Wednesday, 11 April 2012

Visual Effects Project

An aim of this project was to gain understanding of the basic principles of visual effects in production:

  • Lighting and shooting video
  • Importing video shots into the digital environment
  • Camera tracking / match moving
  • Compositing
  • CGI Lighting

We were required to produce a visual effects shot: edit and post-produce a short sequence of video that includes live action, camera tracking, compositing of a 3D object into the scene and final output as a digital movie file. The finished piece had to be 15 to 30 seconds long and contain a minimum of 2 shots with 15 seconds of visual effects. During this short production I used: Matchmover, Maya, Nuke, Final Cut Pro, Photoshop and the Sony the HVR-Z7E  -  handheld camcorder with native progressive recording, and digital back; a fig rig/video tripod; a chrome ball and Nikon D70.

On this project I worked with Laura McNaughton. 'Our idea was to create something unusual in an everyday setting e.g. jellyfish pet. During my research I discovered and fell in love with  this video:




During development stage we worked on pet jellyfish, owner and storyboards. 'We did a short reconnaissance around Dundee and decided to work around the McManus Gallery, where we found lots of  horizontal and vertical patterns. We planned 2 basic shots – 360° pan round figure and other shots eg, a straight pan, or locked off shot



During the pitch we received a feedback from Chris Rowland. He suggested we tackle the 360° shoot in a few attempts, in several cuts, not all in one go and also to try to use rotoscoping to cut out the balloon. Otherwise, no problems were spotted.
 
We ended up with a lot of shoots but we went with these ones for tracking:
This is a test shot for the pet jellyfish tapping Sarah on the shoulder


 Laura modelled some jellyfish:


Problem 1:The Balloon

Just before starting on working with the footage in MatchMover, Malcolm spotted instantly that the use of the balloon was a bad idea. This was because: it was moving too fast and it was too flat, not enough information – despite drawing markers on it.
  
Problem 2: Exporting image Sequence
When attempting to export the image in Final Cut Pro another problem arose; the aspect ratio was wrong. This meant the entire footage had to be re-exported with the correct settings. This took some time.

Problem 3: Maya and Targa files  
Targa files were not visible in Maya 2012. They had to be taken into after effects and exported as jpegs. Then re-imported once more into Maya. This took more time away.

Problem 4: Problem with footage
As the idea of the pet jellyfish was no longer feasible, we decided in the interest of time saving it was best to place something else into the footage we had, instead of re-shooting everything and starting from scratchWe decided to place different objects into our footage, a sign, a traffic cone and a UFO. This meant that what we needed from the footage had changed. Now, the footage was difficult to use because we hadn’t paid attention to things like reflections in the street from the rain, people walking past (because the focus was on the balloon). There was also not enough usable tracking points for the X & Z axis because of the reflection on the street. The moving objects in the foreground made track points unusable.


Problem 5: Track Points 
I tried to work with manually placing tracks into the footage, but Maya didn’t allow the co-ordinate system because “Tracks were not computed” I had more general problems with the  co-ordinate system, like trying to get it to work in our scenes by redoing it several times, re entering track points in different combinations, and trying different options. For reasons I could not work out, Manual tracks became visible behind the camera. So I decided to scratch the idea, and start again.

 Problem 6: Idea Scratched! Time to refilm
 
New Idea, New footage: I filmed several new shots in the car park outside DJCAD, a few good shots were obtained 

New Footage, New Problems 

  • Fighting to get the co-ordinate system to work
  • Problems with tracks “not being computed” again (not using manual co-ordinates)
  • It froze at 66% when solving for camera, and stayed like that for 45 minutes.
So here  you can see our final, not at all perfect test ;) 




No comments:

Post a Comment