OptionalformationsFormation design ids framed together with those nodes.
Each id must name a formation the shot compiled. It contributes the unit's slot footprint under the cue playing at the framed instant, widened by a member's radius and raised by a member's height. Only a frame subject, a frame focus or a coverage subject may name one. Omit it for a group of staged nodes.
Selects a multi-subject extent as the target representation.
Staged scene node ids, cameras included, taken together.
A target that is several things at once, such as a two-shot or a crowd.
A camera frames the union of members' measured portable rest/proxy boxes at staged placements and sampled formation footprints, with rig/stand-in fallbacks for missing model extents. This does not measure deformed rendered pixels. A compact formation is addressed by its design id rather than by expanded scene nodes. The camera follows a group only when a formation has a cue, and then updates the base while keeping opening height/radius; member nodes remain at staged positions in this group measurement.
Every other verb (
lookAt,reach,locomote, apointorstrikegesture, alaunchaim) needs one point on one body. For those, a group that names any formation is refused, and a node-only group resolves to the centroid of every named member. A missing member makes the target unresolved rather than changing the group by omission. An empty group, a blank id or a formation the shot did not compile is refused. A launch with onHit needs one struck actor, so a group aim is allowed only without onHit.