The mocap recordings
We had the fortune to borrow a motion capture studio from the university of Aalborg for our recordings. The studio has a complete Vicon system with 8 cameras and a suit for the mocap actor. After having made several tests to learn the process of mocap, we spend 3 days recording the different takes required in the film.
In this picture you can see Sira, our actor for all the mocap recordings, acting as the dad on a bicycle. Benjamin, one of our animators is steadying the bicycle. The cameras can be seen recording in the background.
Vicon Clean Up
On the picture below you can see what each camera sees during a recording. The cameras are adjusted, so they only can see the extremely light reflexive markers on the mocap suit.
The Vicon program is used to compare the different cameras' recording of the markers. The markers are then converted from the 2D space of the individual camera to the 3D space of the combined cameras. Each marker is then placed on an according area on a 3D skeleton, so the program knows where each marker belongs to the body at alle times.
Motionbuilder Clean Up
After having created a workable 3D version of how the markers are moved around during a recording, we imported the data into the program Motionbuilder.
Motionbuilder is used to clean up the data and set the markers to a Skeleton of the final 3D model created in Maya.
In the picture below you can see the various stages in Motionbuilder.
-
To the left are the recorded markers as imported from the Vicon programs.
-
In the middle the markers have been placed on a dummy body to see how a human being looks in motion. Here there is potentially a lot of clean up if the recordings have made any mistakes, like switching two markers around if they are close to each other.
-
To the right a skeleton of the Maya model has been attached to the dummy body and now moves according to the first markers.
In the picture below you can see the process of placing the dummy body exactly where the recording markers are. As each area of the dummy body is bound to the markers the body follows the markers around.
Maya Model
After the marker data has been attached to the Maya skeleton, the skeleton is imported to the Maya model. In the picture below you can see the model with the skeleton under the skin.
The model in the scene
The model is then inserted into the scene with props and scenery. As you can see there are several handles in the model, which are used by the animators to adjust the recordings to the surrounding environment.
The final lay-out
After a long process the model can finally be seen animated inside the finished environment. The shot below show how the dad is seen in the final movie - though without textures and correct lighting, which still have to be added. Tjeck out our discussion on the 3D pipeline for more details on this.
| 
