How many hours vanish when captions arrive after approval?
The edit is approved, the client is happy, and the producer is already thinking about delivery. Then one message lands: "Can we just add subtitles too?" The word "just" does a lot of damage. Now the team needs a transcript, terminology checks, timing, line breaks, exported SRT or VTT files, a burned-in version for social, maybe a translated version, and another approval pass.
Captions are not hard because the files are mysterious. They become hard because studios often start them after the video feels finished. At that point every typo means another export, every term question blocks delivery, and every new language becomes a small project inside the project.
In 2026, captions and subtitles are no longer optional polish. They support silent viewing, accessibility, international launches, product demos, recruiting videos, training content, event screens, and social edits. For a production studio, they are not a small postscript. They are a deliverable layer that needs ownership.
Why captions are now a production workflow
Teams used to treat subtitles as a text track added at the end. That breaks down when a single video has to work on a website, in a sales deck, in social feeds, on a booth screen, and in a second market. The client may not know the difference between SRT and WebVTT, but they expect the studio to deliver files that play correctly in each context.
Automatic transcription has also changed expectations. It makes the first draft faster, which is useful. It does not remove producer responsibility. Auto captions still miss names, brands, acronyms, accents, numbers, product terms, and short phrases. One missing word can reverse a sentence. One wrong term can make a serious video look careless.
Captions also affect layout. They can cover lower thirds, disclaimers, product UI, hand gestures, or on-screen text. If the team discovers this after final approval, the editor has to rescue the frame by moving graphics, shortening lines, changing text size, and exporting again.
What usually breaks
- Captions are not in the estimate - the team treats transcription, timing, editing, export, and QA as free cleanup.
- Text approval is mixed with picture approval - the client approves the video, then rewrites copy inside the final file.
- No one owns terminology - names, job titles, product terms, and legal lines change between versions.
- Translation arrives too late - the second language is longer than the original and no longer fits the timing or frame.
- Caption files drift away from video files - the SRT, VTT, and burned-in export no longer match.
- QA checks the text, not the viewing experience - the file is accurate in a document but unreadable on a phone.
These failures look small when a studio delivers 1 video in 1 language. They become a real post-production cost when the project has 5 videos, 3 formats, and 2 languages.
How to build a caption workflow
1. Scope caption deliverables before editing
Captions should be part of the brief or kickoff, not a surprise on delivery day. Ask what the client actually needs: closed captions, burned-in subtitles, SRT files, VTT files, translated subtitles, a transcript, clean exports without text, or larger captions for vertical versions.
A good estimate does not say "subtitles" as one vague line. It names the files. For example: clean master, master with open captions, Russian SRT, English VTT, vertical version with adjusted caption size, and transcript.
What to do: add a "caption and localization deliverables" line to your brief and estimate template. Even when the client declines it, the decision is captured before work starts.
2. Create a rough transcript during rough cut
Do not wait for final sound. Once the rough cut has a stable structure, create a draft transcript. It will reveal long lines, unclear names, awkward phrasing, legal statements, product terms, and places where the final design needs space for text.
The transcript does not need to be perfect yet. Its job is to expose scope and give the client a text layer for early feedback. If the client wants to change wording, it is better to know before color, sound mix, and final graphics.
What to do: after the first working cut, create a transcript task with 3 statuses: draft, client review, locked.
3. Separate picture approval from text approval
A video can be visually approved while the caption text is still under review. If those states are mixed, everyone reads the project differently. The editor thinks the video is finished. The client thinks text tweaks are harmless. The producer cannot explain why delivery is slipping.
Separate statuses make the work visible. A video can have picture approval, sound approval, and caption approval. For a small studio, this can be a simple checklist inside the project. For a larger team, it should have owners and deadlines.
A PMS like Basalt gives producers one place to keep caption status, files, and revision notes connected to the actual video project.
What to do: set one rule: no final caption export starts until the transcript is locked.
4. Assign one owner for terminology
Every project has words that should not be left to automatic transcription. Product names, speaker names, roles, industry terms, legal phrases, city names, discount codes, figures, and campaign tags need a small glossary before final editing.
The terminology owner does not have to be a linguist. It can be the producer, account manager, or someone from the client team. The point is to create one source of truth. If one file says "VP People" and another says "Head of HR," that is not just style. It is version drift.
What to do: create a 3-column glossary: term, approved spelling, notes. Attach it to the caption task.
5. QA captions inside the final viewing context
Accurate text can still be bad captions. A line may be too long. White text may disappear on a bright background. Two lines may cover lower thirds. A translation may read well in a document but fail inside the timing.
Caption QA needs 2 passes. The editorial pass checks meaning, names, numbers, punctuation, and terminology. The visual pass checks readability, line breaks, contrast, position, and whether the captions fight the frame.
What to do: review the final export on a phone and a laptop. If the video is going to an event screen, test one export at a similar viewing distance.
6. Price captions as real work
A caption workflow uses time from several people: editor, producer, translator, copy reviewer, client stakeholder. When that time is hidden inside "final tweaks," the studio gives away margin and trains clients to add versions late.
Separate basic and extended caption work. A basic package might include 1 language, 1 caption file, and 1 burned-in export. Extended work can include translation, platform-specific formatting, vertical caption adaptation, transcript delivery, urgent QA, and extra text revision rounds.
What to do: define how many caption text revision rounds are included. Anything beyond that should be scoped as additional work.
A minimum process for this week
If you do not have time to rebuild the whole workflow, start with one rule: every project with captions gets a caption checklist before final edit. Keep it next to the post-production tasks, not in a separate thread.
Minimum version:
- List the required languages and caption formats.
- Create a draft transcript after rough cut.
- Build a glossary for names, brands, terms, and numbers.
- Lock the transcript before final caption export.
- QA the final video on the target devices.
- Use clear filenames for every caption and subtitle version.
That is enough to prevent most late exports, missing translation files, and "which SRT is final?" moments.
Checklist: are captions under control?
- Captions and translations are listed in the estimate as deliverables.
- Each video has a clear need for SRT, VTT, burned-in, and clean versions.
- A draft transcript exists before final export.
- The client has a separate text approval round.
- Names, brands, terms, and numbers are captured in a glossary.
- Translation has been checked for line length and timing.
- The final version has been reviewed on target devices.
- Caption files are named clearly and tied to the correct video version.
Frequently Asked Questions
Can we use automatic captions?
Yes, as a draft. They save time on the first pass, but they do not replace editing, terminology review, and visual QA. For client work, unreviewed automatic captions are too risky.
When should a translator join the project?
After rough cut, before final export. The translator needs context, and the editor needs enough time to see where translated lines are longer than the original.
Should caption files stay in the project archive?
Yes. Captions often come back when a website changes, a new platform launches, another language is added, or the client reuses the video. Store them next to final exports and record which video version they belong to.
Summary
Captions are no longer a tiny task at the end of post-production. They affect accessibility, localization, brand quality, timelines, and studio margin. The earlier a team treats them as a workflow, the fewer emergency exports appear before delivery.
Start with one step: add a caption checklist to the next project before editing begins. It will show the client that your studio manages not just the image, but the full path from final cut to viewer.