I will shoot 3 sequences of high resolution progressive footage:
1. An airplane flying across the sky and disappearing behind the edge of a building.
2. The same building as in #1 with no airplane.
3. An airplane flying across the sky all by itself.
I'll remove the background from #3, leaving an airplane flying against a transparency, and color correct it as I see fit.
I will arrange time at a video editing room with an AVID symphony or similar. I'll arrive with the 3 video files. My goal will be to, in real-time, using real-time luma key, insert the airplane in #3 into video #2, and make it appear to fly across and disappear behind the edge of the building. We'll convert the composited output to NTSC, and also convert video #1 to NTSC.
I'll end up with two NTSC interlaced videos, both showing an airplane crossing the sky and disappearing behind a building edge. The test will be to see if Steve Wright can identify which video is the composite, and to point out any:
1. Mismatched interlace scan lines.
2. Color disagreement.
3. Harsh edges.
4. Differences in video noise between airplanes and building.
Subscribe to:
Post Comments (Atom)
2 comments:
very nice challenge, I hope he accepts it and that we may see the results.
Keep up the good work Ace
Sounds fair enough, for the expert.
Might be a good experiment to do varying levels of compression and some vhs dubs to see at what point
the ability to distinguish real from composite is lost.
Peace
Post a Comment