Lots of great pictures and lots to learn.
Again, this is not macro Raphael.
For those who can ‘do’ cross-eyed, I thought I’d try another slant along the stereo branch. [There could be a pun in there somewhere.]
Over the years, having gone through iterations of moving the camera and/or small objects around, it became clear that rotating the object on its vertical axis was a better option. (I feel sure this would be obvious to those into microscopy etc?)
However, it occurred to me that the CAD viewing package I have, might allow me to manoeuvre my skeleton clock file. The original file is quite ancient, and the CAD viewer ‘knows’ this. It always opens my files with a typical reminder of the fact.
After a bit of fiddling, I ‘rediscovered’ the X, Y, Z manipulation device that allows rotation of the image about any (or all) axes.
Chancing getting it right first time resulted in this.

When viewed, it became clear that the respective image differences are not so much position, but artefacts from the outlines and the ‘false’ reflections from simulated metal.
Your persistence payed off Nick. The above should jump out.
Good viewing guys.
Sam
PS – Take a look at the posts and the interest. Brilliant subject Raphael.