Writing Samples

Visual Effects “How-To”

An example of an article written for the reader with no knowledge of VFX:

Featuring: POCARI Sweat

“Underwater Tennis”

Tennis?  Underwater??  Well, it can be done!  Here’s how:

We began this process with the director and his storyboards.  These were very specific and detailed each shot.  These boards were evaluated as to how they could be done, and how they SHOULD be done.  Our general policy is that “ If it CAN be shot, it SHOULD be shot”.  All elements that could be handled as live action were then deemed live.  All other visual elements such as trees, fog, location, bubbles, ball trails and particles in the water would become CG or composited elements.  We then went through the boards and decided which frames, or elements in the frame, needed to be shot “Wet-for wet”, or underwater, and which could be shot “Dry-for-wet” on a stage.  That left the remaining shots that must be shot underwater.  Once this process of distinction between practical filming and post-production effects has been completed, we can begin to execute the work.

There are many components to the spot that may look like real things but in fact are digital.  Some are a combination.  Some digital elements started out as live images.  This sounds confusing, because it is.  Let’s look at one typical shot:

Here we see a foreground player on a court with his opponent, a ball trail, and a background with fog.   The foreground player had to be shot in the water for realism.  He was shot on a greenscreen (they DO work underwater too!).  The bubbles on his racket are real plus digital.  We fed air into the racket frame, and added CG animated bubbles to make them denser.  The ball, trail and court are CG animated, and the net is a photo of a net “projected” onto geometry in the computer and “waved” gracefully for interaction.  The opponent is shot on a stage with air blowing strongly on him from below, and shot at a high frame rate for that slow underwater motion.  The background is primarily CG, with the trees from still photos shot at a park near Digital Domain, and the hills and grass digital.  The trees are simply “projected” onto geometry too.  The fog is digital, as are the rays.  Each shot has it’s own needs and breaks down differently.

Here we see a player underwater, and a ball that is a filmed element (shot hovering on an air stream spinning) and a CG bubble trail.  The court and net are just photography on stage, and the background added in compositing.

In this shot a foreground net is shot on greenscreen, and the player underwater also on green.  The background is handled the same as the others.  The rack focus is handled in compositing.

There is another piece of this puzzle to be explained:  CG Animation!

The tennis ball is often CG animation in this spot.  The process of creating a CG animated image starts with design.  A design is created I any medium preferred.  From paper and pencil, to 3D modeling and Photoshop paintwork: All are acceptable.  Once a design is approved a model is built.  This consists of pure geometry.  It is a mathematical description of a 3-dimensional object, or ‘.obj’.  The modeling process can take several weeks for complex objects such as a detailed building or a sleek and detailed car.  This file is then surfaced and textured.  Surfacing describes the qualities of the surface.  Glossy?  Sandy?  Furry?  The details of how the fuzzy tennis ball will react in the water must be considered.  Highlights are suppressed underwater, and there is a certain softness to the light.  These must be created too.  The color and pattern are texture maps, literally maps of the globe wrapped around the ball. This can be quick, or take a week or longer for things like hair or fur.  The motion of the ball is controlled by a different part of the program (Lightwave in this case!).  The motion of an object is potentially complex.  Every subtlety of movement must be created.  There are three dimensions, and three axes of travel and rotation.  If the ball is meant to travel in a certain space, that space may need to be created as well.  We made a CG court for the ball to bounce off of, and the net to react to the wake.  Though this example is simpler, complex animation can take well over a week or two per 2-second shot.

Tracking is a critical part of the process.  It refers to the process of reproducing the exact behavior of the camera in a live shot, including motion, rotations, focus, depth of field and lens distortion.  The photography on stage was done on a small court set with greenscreens all around.  We surveyed the dimensions of the court and net and used that data to reconstruct the court in CG.  This is because when animation is rendered in the computer, it must match the scene for which it is intended.  Animation is rendered to a CG “camera” and it must match the live camera precisely.  If it does not, sliding, weird scale issues and lack of realism can result.  A moving camera has presented challenges, but is now much easier to achieve with CG animation because tracking, and especially our award winning software, has become much more powerful.  Tracking takes a week or two but can run parallel to other phases of post-production effects work.  Animation cannot begin in earnest on a photographic plate without a track.

Tracking yields a “camera” that tells the computer how to see the 3D object.  The render will need a lighting set-up to illuminate the object.  The lighting of the render can be set up in a few ways.  There are two most common.  An artist can set up individual lights to approximate the lighting in a scene, perhaps guided by photos and notes.  This can be successful, but for more complex lighting effects a special form of photography is used called “High Dynamic Range” photography.  This is a set of digital photos shot during production at very high resolution.  They usually are taken with a 180-degree fish-eye lens in two or more directions. They are shot in two or more directions to create a complete sphere of images to capture the entire environment. All exposures are captured, from completely overexposed to black.  These are then compiled in a special program or plug-in to make an H.D.R. set-up.  This will provide the exact lighting from the scene and reflection maps too.   The results of these renders are often striking in their realism.

Renders are created on black, and come with one or more mattes.  These mattes are used during compositing to prevent the rendered object from appearing transparent.  Special mattes are often created to control different parts of the render.  Compositing, or the combining of images, is an art.  Here the different elements of the scene are combined to form a convincing reality.  Here is an example of many different elements composited together to form the scene.

Here we see the players about to begin the volley.  The foreground player is shot wet, greenscreen.  The court is a render of a CG court.  The trees are on “Cards” digitally rendered, but come from photos.  The opponent is shot at high frame rate,  “Dry” on stage with water rippled lighting and air.  Bubbles shot as separate elements on a black background are added as highlights only and tracked to the racket, which has some too.  The rays are made in the Flame compositing platform, and are not 3D.  These are all color corrected to match, blurred and rippled, fog effects added as a function of distance from the CG camera, and blended with the image elements of the shot.  These are then rendered as a composite and the result of many iterations, taking up to a week or more for a complex shot, to yield this comp’.   Other things a compositor can do is move on the frame, like a push-in or slide-up, but these are limited by resolution because a shot will have to be enlarged first to allow room for movement.   They can also add vignettes for drama and focus, soften, diffuse, and add a myriad of 2D visual effects available as filters, “sparks” (special Flame dedicated effects plug-ins), elements like dust, rainfall, dirt, scratches, particles, and debris, or any of hundreds of other enhancements.  You can see how the possibilities are endless here too.

The final step in finishing a spot is often laying in titles, end cards, supers and legal text.  In this case, the final end animation sequence was provided to us, as is often the case, by the Client.  The Flame compositor superimposes the animation or image, does any adjustments necessary or requested, and adds legal text in the appropriate places and sizes, supers like website addresses, and any other graphics provided.  This completes the work, and final sound is laid over the picture is available at that time.  The final spot is laid to tape or uploaded, satellite transmitted, or burnt to DVD.

This spot, “Pocari Sweat” (the product’s name in Japan), took several weeks of post-production effects and about 12 top-level artists to complete.  This is not at all unusual, and can take much longer.  For an all CG animated spot that must look photo-real, the process can take two or three months.  The end result is a spot that is arresting.  It will stop a viewer (even on a TIVO!) and captivate their attention, be burnt into their memories as a compelling experience, and be a topic of conversation with others.  A spot that stands out due to amazing visual qualities can achieve so much that a normal one can’t, especially in the age of hurry-up and easy distraction.  Visual effects are one fantastic way to bring to a viewer an experience they enjoy and remember, and to create a brand that transcends the ordinary.  We relish these opportunities and collaborate on a creative and innovative level.  It’s what defines a good team, and we are team players to the end.

The end.