Portrait photographer reframes off-center shots with Photoshop Generative Expand

What they did

They cull in two rounds (first asking "if I were the model, would I have liked this picture?", then picking the best of near-duplicates) and share the shortlist in Google Drive so the model can remove any before editing starts. In Photoshop they widen the crop on the side that's too tight, choose Generative Expand with no prompt so it extends what's already in frame, and compare the three variations for stretched objects. They select distractions such as gate grills or a muddy patch and run Generative Fill with the prompt left blank, then merge each generated layer.

How it went

The subject ends up centered in a vertical frame and the distractions are gone. They pick whichever of the three variations looks believable, since some stretched a car or made odd structures.

Worth knowing

• They point out that the Photoshop AI features are paid. Variations often distort objects, so each one needs checking before merging.
• Price check: Adobe Photoshop: free plan unclear; Photoshop: US$34.49/month, or US$22.99/month on a yearly plan (US$275.88/year) (checked Oct 26, 2026).

Try it yourself

Paste this into a chatbot, then add your own details underneath.

I'm a portrait photographer. Give me a two-round culling checklist that looks at each shot from the model's point of view (pose, eyes, hair, outfit, sharpness), then a short message I can send the model with a shared folder asking them to flag any photos they don't want edited.

Original post

See The Tabletop Studio's original post on YouTube → (Sep 20, 2026)