How do you choose the location image for a generated scene?
On Higgsfield: Higgsfield generates each location empty in Soul 2.0 at 16:9, with a fixed three-quarter camera and a clear open area for the characters, and leaves out people and animals so the reference never adds an extra figure.
Without Higgsfield: Make every location at a three-quarter angle, whose depth gives the camera movement something to grab onto, and pick from several batches the few places you would actually want to film in. Use the location image to lock lighting, layout and background before generating, choosing one with room for the action, such as a long pier that keeps the middle clean.
Axis: pick the location by the space the action needs ... pick it by its look
Answers by date: 7 records in 6 sources, newest first; the answers above are the current ones. The advice was refined (2026), never reversed, in these records.
| date | who | answer | source |
|---|---|---|---|
| 2026-10-24 | Adil | For the alley location Adil asks for a three-quarter view for depth and a 24mm anamorphic look (wide framing, softer edges, oval bokeh); Claude expands it into a prompt he generates in Soul Cinema. | The Complete AI Filmmaking Masterclass Everyone’s Missing (Full Tutorial) 00:04:40 |
| 2026-10-22 | a guest | For a location used in only one shot (the dressing room), the presenter skips a separate location image: ChatGPT uses the first room as a style reference and the new space is described in the prompt. | 5-Step Workflow: How to Make Viral AI Short Dramas (60M Views) 00:14:43 |
| 2026-10-16 | Adil | After three rounds of 10 street locations, Adil keeps one image from the latest batch and two from an earlier one as location references, and advises choosing places you would actually want to film in. | AI Creation: First Steps: Product Images and Ads 00:04:39 |
| 2026-10-13 | Higgsfield | Higgsfield's blog generates each location empty in Soul 2.0 at 16:9, with a fixed three-quarter camera and a clear open area for the characters, and excludes people and animals so a location reference never adds an extra figure. | How to Put Two Consistent Characters in One AI Video Scene |
| 2026-10-07 | Adil | Adil's pro tip: always make a location at a three-quarter angle; the depth gives the camera movement something to grab onto and gives better video renders than a flat head-on shot. | How To Prompt Like a Director Using Opus 5.5 + Seedance 2.5 (Full Workflow) 00:03:55 |
| 2026-10-07 | Adil | As on a live-action shoot, Adil uses image references to lock down the lighting, layout and background of a location before generating. | How To Prompt Like a Director Using Opus 5.5 + Seedance 2.5 (Full Workflow) 00:03:31 |
| 2026-08-21 | Adil | Adil picks the harbour image whose long pier leaves room for a crowd, a crash and a full sprint, with the sides holding the mess and the middle clean for the action. | How to Make a Wedding Love Story Using AI (Full Tutorial) 00:06:49 |