Text, Image, or Video: Choose the Right Starting Point

Separate decided information from open questions#

Before preparing a prompt, write two short statements: what the work must preserve, and what you are willing to invent. This establishes the role of the source material.

Suppose your idea is a tiny greenhouse on a city rooftop. If its shape and setting are open questions, text is a sensible beginning. If you have drawn the greenhouse and care about its curved windows, the drawing is valuable visual evidence. If you mainly want the feeling of a camera rising from street level to reveal a rooftop, a video reference can communicate that movement.

These are three versions of one idea, but the uncertainty differs. The strongest input is the one that makes the essential requirement explicit. More source material is not automatically more helpful; references that disagree can make the brief harder to interpret.

Use text when the scene is still a proposal#

A text-first brief should name a subject, a visible action, and a setting. Add the camera's role only after those are clear. An illustrative direction might be: “A gardener opens the curved door of a small rooftop greenhouse. Inside, warm light falls across rows of seedlings. The camera watches from the path outside.”

This description leaves room for visual exploration while defining an event. It is more useful than a cluster of adjectives such as “beautiful, cinematic, magical, detailed,” because it explains what the audience will see changing.

In MyloStudio, text can lead directly to video or first to a still image. The image route is useful when you want to decide the greenhouse's appearance before introducing motion. Reviewing a still lets you resolve a crowded composition or an unclear door without simultaneously judging movement, timing, and sound.

Choose direct text-to-video when visual exploration is part of the task and matching a particular existing design is not the primary requirement.

Use an image when appearance is the anchor#

An image supplies relationships that are tedious to reconstruct verbally: the gardener's position, the greenhouse proportions, the doorway's shape, and the distance to the camera. Your motion description can then focus on change over time.

For example: “The gardener gently opens the greenhouse door and pauses at the entrance. Leaves inside move slightly. The camera remains outside, keeping the doorway visible.” This is a desired direction, not a guarantee of how a generation will behave.

Inspect whether the source actually shows what the movement requires. If the hand is hidden behind a wall, opening the door requires inventing a hand, its contact with the handle, and the intervening motion. You can reduce that uncertainty by changing the action or editing the image first.

MyloStudio's image editing and image-to-video capabilities make this preparation meaningful. Instead of repeatedly explaining a missing element in the video prompt, you can develop a source image that presents the important action clearly.

Use a video reference for a temporal idea#

Some requirements become clear only as they unfold. A slow reveal, a sudden pause before an action, or a camera movement around a subject is a temporal idea: its meaning depends on a sequence rather than one frame.

MyloStudio can create new work based on the style, motion, or camera approach of a video reference. Assign the reference a specific job. “Use the rising reveal as the camera idea; create an original rooftop greenhouse scene” is more precise than “make something like this.”

Do not treat every detail in the reference as a requirement. Its setting, people, colors, and objects may be irrelevant to your scene. Naming what to borrow and what to replace reduces conflict between your own idea and the source.

If you already filmed the gardener and need the greenhouse's appearance changed, that is an editing task. The question is what must stay recognizable in the original action, not how to reinterpret the reference as a new scene.

Change inputs when the problem changes#

A project does not need to remain on its initial route. A text idea can become an approved image, which then becomes a video scene. A video reference can establish camera intent while a separate image defines the subject's appearance.

When combining material in MyloStudio, describe these roles explicitly. A character reference answers who appears; an environment reference answers where; a motion reference answers how the view develops. If two sources contradict each other, choose which one controls the disputed detail before generating.

The useful question is always the same: which uncertainty am I trying to resolve next? Start your scene in MyloStudio with the text, image, or video reference that answers that question most clearly.