Mocap (Alpha)

This feature takes poses from images or video sequences and applies them to the character in the scene. This way, you can quickly create animation drafts.

To use Mocap:

1. Prepare  a sequence of images.
You’ll need several image files with different poses on them. Video files are not yet supported.

2. Select Add → Plane from the Commands menu to add a plane to the scene:

3. Select the newly created plane.

4. Click the Bind texture button on the Toolbar:

5. Select the image sequence you’ve prepared.
The sequence will be applied to the plane. Now, you can use it to recognize the poses.

6. On the Timeline, select a part where you'd like to place poses from the reference.
The number of captured poses will be equal to the number of selected frames. If you, for example, select three frames, you'll get three poses even if there are more frames on the reference.

7. On the Toolbar, click the Mocap button:

If you are using this tool for the first time, the program will need to download an additional package about 700 megabytes of size:

This is done to reduce the size of the main Cascadeur install file.

The package is located in the resources subdirectory of the main folder the program uses for storing application data (see Application Data section of the File Structure page to learn where this directory can be found):

The package itself is a single file called mocap.pt:

If you don’t plan to use the feature anymore, you can simply delete this file.

8. Wait until images are processed.
Recognizing poses is not an easy process and can take some time.

In the end, you should get several keyframes, one per an image. Each one of these keyframes will contain the pose from the corresponding image.

Limitations

Sometimes, Mocap feature might produce results that don’t fully coincide with the poses on the images you feed into it. The main reason for this is that 2d images only show poses from one angle, so the tool might not be able to recreate every nuance of a more complex pose.

Also keep in mind that recognizing poses can be a time consuming process. Often, it takes about 20 seconds per frame.