Localize a video

What this is about

Translated text rarely keeps its length. German expands against English, and a lower third that fitted perfectly now runs past the safe area or forces a type size nobody can read on a phone. On top of that sit the things translation cannot fix: a claim that is permitted in one market and regulated in another, a gesture that reads differently, a product name that is not used in that country, a person who is not the local contact. Localization is therefore an editorial decision made at treatment stage, not a service ordered after delivery.

The order matters more than the tools. Skip a step and you usually notice two steps later — by which point the fix costs several times as much.

What runs differently here

Where this differs from the general case:

• Design text areas for the longest language you will ever need, not for the one you are editing in
• Keep on-screen text in editable layers or as separate graphic files; text baked into a shot is a reshoot
• Decide subtitle, voiceover or full dub per market and per purpose — internal training and paid campaigns rarely need the same
• Have claims, figures and legal lines checked locally; translation is not compliance
• A music and effects mix without dialogue is what makes every later language cheap

Step by step

This order gets you there fastest:

1. List the target markets at treatment stage and mark which ones need full localization, which need subtitles only, and which get the original. That list changes how the film is written.
2. Write the master script with translation in mind: avoid puns, idioms and sentences that depend on word order, and keep the strongest claims in the voice rather than in a graphic.
3. Design every text element with expansion room and keep the project files editable. Titles rendered flat into the picture turn a two-hour language swap into a graphics rebuild.
4. Shoot with localization in mind: keep visible signage neutral where possible, and cover moments where a spoken product name would tie the shot to one market.
5. Order the mix with separated stems, including a version without dialogue. Every additional language then reuses the music and effects instead of paying for a new mix.
6. Brief translators with context, not just text: the audience, the tone, a link to the picture, and a glossary of product and role names with their approved local spelling.
7. Have a native speaker from the market review the assembled version, not just the script. Timing, emphasis and on-screen text problems only appear once the language sits in the film.
8. Version and name the files by market and language from the first export, so nobody publishes the wrong cut in the right country three months from now.

Common pitfalls

What most often goes wrong in practice:

• Discovering after picture lock that every title was rendered into the shot
• Ordering a translation of the script and treating the result as a finished localization
• Assuming a claim approved in the home market is approved everywhere it will be seen
• Paying for a full new mix per language because no dialogue-free stem was ever requested

With TillyGen

TillyGen holds the full deliverables list from the start, so language versions appear in the plan and the budget as their own items — the moment they are needed, the edit is already built to accept them.

Change one constraint and the consequences travel through the whole plan: affected shots are flagged, the call sheet is regenerated, and nobody keeps working from yesterday's version.