These are super-helpful, James. The time and effort you have put in here, as elsewhere, are greatly appreciated! 
Returning to Frazier's reflection, I'm still troubled by the lack of distance between it and his head. On the simulation you posted a few pages back, the greater similarity of head-to-reflection relationship seems to be not with Frazier at edge of landing but with Frazier in deeper shadow from the lintel:
And yet, you seem to have gotten the lintel shadow spot on.
Why the lack of distance between Frazier and his reflection in Darnell? It just seems off to me...
Something to do with blurring perhaps? 
Alan,
I've misplaced my master for that GIF and redid one, this time using blue registration lines to distinguish from the earlier one. There could be a slight difference in the two as I had to remake it from scratch.
As to reasons for your noted difference I believe the major source could be due to how I generate the figure. My mannequins are basically cookie-cutter, with provisions for shoulder and elbow articulations for individual cases. What is available for individual cases (e.g. Frazier) is ability to scale the basic form in a way that makes the desired height. I can also change the width and breadth of the form. This I did for the Frazier stand-in by changing his width and breadth to 80% of normal ? to give the tall, lean look. Unfortunately,.that scale change also resized the Head. So we have a tall, lean model with a miniature head. For the next gif have changed his body form (and head) to normal proportions. This gives his head to reflection size more like we see in the Darnell frame. The animation displays the Darnell reference frame, then with Frazier at the landing, back 6? and west about 2? and back 12? and west 4-5?. To me, the landing and 6? back give results hard to distinguish, with 12? back seems to much for me. Of course, the blurring that is present in the source frames never help.
I will try the imgur.com link to see if it works here.
