top of page

AI Video Shot Sizes: Wide, Medium or Close-Up?

Oct 16
2 min read

Updated: Nov 5

Shot size determines how much information a viewer sees at once. Choose it according to the message before deciding how the camera should move.


Wide shots explain location


Use a wide view when the relationship between a subject and its surroundings matters. Keep the main subject easy to identify; a busy background can overwhelm a short clip.


Medium shots show an action


A medium view balances context with detail. It is useful when the audience needs to understand what someone is doing and see the nearby objects involved.


Close-ups direct attention


Use a close-up for a texture, expression or small feature. The viewer may lose context, so establish what the detail belongs to when it is not obvious.


Choose a new shot size when the viewer needs new information

Use a wide view to establish where the subject is, then cut closer when a specific hand action or detail matters. Compare the two shots side by side and make sure the cut adds information.

Example: a ceramic workshop


A wide shot introduces the bench. A medium shot shows the maker shaping a bowl. A close-up highlights the rim. If generating from a single reference, do not assume all three views will preserve exact hands and object geometry.

Write the frame before the motion


A useful brief says “close-up of the bowl rim, centred, with a still camera” rather than only “cinematic shot.” Plan the frame in AI LEMNISCATA and inspect whether the generated result communicates the intended detail.


How do you choose just one shot?


Ask what the viewer must notice first. Choose a wide shot for location, a medium shot for an action and a close-up for a detail. If all three are necessary, plan a sequence rather than squeezing several instructions into one brief. Keep the same priority when checking the result: the important information should be visible without explanation.


Choose a shot from the viewer's question


For the same fictional bowl, “Where is it?” calls for the bench and room; “What is the maker doing?” needs hands and the bowl; “What does the rim look like?” needs a closer view. Write the question beside the shot before choosing the framing.


Check whether the answer is visible in a small preview without reading the brief. If a close-up is attractive but the object is unrecognisable, establish context in the preceding shot or widen the view. Avoid asking one frame to show both a tiny detail and an entire room clearly.


Put your plan into action


Bring your idea and reference photos to the video planner. Review the plan and available generation settings, then check the price before checkout. Generate and download the result on AI LEMNISCATA.


bottom of page