A finishing technique: the final shot doesn't have to come from one generation — it can be composited from multiple separate generations, taking the best-performing lead actor from one take and the best-performing supporting actor from a different take. Builds on the multiple takes produced by Actor-Directing Prompting Method (Takes and Incremental Notes), and is distinct from Multi-Model Specialization and Compositing (which composites across different specialized models rather than across takes).
Rather than compositing from a small number of takes, generate a burst of five to seven unique shot variations for a single scene beat, then hand-pick the strongest individual takes and assemble the final sequence from across the burst.
For any beat where performance quality matters, generate the beat five to seven times as distinct variations before selecting, rather than iterating on a single take in place.
Compositing doesn't have to stay within one generation's set of takes. When no single full generation nails every actor's performance, pull the best-performing lead from one generation and the best-performing supporting actor from a different generation of the same shot, then composite the two together into the final frame. This extends the idea from 'pick the best take of a shot' to 'assemble the best shot out of parts of several separate generations.'
Из тем: Unsorted, Directing Performance, Dialogue, and Sound