Corsair, some nice observations,like the plane disappearing. I think the camera was tripod mounted, not handheld.
I think this fake video might make a nice study, to see if stabilizing it smooths out the velocity line, or makes it more jagged. I'm not sure, but I think that might be real video of a real jet at an airshow, with background removed and laid on top of the video of the buildings. I'm not sure.
You mentioned lighting not quite being right. Hmmmmm. . . .
What color was UA175? And what color should it appear in full sunlight?
I have a few moment to make a point or two. IMO the action was deliberately filmed by an amateur (or pro in his spare time) for the purposes of adding the fake plane. The shot was most probably created either as a student project or just to show off on an online CG forum.
The shot is handheld - there is some roll of the camera about the lens axis which is unlikely on a tripod. But, there is practically no parallax, which means the operator is holding the camera in a fixed spot. Having no parallax makes the shot significantly easier for the artist, but such a camera move rarely happens by accident. That leads me to beleive the artist shot the footage himself.
The wrongness of the smoke, lighting on the plane, and its movement etc. are all quite subjective, especially given the compression. I wouldn't say any was proof of fakery but it all adds up. I couldn't select the frame where the plane disappears but I doubt that this happens. There's no practical reason why a half competent FX artist would make such a mistake and so it's probably just an illusion caused by the compression and the plane colour matching the building.
For me the biggest giveaway is the lack of reaction. That's one thing out of the control of FX artists. Also of note, one of the explosions at the front of the building is quite elongated. This could happen for real, but looks to me like either deliberate distortion by the artist, or it's an explosion shot in controlled conditions elsewhere and added in (you can buy these online for a few dollars). Also the smoke and debris behind the building seems to have a slightly mismatched black level (in english - it's too dark).
It's a good effort though.
Now Ace, you're guessing that the plane was taken from real footage. You making the same guess for the WTC shot too. This is really not a wise move for the artists.
First up. Planes are easy to make as computer generated elements. Nowadays you can buy useable models for peanuts (
some examples)
Second. Real footage of planes will probably have a moving camera. This movement would have to be removed via stabilisation. Stabilisation is not perfect and can introduce blurring, severe blurring in some cases. On interlaced footage you must halve the resolution of the clip before even starting so the problems are compounded. Using a CG plane the artist can use any camera movement or none at all, to create the element he/she needs.
Third. Real footage of a plane comes with lighting already on the plane. It cannot be changed once the photograph has been taken. Correct lighting is not just a matter of correctly aligned shadows. An artist needs to match shadow density, sharpness, bounce light, environmental light, colour temperature as well as a whole stack of camera specific attributes. These things rarely match in two totally seperate shots. For CG planes however, the artist can light the subject however he or she desires.
Fourth. Real footage has a fixed motion and a fixed start and end. CG planes can be animated to do anything for as long as the artist needs it.
Fifth. A real plane would need to be extracted cleanly from its background. Although chromakeying has been around for decades this is sill not a trivial process if perfect results are needed. There are some excellent intelligent solutions for keying but they all require good quality blue/green/magenta screens. It IS possible to pluck something out of a blue sky but the result can be messy, especially if that object has a tendency to aquire the same colour as the sky (shiny or white objects). CG planes can be rendered without a background at all, so this process is unecessary.
Sixth. The footage of the real plane would have to roughly match the expected action in the completed scene, with a roughly similar camera move. (Stabilisation would fix the small differences). This might not be a problem, but if you don't know what the background footage will be in advance, or how many cameras will capture the scene, you have a big problem knowing exactly what to film beforehand. CG has the same problem, but the action or camera move can be edited in minutes in order to match the incoming footage.
If you think 2001 was too early for fully CG planes, look at this homemade movie from 2000/2001:
http://www.ifilm.com/video/204155 (The page says 2001, but I'm pretty sure I first saw it in 2000).
In 1996, a largish Hollywood movie featured a shot of a military plane that had been cut out of stock footage and digially inserted into the production's background. I'll let you figure out which film it was. Despite ideal conditions, the effect is not particularly successful, even when viewed on TV, IMO mainly because of the third point above.
So, to conclude all that. Any sensible evil CGI artist would have chosen a computer generated plane, not a real one. Therefore what you claim are rotoscoping artifacts around the plane (inaccurate cutting & pasting) would never exist. Let's ignore for now the fact that your pink halo looks nothing like any roto error that I've ever seen yet looks exactly like JPEG artifacts.
Also your explanation for the jerky movement you have found makes little sense. The fake plane would likely have been rendered performing a perfectly accurate, smooth flight (since that's what computer animation does by default). The smoothly moving plane would have been tracked into the shot by computer.
Now, contrary to some debunking, tracking can be done live, and this technology did exist in 2001. It is however imperfect.
This is where I have to give you a little credit, Ace. Stabilising a shot and looking for anomalies in the movement does work. I've used it myself to show that a UFO video was fake. As great as an artist's tracking software may be, it can still slip up and can be fooled into making mistakes. Therefore the jitter you claim to have found, might really exist in fake footage.
However, as has been stated many times, you can 'find' this jitter by simply not measuring accurately enough. Humans in my experience are not accurate enough, so no conclusion can be reached.
Working against your theory, amongst many other things, is the very tight integration bewteen plane debris, fireballs and the building, which are very very difficult to fake in CG, even with ample time.
I don't expect this to be the end of the matter, (since,
'it don't look right to me' can't be debunked) but I can't have people trampling all over my turf without saying
something.