Just to make sure I wasn't fooling myself with an isolated overlay example, I chopped out all segments with 2x magnification in IR and TV and overlaid them with rigidly fixed overlaid positions. There are three 2x IR segments that are much shorter than the one 2x TV segment. So I rerun the three IR segments over the one continuous-running TV segment. I didn't use the last few frames in 2x where the object leaves the screen. The purpose of this is to see how the ATFLIR system localizes the TV and IR signals during smooth tracking, thus off-center motions (ie, loss-of-tracking moments) are irrelevant. It's apparent that when the target is smoothly tracked, the ATFLIR system localizes the IR signal to a position corresponding to the upper right quadrant of the TV-imaged 'tic tac'. That happens to be the exact location I suggested the IR signal would fit if it was a right-side-biased solar-warmed upper edge of the object. I suspect this is not a coincidence. Very likely the ATFLIR system builds a composite model of tracked targets, assigning locations to the target's hot spots where those hot spots would exist on the visible (ie, TV imaged) object, and maintaining during smooth tracking a static position for the entire composite object.