• Diana Castellvi

Week 4: plate clean-up and rough composite

My goals for this week were to clean-up the plates for shot #2 and #3, and to have a rough composite of all the CG elements in both. The clean-up for shot #2 was finished earlier in the week and a blog post dedicated to it can be found before this one.

Shot #3 clean-up

For shot #3, I had to clean the car in background and its reflections on the theatre doors. To accomplish this, I used footage without the car that we had filmed along with the plate (it isn't being used in the project because the camera movement is not the one we wanted).

I tried several methods to remove the car, like rotos and tracker/matchmove nodes. The method that ended up working the most was projection in nuke. I projected the footage from a static camera and viewed it through the camera obtained from the tracking of the plate so that it would move accordingly. I did this three times: one for the car, another one for the reflections on the door and a third for the reflections on the glass window on the ticket cabin.

For all of them, I chose a frame from the rejected footage that matched the plate in perspective and then I would hold that frame. I would also duplicate my tracked camera on that same frame and delete the animation so that it was now static. I then projected the clean footage on a card from the static camera and rendered it through the animated one. The result was that now the frame form the clean footage moved with the camera movement, sticking to where it is supposed to be. From that render, I would also roto the part that was covering the problematic areas and would merge it over the plate, effectively hiding what needed to be removed.

I also tried to clean the debris on the ground but, unfortunately, I'm having some issues with getting the roto paint I'm using to clean the plate to stick to the plate. I will do more tests during next week to successfully get rid of the debris.

Original plate

Plate after clean-up

Shot #2 compositing

This weekend I started compositing all the elements into the plate. Unfortunately, we had some issues with the files and rendering them and that affected the amount of time I had to spend on compositing before the deadline. I did establish a very basic nodegraph with all the layers we have that will undoubtedly grow as I spend more time tuning the composite. The amount of layers will surely grow too as we are in the process of importing the FX as an .ass file into maya. That will allow for more control over the elements and how they affect the car.

Shot #3 compositing

For this shot, we ran into some problems rendering the portal so it does not appear in the final composite. However, Mitch was able to render a couple of frames to get a sense of how it would look all composited and I have included them below.

Final thoughts

This was a very stressful week as we attempted to bring our 40% completed project to a 75%. However, this was a good opportunity to realize in which areas we are having problem and how we are gonna bring all the elements together. Personally, I'm very happy with the clean-up of the plate for shot #3, except for the ground part, which I will continue to address next week. Rushing to do the compositing for both shots, also caused me to make some mistakes which I will definitely change for next week's check-in.

3 views0 comments

Recent Posts

See All

This week was the last, which meant addressing feedback and making final adjustments. However, I feel like my team made a lot of progress in the last week. We had a new version of the sparks and the p

For this week, I addressed the feedback I got last week from our mentors at the Mill. This included fixing the glow of the headlights and the brightness of the lightning in shot 0, adding defocus in s

For this week's update I was able to spend more time on the comp for shots #0, #2 and #3. We are getting close to the end of the project and we are starting to do more refinement than actual big chang