
| Function | Flame Shortcut | Smoke Classic Shortcut |
|---|---|---|
| Create Circle GMask Selective Preset | Space - S | Not assigned |
| Create Square GMask Selective Preset | Ctrl - Space - S | Not assigned |
| Create Horizontal Line GMask Selective Preset | Shift - Space - S | Not assigned |
| Create Vertical Line GMask Selective Preset | Shift - Ctrl - Space - S | Not assigned |



The Semantic keyer uses a machine learning model to automatically extract a matte out of skies, human bodies or heads.







The Motion mode requires access to Motion Vector data. The data can come from a Motion Vector texture map (if connected to the Surface) or from the Action Motion Vector output (if connected to the Camera). The Motion mode isolates part of the Motion data using direction and motion min/max fields, and then create a matte out of it.
Since the process of creating Motion Vector results includes a lot of small motion, you'll probably need to increase this value to get a precise motion segmentation result.

Use a primitive to precisely extract an object from the depth, something that cannot be done with other depth-based segmentation tools.