In response to some trigger, the Crowd Transition DOP switches the agent to new state, and switches the agent from playing the clip associated with the old state to playing the clip for the new state.

The Crowd Transition node has simple controls for blending between the animation clips of the "before" and "after" states. If you need more control (for example, waiting for a certain frame to start the transition, or transitioning through intermediate clips between the start and end clip), you can set up a transition graph.

A transition graph stores detailed information about how to transition between different clips. For example, the graph might specify that the idle → walk clip transition should start at frame 16 and transition from the idle clip to the idle_to_walk clip, then to the walk clip.

Transition graphs are more work to set up, but give you much more control over transitions. If the default blended transitions look good enough at your level of detail, you don’t need to bother with transition graphs.

1. After you load/generate the agent primitive in SOPs, append an Agent Clip Transition Graph SOP to the agent. Set up the different transitions in the parameters.

2. In the crowd simulation network, on the Crowd Object DOP, set the Transition graph parameter to point to the Agent Clip Transition Graph SOP.

3. On any Crowd Transition DOPs for a state changes that should use the graph, turn on Use clip transition graph.

• The Agent Clip Transition Graph SOP stores the relationship graph as geometry, where points represent clips and polylines represent transitions, with data stored in attributes. This allows the possibility of creating your own custom tools to generate or edit the graph.