HiggsfieldImage-to-videoParameters & settings
Higgsfield Camera Presets: Dolly, Orbit, Crash Zoom
Higgsfield camera presets are one-click cinematic camera moves, such as Dolly In, Crash Zoom, 360 Orbit, Bullet Time and FPV Drone, that you apply to a start image to turn it into a short video. You upload the still, choose the preset, add a brief prompt for what happens in the scene and generate; the preset controls the camera so the prompt only has to describe the subject.
Higgsfield (higgsfield.ai) is built around one-click camera presets for image-to-video. Instead of describing a camera move in words and hoping the model understands, you upload a start image, pick a named move such as Dolly In, Crash Zoom, 360 Orbit, Bullet Time or FPV Drone, add a short prompt and generate. The preset handles the camera; your prompt handles what happens in the scene. For anyone who has fought with "slow push-in" in a text box and got a pan instead, that split is the whole appeal.
How camera presets work
A preset is a predefined camera behaviour that the model is tuned to reproduce reliably. When you select Crash Zoom, the output zooms hard into the subject; when you select 360 Orbit, the camera circles it. You still provide the image and a prompt, but the prompt no longer needs camera language, and in most cases it should not contain any, because text that argues with the preset produces confused motion.
The library of presets changes often. New moves are added, names are adjusted and some are retired, so treat any list, including the one below, as a snapshot. Browse the current interface before planning a shot around a specific name. Higgsfield works on a credit plan, and the cost of a generation can depend on the model, duration and resolution you choose; check the platform's current pricing page.
The main preset families
Presets group naturally by what they do to the camera. Names below are examples seen in the interface at the time of writing; the exact labels may differ.
| Family | Examples | What it does | Best for |
|---|---|---|---|
| Push and pull | Dolly In, Dolly Out, Crash Zoom | Moves the camera towards or away from the subject, slowly or violently | Reveals, emphasis, product heroes, dramatic beats |
| Orbit | 360 Orbit, Arc | Circles the subject while keeping it centred | Products, cars, characters, architecture |
| Time effects | Bullet Time | Freezes or slows the subject while the camera moves around it | Action stills, sports, fashion poses |
| Flight | FPV Drone, Fly-through | Fast first-person flight through or past the scene | Landscapes, interiors, architectural renders |
| Handheld and lens | Handheld, Dolly Zoom, Snorricam-style moves | Adds organic shake or shifts focal length during the move | Documentary feel, unease, music-video energy |
Choosing a preset for your image
The best preset depends less on taste than on what the image can support. Ask three questions:
- Is there a clear subject in the centre? Orbit, Bullet Time and Crash Zoom need one. A busy wide landscape with no focal point will spin or zoom into nothing.
- How much of the scene is missing? Orbits and drone moves reveal parts of the scene the still does not contain, so the model invents them. The more a move reveals, the more it has to make up, and the more likely you are to see warping. Dolly In and Dolly Out reveal very little and are the safest choices.
- What is the emotional job of the shot? Slow dolly in for intimacy or tension, crash zoom for punch, orbit for admiration, FPV for exhilaration. Match the move to the beat it will sit on in your edit.
Portraits take Dolly In and Bullet Time well. Products take 360 Orbit and Dolly In. Architecture and landscapes take FPV Drone and slow pulls. Fast action stills suit Crash Zoom and Bullet Time. Start there and deviate once you know how your images behave.
Settings at a glance
| Setting | What it does | Start with |
|---|---|---|
| Camera preset | The predefined camera move applied to your start image | Dolly In for your first test; it reveals the least |
| Start image | The still that becomes the first frame | A sharp image with the subject centred and room around it |
| Prompt | Describes what the subject and scene do during the move | One sentence, no camera words |
| Preset mix | Combines two moves in one generation where the interface allows it | None until single presets behave |
| Duration | Clip length; options depend on the model selected | The shortest option |
| Model | Which underlying video model renders the clip; names change | The current default |
Mixing presets
Higgsfield lets you combine presets, so one generation can, for example, dolly in and then orbit, or add handheld shake to a crash zoom. Mixing is powerful but it multiplies the amount of scene the model has to invent. Three habits keep it under control:
- Pair a big move with a small one. Orbit plus subtle handheld reads as a real camera operator. Orbit plus FPV reads as a fairground ride.
- Get the single preset right first. If Dolly In alone warps the subject, adding a second move will not help.
- Cut rather than combine when in doubt. Two short clips with one move each, joined in your editor, often look more professional than one clip trying to do both.
Common mistakes
- Camera words in the prompt. "Slow zoom out" with Crash Zoom selected gives the model two masters. Describe the subject only.
- Subject touching the frame edge. Any move that pushes in will crop it off. Leave margin in the still.
- Expecting the orbit to show a correct back view. The model guesses what is behind a person or product. Use half-orbits or arcs for anything where the back matters, or supply a start image that already shows a three-quarter view.
- Text and logos in the image. They smear during movement. Add them afterwards.
- Judging on one run. Presets are consistent but not identical between runs. Generate two or three and pick.
For the full still-to-clip process, including how to prepare images and organise takes, see the Higgsfield image-to-video workflow. If you want to create the start image inside the same platform, the Soul image presets guide explains the image side.
Try it
Start image: a single trainer on a wet concrete floor, neon reflections, centred, plenty of space around it.
Preset: 360 Orbit.
Prompt: Water droplets glisten on the shoe, faint mist drifts across the floor, neon light flickers softly. Everything else stays still.
Duration: shortest available.
Run it once as written, then swap the preset to Dolly In without changing anything else. Comparing the two tells you how much the orbit is inventing and whether that invention is acceptable for the product you are showing.