"IR signature of a physical object" is perhaps a poor descriptor. If you point a camera at a jet then what you see is the IR signature. A glare is arguably the IR signature, within the limitations of the camera system. And idealized IR signature would basically be a crystal clear image of the jet and the exhaust in the MWIR spectrum.FLIR is not a glare right? It's an IR signature of a physical object.
The FLIR1 video starts out as glare.
Later in IR though there's very little glare, and we are probably seeing something more like the true shape
I've argued that there's a glare on the right side, and that you can even see it rotate during the gimbal roll, but it's a hard sell at this resolution.
My suspicion is that it's a plane more tail-on at the start, hence the glare, and then it turns more to the left. Like a much less extreme version of this: