|Casting for Motion Capture|
DETERMINE CASTING REQUIREMENTS
Casting for Motion Capture performers (the 'talent') is very similar to traditional casting, except that it is less important (although still very desirable) to cast for acting ability.
Do not cast people for what they look like facially, unless it is for facial capture.
Cast people by their body type and by how well they perform the part.
When casting for Motion Capture talent, it is essential to cast for individuals that have excellent body control and a highly developed sense of physical awareness AS WELL as an ability to act the role.
While every resource that is typically employed to cast a traditional production (such as casting agencies or trade publications) can and should be utilized to find good performers, care must be taken to specify and cast for the critical skills necessary to perform the role.
It has been my experience that the more professionally someone does something with their body, the better the motion capture talent they are.
What do I mean by that? Well superstar athletes are so practiced at what they do they are almost robotic in their precision in executing movements. Other people who use their body as the tool of their profession are likewise usually good motion capture talent.
Categories of people to audition can include (but should not be necessarily be limited to):
· Stunt people
· Martial Arts experts
· Gymnastics experts
· Professional Athletes (usually pro athletes are 'special case' talent cast specifically to perform some athletic motions)
· Costume puppeteers (People who regularly make their living acting within large costumes)
· Pro Wrestlers
· Fight Experts
· Theatrical Stage Actors (who possess an excellent physical self awareness)
· Actors used to working with green screen.
Unlike traditional film making, it is possible, and often necessary, to cast several performers in the same role
Why is this? Because one person may not be able to do everything that the character is called upon to do in the script.
(in the example given later in the casting call sheets, we may have to cast one actor in the main role of the child phase of the character, with one or two actors cast for teenage and adult phases of capture, while yet another single mocap actor might be able to provide the dancing for all three ages. There is even the possibility that one person could fulfill the requirements of all three parts and only one actor would be required. Mocapped callback testing will help to determine this later on, if there is any doubt)
It is preferable (but not critical) that all the actors who will be cast in the same role be roughly the same height and body dimensions as one another.
This just makes for less tweaking and less overall data massaging which will be required to make different peoples data fit one single CG model.
Any differences between the actors and the character that they play will make for some amount of retargeting to make the character reach the same goals (do the same thing). If several actors play the same character, the more differences between the actors the more retargeting will also be necessary. The best case scenario is to have an actor’s captured motion go back onto a digital version of themselves. The more the actor differs from the character the more it can make work downstream in the pipeline.
IMPORTANT ITEMS OF NOTE:
The motion capture system is so accurate that it can reveal imperfections in people's skeletal or physical structure which may not be visible or readily apparent to the naked human eye. Old injuries, such as knee problems, other sports injuries, previous surgeries or back or skeletal problems, even though they may have healed long ago WILL SHOW UP IN THE DATA.
It is important that all talent disclose any previous physical ailment or injury which caused impairment to the person’s ability to move AT THE TIME THE PROBLEM OCCURRED, no matter how long ago the injury or problem may have occurred.
While it is perfectly acceptable to use previously injured or physically impaired talent, it is important to know about the problem in advance so that it can be examined and compensated for, or to screen out talent with bad physical traits, if the problem effects the persons movement (and therefore their data) in an undesirable way.
The Casting process should be carried out exactly as it would be for a traditional film role.
Create a ‘usual suspect’ character lineup with ‘real world measurements’ and how they would relate to normal human heights and weights, use this in casting.
It is important to keep relationships correct amongst actors to make your life easier. The closer the actors match in height relationship to the character they are playing the less overhead there is with things such as eye-line being wrong, contact between characters being off, or character to set ratios being hard to figure out.
While it’s impossible to cast cg character sized people in every case, you can cast someone who is proportionally shorter than the others, to play a child role and someone who is taller than that person as the parent. As long as the ratios to one another are as close as possible to the ratios that would exist size-wise between the characters,…when the parent character interacts with the child character, it won’t cause as many problems. Decide on a range of heights that might work for the shorter character and send this out to casting agents as part of the character spec sheet ‘eg. Looking for actors to play the role of a child. Actors could anywhere from 3’6 to 5’ with shorter people preferred’)
Once you weed out the first round of potential actors, have a second round, ‘kabuki callback’
Kabuki callbacks are something we developed to determine who was a good physical performer, and who was not. Kabuki callbacks are called this because a large white sheet is hung behind the performer, and a bright, diffused light is shone that illuminates the entire sheet.
During the callback you videotape the performer doing their audition in front of the sheet facing you. While you will be able to look directly at the performer and see their facial features, the video camera will only record their silhouette. When you watch the video playback it will be easy to see who is a good physical performer and who was acting with their face.
I found that this was important because as humans, we all tend to look at one another’s faces to judge what they are doing or saying. While we would really like some actor or other, when we would watch the kabuki callback tape we would find out that the actor we had like had merely been standing still acting with their face, while other actors would really stand out as acting with their whole body.
there are multiple actors who are ideal for a part, after second round casting, you might still want to have a fully mocapped callback.
It may be that the only way to tell who gets a part is to see them act it out as the character. In this case have as few people as is necessary come back and do a full audition while being mocapped. Since it is a pain to capture a whole bunch of people in this way, try and keep the number of people who need this kind of callback to a minimum.
It should really only be required when more than one person is very good at motion capture acting and are suitable for the same role. Seeing each one act the part and applying the motion to the character can sometimes make the decision much more of an obvious choice.
Motion Capture is so precise that it is possible to recognize people. This is most evident when more than one person’s motion is applied to the same character.
Whenever star talent is captured, it is advisable to have them play themselves as much as possible. It is possible to tell when more than one person plays a role (but this is not a ‘deal breaker’ and is just something to keep in mind).
Obviously this is not possible in all cases, especially when stunts are involved.