I decided to do another camera projection exercise as part of the challenge posted on The Challenge Jar website.  I visited Houghton Lake earlier this week where I had taken a lot of snowy landscape pictures.  The image I chose for this project was taken at Hartwick Pines State Park.  This time my goal was to create a camera projection from the snowy woods image below.

Snowy Woods at Hartwick Pines

As with any camera projection project, the first step is to breakdown the image into components that can be separated in 3D space.  I chose to use the two trees in front as the foreground.  I found two more trees that I designated as the midground, and then relegated the rest of the woods as the background.  I also separated the ground up to where the bulk of the background trees begin.

The challenge called for monsters.  While walking in the deep woods at Hartwick Pines, my first thought was ‘What a great place for a Dogman encounter.’  Northern Michigan has its own werewolf folklore.  If you ever meet someone from Northern Michigan, ask them about the Legend of the Dogman.  So I threw a Dogman picture into the list of images to be projected.

Image Breakdown in Gimp

Once components from the image have been chosen, the next step is to extract each piece and paint in the missing imagery.  I extracted the images with Gimp since I didn’t have Photoshop installed on my laptop.  I later finished the matte painting chores with Photoshop.  With images in hand, the next step is to create the 3D environment in Blender.

Mapping Images to Geometry in Blender

As with my last camera project exercise, I created simple geometry in Blender, and added a slow camera dolly into the scene.  Camera projections only work with simple camera moves, so I was limited in how I could maneuver the camera.  However I decided to try something different.  Since the challenge called for a ’stop motion style’ animation, I added jitter to the camera movement.  This was accomplished with a python script I found for Blender camera moves.  The intent here was to create a jerky, hand-held style camera move.

Final Composite in After Effects

My goal was to use the edit nodes in Blender for all compositing.  However, after a number of failed attempts, I realized I need to learn more about nodal compositing to be effective.  So I rendered out the image layers in a series of passes, and composited the shot in After Effects.

Dogman Encounter from Fourth Order Light on Vimeo.