Every suite produces the same heartbreaking frame: the pose is the best of the day, the light is perfect, and the dress belongs to a different pack. Two tools claim it — the editor, which repolishes, and the generator, which recuts — and they cost differently, in credits and in risk. The repolish runs at edit prices and preserves everything you already approved. The recut costs a fresh frame and a fresh round of culling, and it may never land that pose again. Choosing well is a bench skill, and it reduces to one question: is the cut right?
Repolish when the bones are right
If the composition, pose and light would pass on their own, the frame has a surface problem, and surface problems are what the editor is for. Wardrobe changes, background moves, a grade shift to match the suite it is joining — one facet per job, briefed with explicit “keep identical” clauses for everything else. The frame’s cut survives untouched, which is precisely the point: you are not gambling the keeper, you are finishing it.
Recut when the bones are wrong
If the pose fights the crop, the face has drifted off her facet sheet, or the composition leaves the frame nowhere to go, no amount of polishing will save it — polish reveals a cut, it does not redesign one. Recut with a corrected brief: tighten the vague line on the sheet, simplify the scene to one room and one light source, and cut wide again. It feels wasteful for exactly one frame, until you price the alternative: three stacked edits chasing a frame that was never sound, each one destabilising a different facet.
The one-facet rule of thumb
Count the changes the frame needs. One change — repolish. Two changes in the same family (wardrobe and its matching backdrop) — repolish, in two jobs. Anything more, or any change to pose, face or composition — recut. And keep the economics honest: failures refund automatically on both paths, so the real cost is never the lost job; it is the hour spent polishing a stone that should have gone back to the wheel.