Flexible Pipeline

The flexible mocap pipeline consists of the following steps:

• Extraction – Takes a set of markers as input and extracts a set of goals from them. The input (set of markers) is usually imported from a motion capture system and it represents the 3D position (without the rotation information) of the optical markers.
• Goals – Controls a character’s rig. They are used for solving to the final skeleton and to animate the actual character.

You can access the input, intermediate, and output data through these steps. This enhances the flexibility of the pipeline. The following figure shows the flexible mocap pipeline.

You can customize the flexible mocap pipeline at multiple insertion points.

You can modify the input markers for low-level corrections in the motion capture data. For example, you can constrain markers in different ways to enforce a given configuration or to correct artifacts such as, marker sliding. You can also perform complex operations on markers such as, combining several markers into new ones. Additionally, the motion capture hardware manufacturer's clean-up software can apply average or filtering schemes in real-time instead of applying them offline.

You can apply custom logic on the extracted goals. You can consider these goals as temporary IK goals, which the next solving step uses as input. It is better to apply correction in the high-level extracted goals than in the lower-level input markers or output bones. You can perform corrections in the whole character structure easily using the high-level goals than the low-level markers. You can use the default (relation) constraints to add custom logic to the extraction. For example, you can constrain the goals to not allow knee snapping. Additionally, you can use the story tool to add animation on top of the extracted goals, which is very useful. For example, you can use this to fix the mesh penetration issues. It is also possible to use custom constraints on the goals for any type of correction such as, controlling the way rotation is spread on each part of the spine. The constraints can either use the extracted goals as an input and modify them, or override the value from the extraction step.

You can also customize the solving from the goals to the bones. Usually, the solving is performed using the HIK solver or MotionBuilder's internal solver, but you can also use your own custom solver. As a post-processing step, you can modify or override parts of a solve output such as, an arm in a complex rig situation.