A Cell is a volume in space that is automatically generated by the Umbra 3 Optimizer. A single cell determines a volume where all the points share similar visibility.
Gate is an object that can be opened and closed runtime so that its effects in visibility changes. When a gate is closed, it cannot be seen through and when it is open it does not block visibility. Typically doors and windows are used as gates.
The most commonly used visibility query of Umbra 3. A portal visibility query uses occluder data, generated by the Umbra 3 Optimizer, to generate a list of visible target objects given a camera transformation.
Occluder is an object that can be used to hide other objects behind it. Good occluders are typically for instance walls, ceilings, floors and other large opaque objects.
Optimizer is the Umbra 3 SDK component used to generate occluder data from arbitrary input geometry. The Umbra 3 Optimizer is typically integrated into the toolchain or editor.
Umbra 3 Runtime is a small library used to perform visibility queries using the occluder data generated by the Umbra 3 Optimizer. The Umbra 3 Runtime is integrated into the rendering engine.
Target is an object an object that can be hidden by the occluders in a scene. Most of the objects in the scene should be targets. Umbra 3 visibility queries return lists of visible targets.
The occluder data generation process subdivides the input geometry into jobs which can be cached and processed individually. These jobs are called Tiles.
Umbra 3 Tome is a container of the occluder data generated by the Umbra 3 Optimizer. The occluder data of any given 3D scene may be divided into a single Tome or multiple Tomes, which are used by the Umbra 3 Runtime to perform visibility queries.
An axis-aligned box placed in an Umbra::Scene that represents the volume in space inside which the in-game camera is allowed to move. Placing a view volume may potentially yield significantly smaller occluder data, as well as speed up the occluder generation process. However, performing visibility queries from camera locations outside view volumes produces undefined results. There is no limit on how many view volumes can be placed in a scene.