Feature Request: Two further stabilization options

I’ve found the stabilisation function in TVAI to be very effective, but its usefulness (to me at least!) would be greatly enhanced by the addition of two other output choices beyond Auto-Crop and Full-Frame (Auto-Crop is fast, but seems to zoom the whole video by the zoom value required to hide the largest damped movement. Instead of zooming, Full-Frame fills in the edge gaps of shifted frames with pixels that resemble the existing edges of the picture; it is over 10x slower than Auto-Crop on my PC!).

Additional option 1 - Adaptive Zoom:
In this approach, the amount of zoom required to hide the image edges would be calculated on a per-frame basis (during the initial pass?). These values would then be smoothed with a ‘rolling window’ that always exceeded the current frame’s zoom (some approaches are described at python - Algorithm for smoothing a curve strictly above the original - Stack Overflow). A ‘smoothness’ slider could be provided to vary how rapidly the resultant curve changed:- eg a max value would keep a constant zoom factor just higher than that needed to deal with the largest displacement, whereas a value of zero would mimic the per-frame zoom values required (which would be useless, of course). I have used this approach in a stand-alone Python prog with very pleasing results; large displacements are dealt with, but not at the expense of over-zooming into relatively stable parts of the video, and I can set the rate of change to make the zooming hardly noticeable.

Additional option 2 - Flood Fill:
The existing Full-Frame option is computationally expensive (eg 20h per 1h video) and, for me, counterproductive when I want to post-process an unzoomed, stabilised video that requires info about the gaps between the displaced frames and the edge of the video frame. How can I get this info when Topaz has done its best to disguise those gaps (Full-Frame) or thrown away a load of edge-detail (Auto-Crop)!
The computationally cheap (ie fast!) but effective answer would be for Topaz simply to fill the gaps on a per-frame basis with a colour of my choice (green being the obvious one). The video could then be post-processed in something like DaVinci Resolve, using the ‘green screen’ to address the particular needs of the video project.