Mocap data can bring realistic and expressive motion to your 3D animation projects, but it also comes with some challenges. How do you clean and edit the raw data to make it fit your artistic vision and technical requirements? In this article, we will share some best practices for mocap data cleaning and editing, based on our experience and industry standards.
There are many software tools available for mocap data cleaning and editing, each with its own features and workflows. Depending on your budget, preferences, and needs, you may want to use a dedicated mocap software, such as MotionBuilder, or a general 3D software, such as Maya or Blender, with mocap plugins or add-ons. Some factors to consider when choosing the right software are: the compatibility with your mocap hardware and format, the ease of use and customization, the quality and speed of the results, and the integration with other tools and pipelines.
Make sure to chose the right mocap hardware for the right application. Do your research into the different mocap technologies with their upsides and downsides. There is a big difference in data quality, which can lead to a high level of unwanted clean-up that would have easily been prevented by choosing the optimal setup.
One of the first steps in mocap data cleaning and editing is to filter out the noise and errors that may occur during the capture process. Noise can be caused by factors such as sensor interference, occlusion, calibration errors, or human errors. To filter out the noise, you can use various techniques, such as smoothing, gap filling, outlier removal, or resampling. The goal is to reduce the jitter and spikes in the data, while preserving the natural motion and detail. You can use different filters for different parts of the body, depending on the level of noise and detail.
Another important step in mocap data cleaning and editing is to refine the pose and timing of the motion, to make it more accurate and expressive. This may involve adjusting the joint orientations, the root position and rotation, the contact points, the balance, the weight distribution, and the transitions. You can use tools such as retargeting, inverse kinematics, keyframe editing, or curves editing, to modify the pose and timing of the motion. The goal is to enhance the realism and quality of the motion, while respecting the style and intention of the performance.
Sometimes, you may want to blend and mix different mocap data sets, to create a more complex or diverse motion. For example, you may want to blend two different walks, or mix a walk with a jump. To blend and mix the data, you can use tools such as blending, layering, stitching, or blending graphs, to combine and transition between different motions. The goal is to create a seamless and natural motion, while avoiding pops, glitches, or discontinuities.
The final step in mocap data cleaning and editing is to optimize the data, to make it more efficient and compatible with your 3D animation project. This may involve reducing the file size, the number of keyframes, the number of joints, or the complexity of the motion. You can use tools such as compression, simplification, baking, or exporting, to optimize the data. The goal is to improve the performance and quality of the animation, while minimizing the storage and processing resources.
The best way to ensure that your mocap data cleaning and editing is successful is to test and iterate your results. You can test your results by playing back the motion in different views, angles, speeds, and contexts, and checking for any issues or inconsistencies. You can also test your results by applying them to different characters, rigs, or environments, and seeing how they look and behave. You can iterate your results by making changes and improvements based on your feedback and evaluation. The goal is to achieve a satisfactory and final result that meets your expectations and goals.
When you talk about clean-up, the HD Reprocessing feature from the Xsens motion capture systems is a game changer. It fixes contact points, sliding and sinking, among others.