The speeder shot is probably never going to look much better than Harmy has done with it, but if it proves the most popular, I’ll do it.
If the purpose of the demonstration is to show the strengths of the software you have developed, I would suggest also picking a sequence where you expect the maximum benefit of the software.
With regards to the speeder shot, I’m curious to know the maximum image quality that can be pulled from the prints, without reverting to the in my opinion somewhat revisionist techniques employed by Mike Verta to clean up the speeder shot, and essentially remove all the color noise that most likely also existed on the original negative. The downside of removing all the color noise in my view is, that you will never recover the detail that should be associated with such a clean image, and thus the cleaned up image may actually appear less natural without the noise than the image with the color noise in it.
Would stacking multiple scans of those 139 frames (channel aligned, free of warping) help reveal detail?
I’ve watched a few of the videos and am still not sure if verta used more than one print source on that shot to get ride of the noise to get closer to what the original negative looked like.
From what I understand most of the noise and dirt seen in that sequence is not print grain, but also on the negative. Any dirt and grain specific to the print can be strongly reduced by using multiple prints. Additional detail may be recovered by stacking multiple frames. However, you would recover detail, that wasn’t on the negative to begin with, and would thus be outside the scope of a restoration in my view. Personally, I don’t see much difference between stacking multiple frames to recover detail, and to do what Lucas did, which is to go back to the original elements, and recomposite the image digitally. In either case the purpose is to enhance the image beyond what was capable in 1977. So, if you’re going to be revisionist, why then not use the technique that provides the most detail?