical object, but consistent with a glare.
- we see it when the plane banks at various points of the video. The glare does not rotate.
- at 0'30 the plane makes the most important change in banking. This would be the most clear evidence that the object is a glare, because it should not rotate with the background (that clearly does at this moment).
- but because the pod head also rolls at the same moment, the glare actually rotates, and it does it a little more than the background (meaning the head roll compensates for the non-rotation, plus adds some additional rotation to the glare).
- the pod head rolls in "steps" before and after the plane banks, but it rolls smoothly when the plane banks, "hiding" what would be a dead giveaway for the object being a glare, because it would not rotate with the background if the pod head was not rolling. Or it would but with a step (unlike the background), if the pod head was rolling with a step like before and after.
So here we have a pod that makes steps in its rotation, but just not when it would make it obvious the object is not part of the image, but is a glare. This is unfortunate.
And thanks to you and @jarlrmai
for the examples of IR videos.