Lesson Content

Specifying both camera and subject motion.

A prompt can specify both camera movement and subject movement: "camera tracking alongside a running horse, tracking shot keeping the horse centered in frame, background blurring past." The model attempts to honor both motion types, though coordination between camera and subject motion is more complex than either alone.

Reliable vs. complex combinations.

Some combinations are more reliably generated than current AI models can produce well: simple tracking (camera following a moving subject) and simple subject motion with a static camera are more reliable than complex compositions like "camera orbiting while subject also rotates on its axis." Simpler combinations first, adding complexity where the model demonstrates it can produce them reliably.

Managing motion complexity.

Adding too many simultaneous motion descriptions can overwhelm the generation – the model may execute one motion element reasonably while effectively ignoring others, or produce confused output that doesn't clearly match any of the specified movements. When multiple motions are essential, specify the most important one first (the model typically prioritizes earlier prompt elements).

Counterpoint for visual interest.

Cinematographic counterpoint – camera moving in one direction while a subject moves in another, or camera zooming out while a subject walks toward it – creates dynamic visual tension. Including this intentionally ("camera slowly pulling back as the figure walks toward camera, creating increasing apparent size") can produce striking visual results when the model executes it.