When users prompt for a "cinematic" look, what they're actually asking for decomposes into a small set of concrete camera/lens parameters: depth of field, lens compression, and light fall-off. Rather than relying on vague descriptive language, some tools (e.g. Higgsfield Cinema Studio 2's Image Mode) expose these directly as explicit selectors — camera body, lens, focal length, aperture — including preset combos for users unfamiliar with film gear, so the look is chosen deliberately rather than described and hoped for.
This is a specific instance of the broader principle behind Hard-Rule Prompting for Structural Constraints: naming the actual mechanism you want instead of a vague adjective gives the model (and the user) a concrete lever to pull rather than a hope to express.
Из тем: Unsorted, Prompting Foundations: Briefing the Model Like a Director