From shot breakdown to editing

How can you achieve narrative and visual continuity between shots produced using generative AI?

1960 € per session, excl. VAT

In short

This 7-hour day, excluding breaks, turns a video brief into an edited sequence that can be checked and handed over. You define the intentions for each shot, establish consistent visual references, prepare generation requests, select takes and check continuity between shots. The final edit is accompanied by a quality checklist and a traceability dossier.

Why this training?

Shots that are appealing in isolation do not necessarily make a scene easy to follow: an object changing appearance, an interrupted action or a reversed direction can disrupt the viewer’s understanding. Generative AI allows you to explore proposed shot breakdowns and produce variations according to the capabilities available. The method prioritises a consistent reference, targeted tests and a reasoned selection rather than an accumulation of generated outputs. Running case study: The fictitious organisation Avelune Céramique is preparing a 30-second video showing a bowl being packed for dispatch. Six to eight shots must make the action understandable, preserve the product’s appearance and end with a sealed parcel, without any identifiable person. Transfer to the workplace: Reuse the shot breakdown and continuity checklist for an upcoming work-related video, then have the edit reviewed by a colleague who was not involved. Record misunderstandings, corrections and publication decisions.

Objectives of this training

  • Breaking an action down into six to eight justified shots
  • Formalising visual constants and continuity constraints
  • Writing generation requests that are consistent across shots
  • Assembling a sequence with a clear temporal progression
  • Evaluating the edit and documenting priority corrections

Who is it for and what are the prerequisites?

Roles and tasks concerned

  • Audiovisual communications officers producing short videos
  • Editors and content creators incorporating generated elements into their sequences

Prior skills

  • Be able to import, cut and export media in an editing environment.
  • Have experience of making structured requests to a generative AI.
  • Know shot sizes and the basic principles of continuity.

What you will work on

  1. Defining the narrative and intended uses

    Objective: Define the action to be understood and the production limits. Audience, channel, duration; rights to reference material, consent, synthetic content disclosures and deepfake risks.

  2. Breaking down the action

    Objective: Assign a narrative function to each shot. Wide shot, detail shot, insert; 180-degree rule, match on action and ellipsis.

  3. Establishing consistent references

    Objective: Preserve recognisable characteristics across takes. Bowl shape, colour, lighting, setting, orientation and packaging status.

  4. Producing and selecting takes

    Objective: Choose media compatible with the planned continuity between shots. Limited tests, comparison of variations, distortions and inconsistent movements; limitations of the available tool.

  5. Assembling for continuity

    Objective: Make the progression of the action immediately clear. Cut points, pacing, sound continuity and visual harmonisation.

  6. Checking and delivering

    Objective: Check quality and prepare the handover. Viewing without commentary, counting continuity breaks, technical checks and traceability.

What you take away

  • Technical shot breakdown and annotated storyboard
  • Visual bible and generation request log
  • Exported 30-second sequence and editing project
  • Completed quality checklist and media provenance dossier

Method and assessment

  • Short instructional inputs followed by immediate practice on the running case study.
  • Individual work followed by work in pairs, with peer review of the productions.
  • Guided group discussion: what works, what needs correcting and what remains a human decision.
  • Preparation: Check access to a tool authorised by your organisation and the generation capabilities actually available. Prepare fictitious or anonymised reference material that can legally be used; provide backup training rushes.

Observable criteria: The six to eight shots allow the packing sequence to be recounted without an oral explanation. ; Every continuity match is checked: direction, position, action and state of the bowl. ; Every identified anomaly receives a correction or an explicit justification. ; The export meets the 30-second duration, framing and parameters set in the brief. ; Rights, necessary consents and applicable disclosures are documented before distribution..

Frequently asked questions

Do all shots need to be generated?

No. Authorised rushes can supplement generated footage as long as continuity between shots remains consistent.

What should you do if generation fails?

Use the backup training rushes and document the limitations encountered.

Can you reproduce a real face?

The exercise rules this out. Professional use requires, among other things, rights, appropriate consent and no misleading presentation.

How should synthetic elements be disclosed?

Document their provenance and determine the disclosures required according to the distribution context and applicable rules.

Last updated: