Color & LUT: name the grade, not the vibe
A cinematic prompt that says "moody grade" or "teal-and-orange" without the rest of the brief asks the model to color in a feeling. A prompt that names the LUT direction — what is rolled off cool, what is pushed warm, what is crushed — gives the model a reproducible color script it can act on frame after frame.
Choose the LUT family first. Teal-and-orange pushes highlights cool and skin tones warm, classic anamorphic theatrical. Low-contrast lifted blacks softens overcast and atmospheric frames. High-contrast crushed blacks seats noir interiors. Then write the pair: "cool highlights, warm skin" or "lifted blacks, gentle roll-off" — not "cinematic." Midjourney, Ideogram, and ChatGPT all reproduce the LUT direction when it is named explicitly.
Aspect ratio & anamorphic lens
Aspect ratio is the cheapest cinematic lever — and the most often left out. 2.39:1 is the anamorphic theatrical crop, the look of modern widescreen. 1.85:1 is the classic theatrical crop, slightly wider than 16:9. Aspect ratio carries the screen feeling even before the lens does.
The anamorphic lens is what carries the lens flare streak. A 25mm, 32mm, 40mm, or 50mm anamorphic stretches every highlight horizontally — every practical light, every backlit sun source bleeds sideways across the frame. The flare streak is the look's signature, not a defect to clean up. Midjourney, Ideogram, and ChatGPT all respond when you pin the aspect ratio and the anamorphic lens together — they read "2.39:1 + 25mm anamorphic" as a single cinematic language, not two separate nouns.
Mood, depth, and lens flare
Mood is what the depth and flare carry. Blue-hour exterior reads as noir-blue because the sky is still lit and the streetlamps are on. Backlit golden hour reads as warm and saturated because the sun is on the frame edge. Low-key interior reads as compressed and deliberate because the shadows hold shape. The mood is downstream of the lighting and the lens — not the caption.
Depth-of-field is the third cinematic lever. T1.4 to T2.0 dissolves a shallow subject plane. T2.8 keeps the subject sharp but softens the midground. T4 to T5.6 freezes the whole frame for establishing shots. Pair the depth with the lens — a 25mm wide open is a different image than an 85mm wide open — and name them together. Midjourney, Ideogram, and ChatGPT all separate the cinematic vibe into aspect + lens + depth + LUT + mood when each one is named in the prompt.