What’s new
- One look only, and it's fixed: the API docs state Soul Cinema's style is fixed and any client-supplied style parameter is ignored — the cinematic look is baked into the model, unlike Soul 2's swappable presets.
- More convincing grain and depth: Higgsfield describes the grain as closer to real film stock than digital noise, with foreground/background separation and material detail built to cinematography standards.
- Understands different film eras: it distinguishes the look of the 1970s, the 1990s and today without you having to spell out what each era should feel like.
- Works directly as a video keyframe: Higgsfield recommends feeding a Soul Cinema image into video models like Kling or Seedance as the opening frame.
What it’s good for
- Storyboard and keyframe previews, and mood references for film projects.
- Visual concepts for music videos and commercials.
- Gallery-grade artwork that needs a cinematic look.
Prompting tips
- Describe the scene, subject and mood: that's Higgsfield's own guidance — no need to stack cinematography jargon on top.
- Use a trained character ID for a recurring face: it shares Soul 2's Soul ID mechanism — train once, then call the character by ID with an adjustable 0-to-1 strength.
- Turn on auto-enhance when you're not sure: if your own prompt is thin, the enhance toggle fills in detail for you.
- Name the light source: terms like twilight, blue hour or soft-focus carry more cinematic information than just "nice lighting."
- Save the seed for a result you like: the same seed with the same prompt gets you back to something close to a result you liked.
Blue hour after rain in an old town street, a man in a trench coat seen from behind gazing into the distance,
35mm prime lens look, film grain, vignette, close shot, quiet and subdued mood.
Known limits
- No direct reference-image upload: keeping a character consistent relies on a trained Soul ID reference, not attaching a photo as a composition reference.
- No swappable style presets: the API docs state a client-supplied style parameter is ignored — the cinematic look is fixed.
- Batch size is only 1 or 4, with nothing in between.
Using it in Nomi
On the generation canvas, add a image card, pick “Higgsfield Soul Cinema” in the model picker and choose a mode (Text to image). Add your references, write the prompt and generate. In the storyboard you can pick it per shot. The first time, connect any one of the providers above under Settings → Models.
