If I were you I'd make it perfect adding a two modes frame rate
converter
which blends frames or uses morphing in between, optionally capable to
detect any duplicates first (for 8mm, 16mm, or even those 1930s
cameras with
manual film transport).
That would make three modes (blend, telecine,morph).
Option to detect duplicates is a good suggestion.
A motion compensation based de-interlacer and state
of the art intepolation algorithm to choose.
I only detect stationary parts. After all zero motion is very common.
IMO you would need very accurate motion vectors to improve
interpolation or noise reduction. Much more accurate than for
compression or frame rate conversion.
Do you have any references for state of the art interpolation?
