Monday, April 30, 2012

Self-Evaluation

At the onset of this project I had very high expectations. I always go into a video or short film knowing what I want the final product to look like, but often not realizing how unfeasible my desired result actually is. However, after finishing this project, I am happy to say the finished product is pretty close to my expectations. Some issues I have with the video include: frame rate issues (converting 60fps to 30fps turned out to be more difficult than anticipated), lighting inconstancies and issues (as even with 3 lighting kits shots came out very dark), actors not being able to hold certain positions still enough, and not having access to a track dolly system. Some accomplishments of mine however come in making up for the lack of certain pieces of equipment. The camera moves I was able to achieve with a wheeled dolly and in house stabilizer turned out to be quite effective. Continuity between shots also held up well.

All in all, I am very happy with the final product and look forward to potentially revisiting this concept in the near future.


Thursday, April 26, 2012

Production Log

some of the lighting kits used for the project
 The entire short was shot over the course of one afternoon, between the hours of 10:00 and 4:00. The first scene started in my room, which is a quad on the floor. I brought in two lighting kits for the project, as lighting in the towers can be a nightmare. I shot on my own camera, a Panasonic AG-HMC40 Prosumer camera. The camera is capable of various frame rates, including 60fps at 720p. Though it can shoot 60 fps at 1080, it is interlaced footage, which would have defeated the purpose of slowing it down to 30.
motivating the actors to convey proper emotions 



The hardest part of this particular shot was making sure each actor could hold his pose extremely still. In order to sell the effect of stopped motion, the process used was fairly straightforward. I had each actor hold their pose as emotionally active as possible, but make sure not to move. Then, with very exact camera movements and the help of Andrew Kuserk and some 3D tracking programs 3D objects were added. I tried to make sure my narrative focused on objects that could easily convey frozen time.

one of the many takes we shot
As the shots continued forward, I had to be very exact about the path taken. In order to try to sell the effect of a single tracking shot, each individual take had to start and stop at corresponding points. This process proved to be quite difficult for several reasons:

  • maintaining the same lighting and color from shot to shot
  • making sure the angle of the camera stayed consistent
  • maintaining the same placement of objects from shot to shot 



Erik's perfect reaction to the situation unfolding
As the narrative moved into the hallway, we see two friends playing hockey, and it seems something went wrong. Though the physics of the shot are not realistic (as the glass would shatter out not in from the impact of a hockey puck) this shot proved to be the most difficult in post production. 

Alyssa's addition to the scene sold the shot, however proved to
 be difficult in post







Not only were the refractions of the added glass difficult for Andrew, but also the ability for the glass particles to move in front of the actress Alyssa.


the twitch effect used to stitch individual shots together
In order to have the glass particles move in front of Alyssa (rather than pass right over her since they are just a PNG image sequence with an alpha channel) she needed to be rotoscoped out of every frame. Rotoscoping is the process of painting pixels out of a shot frame by frame. The process is often used to remove wires from actors in post production or imperfections in a shot. In this case, Alyssa had to be "removed" from each frame the glass particles had to move around her. After abut 3 hours at 60fps, a mask was used instead as the rotobrush in after effects "wobbled" too much around the actress. 




Moving down the hallway the shots begin to use an affect effects plugin called twitch, to stitch each shot together. This allowed imperfections like color and lighting differences between takes to be hidden. 

the two individual takes, one darker in color
the two individual takes, one lighter in color














one of the best shots in the short, the frozen effect was sold quite well here


The next shot features actor Ben S. and actress Sharmin frozen as the camera moves through the papers frozen in midair. This shot was on of the best in the short as the actors held incredibly still and the floating paper effect held up well. 





The last shot of the short involves actor Ben L. breaking the frozen time effect and yelling into the camera shut up. This ending seemed logical and a quick way to break the action ensuing. 






Time Freeze Final-Abstract/Storyboard/Concept

beer spilling and gunshot frozen in time
 When it came time for me to select a theme for my IMM 412 final, I decided on an idea relatively quickly. I wanted to freeze time. There was this one shot in the film "The Other Guys," starring Mark Walhberg and Will Ferrel that confused the hell out of me. The shot appears to be a single take tracking shot following the Walhberg and Ferrel through a crazy night they have in a bar. However, what makes this shot extraordinary is the fact that time is frozen. The camera moves freely through 3D space as frozen pieces of popcorn and spilling beer remain frozen in time.

frozen popcorn pieces


I instantly wanted to delve into how this shot was created! I searched the web and found a few similar sequences that resembled this bullet time frozen effect. The first instance I stumbled on was a commercial Philips had produced about 3 years ago. The video followed a heist situation gone wrong, one tracking shot of an entire hospital complex...oh yeah, and everything was frozen in time. In order to achieve the amazing sequence, 3 different motion control rigs were used to keep shots consistent and exact. Then, two shots were filmed for each segment of the video,
one as a clean pallet at 24 fps, and then a second at 48 fps with actors in the shot so their slight movements would become less noticeable. Another technique used was to shoot the scene at an extremely high frame rate (about 750fps) so that the actors slight movements were not picked up in the final composition.

Realizing I had none of this equipment or capability, I looked to alternative ways to sell my shot. I knew I wanted a single tracking shot of a college freshman floor. The shot would have total chaos ensuing on the floor, and then end with one student trying to study and breaking the frozen shot when yelling at his floormates to shut up. However, achieving this shot proved to be difficult even before sitting down with my friend, Andrew Kuserk, for the 3D post production.

I started testing about two months before even starting production in order to see what type of camera I should use. After several failed attempts using a high speed point and shoot camera I own, I decided it would be best if I used a high definition camera at 60fps. Though the high speed camera could shoot at 220 fps it was just too low a resolution. After some testing with actors holding positions at 60fps and slowing that footage to 30 fps in after effects, I found a pretty good balance between semi high speed and quality. I also utilized the warp stabilizer in after effects as well as other tools to help clean up my footage. Next came production.