Brooklyn, New York
I’ve started translating DOG from FCP to AE. I’ve found a free script to do this, which saved me from having to spend $500 bucks on Automatic Duck. Much as I would like to own AD, I’ll have plenty of other chances to spend money in the weeks ahead.
I’ve got act 1 in AE now.
In a previous episode, I discussed relighting the red corridor. I’ve done several tests so I know that there is a look I’ll like, that this effect is going to work. Following Stu’s order of operations, (see below) I’ve stepped back to fixing image flaws and adding visual elements like box labels, posters etc. The red corridor requires significant motion graphics intervention, so it’s a good place to start.
I’ve applied my limited knowledge of After Effects to rebuild the reaction shot of c as she opens the door for j. This shot is flawed in that the cabin window behind her did not have CTO applied to match the tungsten lights used in the shot, so there’s some blue light bouncing around. Also, though a polarizer was taped to the glass, the waves outside are slightly overexposed. Multiple NDs would have done the trick.
Here’s the original shot before repair. The waves outside the window are hard to see, they are about 1 to 2 stops over exposed.
Also, the reflection from the window on the cabinet is very blue. That will be adjusted when I color correct.
My approach to this situation may not be ideal, but it seems to work for now.
I’ve broken the image into 5 layers. The unaltered image is the background. On top of that are three layers pre-composed – the new waves, the window opening with surrounding wood trim and on top of those two is the shoulder of christina that passes in front of the window. Finally on top of everything is the door frame, (which eventually will have some red light falling on it when I get to re-lighting).
Here’s the shot after. Notice the brass door latch (lower left) which has a bit of christina’s dress showing through.
As she moves to the right, we see the waves through it. That’s part of the top most door frame layer.
Here’s the pre-composition showing the three mid ground layers. Black areas are transparent when placed over other layers.
The black lines are masks for rotoscoping the shoulder as it moves across the window. As the shape of the dress and position of
the shoulder change over time, these masks must be animated. That’s the time consuming part. The blue latch mask shown
here is not needed, it’s been applied to the doorframe layer discussed above.
The waves I’ve composited are not the most exciting, and maybe that’s ok. There is actually a lot going on in this scene already what with the dead uncle showing up bathed in hellish red light and all. Because I’ve done all the heavy lifting with the animated masks, I could just as easily slip in one of the subtle apocalypse shots such as the cliche flooded empire state building or wrecked cruise ship. We’ll see.
I have 5 of these reaction shots to roto and reconstruct. C’s hair is not a trivial issue by the way, that also has to handled if it passes in front of the window. Most likely I’ll just trim out the loose strands from the whole scene so no hair ever passes in front of the window. Otherwise I’ll have to recreate them from scratch because the original hair that passed in front of the window is totally blown out from the blooming light.
Stu Maschwitz suggests onlining in After Effects and provides the following order of operations…
1. remove DV artifacts (DOG was shot in HD, not DV)
2. Add any visual effects elements (you are here)
3. optimize the shot (technical wizardry involving color and contrast)
4. relight and change object colors (red hall, enhance soft focus with light tricks, etc.)
5. gradients, diffusion and other on-lens effects (may not applicable to DOG)
6. vignetting and other in lens effects (maybe not)
7. perform color correction (a big deal)
8. simulate film stock (maybe)
9. dissolves, fades and transitions, with titles (lots of these)
10. perform specific color corrections, resizing and sharpening for specific output formats (DVD for now, but could try a film out too after all this work).
Currently, we are at step 2 on this sequence. Some of the steps we can skip, that leaves us with about 5 major steps. I’ve also got to master the sound, so that’s 6 (major) steps to finishing.
So why write about this process, why not just dedicate all my energy to doing it?
Doing these posts is an attempt to slow down time. When I can consistently articulate what’s happening, I feel more efficient. Not only am I taking stock of my own progress, but I am clarifying what’s most important, what has to happen next. After writing my report, the time I spend working is focused and productive. I am then working super fast, which is in effect slowing down time.
Got a call from my brother and he likes the blog style, thinks it would be of interest to other filmmakers. I think so too and I wouldn’t mind including a “making of” micro-documentary when DOG is finally released. These posts are a start on that. At this stage however I am not quite ready to expose my completion process to the general public. Finishing this project is not exactly a cake walk, so I’ve only invited my top supporters and allies to tune in. Knowing that my peeps are with me week to week keeps me going.
With luck, many of us will collaborate on future projects. Sharing these concepts builds our collective knowledge of what’s possible.
Stay tuned…
Dan Kelly