Outpainting

What is Outpainting?

Outpainting extends an image beyond its original edges by generating new content that naturally continues the scene: revealing more of the environment, widening a composition, or adapting an image to a different aspect ratio.

At a glance

Also known as
Canvas extensionImage expansionUncropping
Used for
Extending images to wider aspect ratios without croppingRevealing more of a scene's environment around a subjectBuilding larger visual environments from a central image anchorAdapting portrait or square images to cinematic or landscape formats
Common tools
Stable diffusion (outpainting mode)Dall·e (canvas extension)Adobe firefly (generative expand)ComfyUIMorphic canvas tools
Related terms
InpaintingMasking / maskAspect ratioDiffusion modelCompositingCanvas

Ready to create?

Direct scenes, design characters, and ship full films

All-in-one AI creative platform with simple, transparent pricing, no speed throttles, and an infinite Canvas for max creativity.

Try Morphic
How it compares

How it compares

Compared with related concepts

Outpainting and inpainting are complementary techniques that work in opposite directions. Inpainting fills or replaces content within the existing frame: removing unwanted elements, repairing damaged areas, replacing masked regions with generated content. Outpainting extends the frame beyond its existing borders, generating new content outside the original canvas area. Both use the same underlying masked generation approach, but inpainting operates within the existing image boundary while outpainting operates beyond it. A complete generative editing workflow may use both: inpainting to refine content within the original frame and outpainting to expand the frame to the required dimensions.

Think of it like…

Outpainting is like unrolling a scroll to reveal more of a painting that continues beyond the visible portion: the new area that emerges was not in the original view, but it belongs to the same world, lit by the same light, following the same perspective, as if the artist always intended the wider composition.

Pro tip

When outpainting to change aspect ratio, guide the extended regions with a descriptive text prompt that specifies what the expanded area should contain. Even a simple extension like 'add sky above' or 'continue the stone wall and floor to the left' significantly improves the relevance and quality of what the model generates beyond the original border, compared to leaving the prompt empty and allowing the model to generate whatever is statistically probable from the edge content alone.

Types and variations

  • Directional outpainting extends the image in a single direction ( left, right, up, or down ) most commonly used for aspect ratio conversion or targeted scene expansion.
  • Omnidirectional outpainting extends simultaneously on multiple sides, growing the canvas outward in all directions.
  • Aspect ratio conversion outpainting specifically adapts an image from one format to another ( square to widescreen, portrait to landscape ) using outpainting to fill the new canvas area.
  • Iterative outpainting builds progressively larger environments by repeatedly extending the canvas, each generation anchored to the previously expanded edge, allowing the construction of environments far larger than any single generation could produce.

Ready to make your first scene in Morphic?

Try Morphic

Common use cases

Outpainting is used to adapt social media images (typically square or portrait) to cinematic widescreen ratios for film or video use, to extend product or portrait photography compositions to wider frames that better suit publication or campaign layouts, to build expansive environmental backgrounds by progressively extending a central scene reference, to reveal contextual surroundings around a tightly cropped source image, to recover or reconstruct parts of an image that were cropped during original capture, and in AI generation workflows to create larger compositional canvases from which specific regions are extracted for use in production.

Ready to create?

Direct scenes, design characters, and ship full films

All-in-one AI creative platform with simple, transparent pricing, no speed throttles, and an infinite Canvas for max creativity.

Try Morphic

FAQs

What is outpainting?
Outpainting is an AI technique that extends an existing image beyond its original borders by generating new content outside the canvas edges. The generated content is conditioned on the original image to ensure continuity of lighting, colour, perspective, and visual style, seamlessly expanding the scene rather than adding visibly disconnected new content.
What is the difference between outpainting and inpainting?
Inpainting fills or replaces content within the existing image frame: removing objects, repairing areas, or replacing masked regions. Outpainting extends the image beyond its original borders, adding new canvas area and generating content to fill it. Both use masked generation, but inpainting operates inside the boundary while outpainting operates outside it.
Can I use outpainting to change the aspect ratio of an image?
Yes, this is one of the most common practical applications. Outpainting can convert a square or portrait image to a widescreen ratio by extending the canvas horizontally, or convert a landscape image to portrait by extending vertically. The extended areas are generated to continue the scene, providing a natural-looking result that avoids the distortion of scaling or the loss of content from cropping.
How do I control what is generated in the extended areas?
Most outpainting implementations support a text prompt that guides what is generated in the extended regions. Describing the content you want to appear — 'stone corridor continues to the left', 'clear blue sky above', 'urban street scene extends to the right' — significantly improves the relevance and quality of the extended content compared to prompt-free extension. More specific prompts produce more controlled and useful outpainting results.
What makes outpainting results look seamless or obvious?
Seamless outpainting results from the model accurately reading and continuing the edge content of the original image: matching light direction, colour palette, perspective, and environmental logic. Obvious or jarring extensions occur when there is a mismatch in any of these elements: incorrect light direction in the extension, perspective that doesn't match the original's vanishing points, or colour grading differences between original and generated areas. The quality of outpainting varies across models, with stronger models producing more coherent and seamless extensions.
What is iterative outpainting?
Iterative outpainting involves extending the canvas repeatedly, using each expansion as the reference for the next. Each generation step extends the image further, allowing the construction of very wide environments or panoramas far larger than any single generation could produce. Each step is anchored to the previously expanded edge, so the content remains spatially consistent across the full extent of the constructed image.
Can outpainting add content in any direction?
Yes. Outpainting can extend an image in any direction: left, right, up, down, or simultaneously in multiple directions. Directional extension allows targeted additions (adding sky above, widening the environment horizontally). Multi-directional extension grows the canvas outward on all sides simultaneously. The direction chosen should be informed by what the compositional goal requires: changing aspect ratio typically requires extension in one or two specific directions.
Is outpainting useful for video production?
Yes, in several ways. Outpainting can adapt still reference images or concept art to the correct aspect ratio for a video project. Extended backgrounds created through outpainting can serve as environment plates or compositing backgrounds. In AI video workflows, outpainting can expand key frame images to wider compositions that provide more spatial context for subsequent video generation. It is also used to create widescreen versions of portrait-format generated images for use in landscape video contexts.
Can't find what you are looking for?
Contact us and let us know.
bg