higgsfield-image-shots

To write cinematic prompts with specific camera angles, use shot framing and angle keywords like Low Angle, Dutch Angle, or OTS to shift power or tension in your AI image model prompts.

The best shot framing keywords for text-to-image models range from Extreme Wide Shot to Extreme Close-Up, letting you control how much of the subject versus the environment appears in the generated frame.

To keep character consistency across different cinematic shots, use the [img 1] token pattern in your prompt to map reference images, ensuring composition remains reliable across a character series.

You can make a still image look dynamic by applying implied motion keywords like dolly, pan, zoom, crane, or bullet-time style phrasing to the prompt instructions for your AI image model.

Yes, this cinematic prompting method works with Soul 2.0 and similar reference-based systems by using provided AI prompt keywords and the [img 1] token pattern to reliably structure prompts.

You should avoid using over-the-shoulder or POV shot keywords when your shot lacks the spatial depth required to support these angles, as they may distort the framing of your generated image.