Consistent visual variants

How can you adapt a visual into square, portrait and landscape formats without distorting its subject or losing its message?

1250 € per session, excl. VAT

In short

In a half-day session of 3 hours you create three versions of the same visual for digital media. You choose between cropping, recomposition and AI-assisted extension according to the options available, then check fidelity to the subject, legibility, dimensions and conditions of use before exporting.

Why this training?

A change in aspect ratio can crop off a handle, reduce the space reserved for text or shift the focal point. Extending the edges with AI can also introduce unwanted objects or alter the appearance of the product shown. Distinguishing elements that must remain untouched from areas that can be modified helps you choose a restrained adaptation. Comparison with the original and checks at display size make corrections more precise than a purely aesthetic assessment. Running case study: The fictitious organisation « Avelune Céramique » is preparing to present a handcrafted cup. Using a visual authorised for use, it wants a square version of 1080 × 1080 pixels, a portrait version of 1080 × 1920 and a landscape version of 1920 × 1080, without altering the cup’s shape or decoration. Transfer to the workplace: For the next request for visual variants, reuse the format matrix and checklist on an image authorised for use. Compare the versions before and after correction, then record the defects avoided and the limitations encountered.

Objectives of this training

  • Identify constraints relating to aspect ratio, dimensions and text placement.
  • Choose between cropping, recomposition and background extension.
  • Formulate an adaptation instruction specifying the elements to preserve.
  • Produce three consistent exports from one source image.
  • Assess visual discrepancies and document corrections.

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Communications officers producing visuals for multiple digital media.
  • Content creators and marketing assistants who regularly produce visual variants.

Prior skills

  • Have already formulated a few simple requests to a generative AI.
  • Be able to open an image, identify its dimensions and save a copy.

What you will work on

  1. Assess the visual and its uses

    Objective: Define what can change and what must remain intact. Aspect ratios, pixels, usable resolution, safe areas; reproduction and adaptation rights, consent from people depicted, required notices according to use.

  2. Develop the framing

    Objective: Choose a workable composition for each aspect ratio. Difference between resizing and cropping; focal point, margins, negative space and risks of distortion.

  3. Adapt the peripheral areas

    Objective: Create background variants without transforming the product. Instructions specifying editable areas, lighting, perspective and exclusions; extension only if available; joins, textures and shadows to inspect.

  4. Check and export

    Objective: Deliver usable files with verified quality. Checks at 100 % and at display size; legibility, compression, choice of JPEG or PNG, file naming; transparency about modifications and prevention of misleading deepfakes.

What you take away

  • Matrix of the three formats and protected areas.
  • Adaptation instructions and correction history.
  • Three files exported at the requested dimensions.
  • Comparison sheet and completed checklist.

Method and assessment

  • Short teaching inputs followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with peer review of the outputs.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Arrange access to a tool authorised by your organisation and check which editing functions are actually available. Prepare a fictitious or anonymised visual with established adaptation rights, accompanied by the graphic constraints.

Observable criteria: The three exports exactly match the dimensions in the brief. ; The cup’s shape, decoration and proportions remain faithful to the original. ; The subject and text remain visible within the areas defined for each medium. ; No inconsistent joins or unwanted objects are detected during the check at 100 %. ; Rights, any required consent and necessary notices are documented..

Frequently asked questions

Does a new image need to be generated for each format?

No. Cropping is sometimes sufficient; generation remains limited to areas where modification is useful.

Can you use a photograph featuring a recognisable person?

Only with the necessary rights and consent. The exercises favour a product on its own and exclude misleading deepfakes.

Should AI recreate the text and logo?

No. Keeping the approved elements and repositioning them separately avoids incorrect characters and distorted brands.

What should you do if the source image is too small?

Find a higher-resolution source or adjust the intended use; enlargement does not restore missing details.

Last updated: