AI Video and Image Generation Glossary

Use this glossary to read a scene brief and distinguish creative vocabulary from the generation choices on SeaDance. Camera and style terms describe what to write; they do not establish extra controls or guarantee a particular result.

Text-to-video

A written scene description is the starting input for a new video. On SeaDance, Text to Video offers seven video models and produces one result per request.

Image-to-video

A still image establishes the starting view for a generated moving scene. Describe what should happen after that starting moment; Image to Video supports a start frame and, on models that allow it, an optional end frame.

Reference-to-video

Images guide the subject or appearance of a new video rather than simply describing it in words. SeaDance Reference mode accepts 1–9 images with Seedance and MiniMax models; the supported input is still images.

Text-to-image

A written description supplies the brief for a new still image. SeaDance offers four image model families; see the AI image generator for the available choices.

Image-to-image

An uploaded image accompanies a written instruction for an image result. This is the image creation workflow, separate from asking a still picture to become a moving video; image limits vary by model.

Prompt

The prompt is the written brief you submit with a generation request. A useful starting structure is a subject, an action, a setting and a camera direction, as discussed in the prompt guide.

Reference image

A reference image gives the brief a visible source for a person, object, setting or appearance. Explain which detail each uploaded image should guide instead of assuming all images have the same role.

First frame

The first frame is the opening view of a video. In Image to Video, the start image defines that starting point, so describe a continuation that makes sense from the pose and composition already shown.

Last frame

The last frame is the final view of a video. An optional end image is supported by the Seedance and MiniMax Image to Video choices, while Kling 3.0 accepts only one image; check the selected model before planning both endpoints.

Aspect ratio

Aspect ratio describes the relationship between picture width and height. SeaDance video requests currently use 16:9, so a prompt written for a tall composition still needs to fit a landscape frame.

Resolution

Resolution describes the output picture size rather than the amount of action in a scene. SeaDance sells Seedance video at 480p, Kling 3.0 at 720p and MiniMax H3 models at 768P; model specifications elsewhere on a page can describe broader capabilities than the settings sold here.

Duration

Duration is the length of the requested video in seconds. The current SeaDance video choices are 5, 10 or 15 seconds, and the credit quote changes with the model and duration.

Frame rate

Frame rate refers to how many individual frames make up a second of video. It is a separate concept from duration; these writing guides do not imply that SeaDance provides a frame-rate selector.

Seed

A seed is a term for an initialization value associated with a generation process. Knowing the term does not mean a seed can be entered in SeaDance, and these examples do not promise identical results from repeated requests.

Prompt adherence

Prompt adherence describes how closely a result follows the written brief. Compare the subject, action and framing separately when reviewing an output; the examples here are instructions to adapt, not measured adherence scores.

Negative prompt

A negative prompt names elements the writer wants excluded from a result. SeaDance examples may include plain-language constraints such as no captions in the main prompt; this is not a claim that a separate negative-prompt field is available.

Subject

The subject is the person, animal or object that carries the scene. Name a few identifying details that matter to the shot, such as an ivory coat or a bamboo basket, before describing surrounding decoration.

Action

Action is the visible change the subject makes during the scene. A lid rising or a courier sliding a letter supplies a more concrete event than a general request for an exciting moment.

Composition

Composition is the arrangement of subjects and background within the picture. Describe where the important object sits and what needs room around it, especially when a racket, hand or coat moves toward the edge.

Close-up

A close-up concentrates attention on a small part of the scene. Use it to describe a face, wrapper texture or product surface, while leaving enough of the action visible to explain what is happening.

Wide shot

A wide shot gives the subject more surrounding space. It can establish a courtyard or keep an entire sporting gesture in frame, but its emphasis differs from a close study of tiny surface details.

Dolly

Dolly describes a camera position moving through space toward, away from or alongside a subject. In a written prompt, specify a short direction and purpose, such as approaching a coat to reveal its folds; this is camera wording, not a dedicated control.

Pan

A pan turns the camera view sideways from a fixed position. Describe what the turn reveals, such as moving attention from a traveler to a departing boat, rather than adding the term without a destination.

Tilt

A tilt turns the camera view upward or downward from a fixed position. Use it when the change in viewing direction matters, for example moving attention from a stone table toward the person standing behind it.

Orbit

An orbit describes the camera moving around a subject. Give the arc a limited extent and explain what remains central, rather than combining a full circle with several unrelated camera moves in a short scene.

Tracking shot

A tracking shot follows movement through a scene. A short lateral track can follow a walking subject without breaking the scene into several separate shots.

Handheld

Handheld describes a camera feel with small movements associated with a person holding it. A travel diary can ask for gentle motion tied to footsteps; the wording does not require a motion reference upload.

Lighting

Lighting describes where illumination comes from and how it falls across the subject. Linking side light to silk folds or lantern light to a letter gives a visual reason for the instruction.

Depth of field

Depth of field describes how much of the scene appears in focus. A prompt can ask for a softly blurred street behind food, keeping the intended focus on the basket; that wording is a visual request rather than a camera setting guarantee.

Credits

Credits are the balance used for SeaDance generation requests. New accounts start with 0 credits, so purchase a plan or credit pack before generating and check the quote for the selected settings.

Generation refund

A generation refund restores the credits for a failed generation. It is distinct from a refund of a purchase; a successful backup generation keeps the quoted credit price, as explained in the Seedance 2.0 pricing article.

Continue with Seedance 2.0 prompts or Seedance 2.5 settings.