Captions That Match Your Script: Correct from Script

Automatic transcription is a strong starting point for captions, but it is not the finished article. A narrator says a place name, a product name, or a technical term, and the transcript hears something close but wrong. Viewers notice those mistakes immediately, because captions are read, not just heard. If you wrote a narration script before recording, you already hold the most reliable reference for what should be on screen. Correct from Script is the step that uses it.

Why captions drift from the script

Transcription in VLPremiere Pro runs on your Mac and produces a transcript with word-level timing. That timing is valuable: it is what turns a block of text into captions that appear at the right moment. The words themselves, however, come from listening. Names with unusual spellings, foreign words, numbers, and specialist vocabulary are exactly the places where a listening system can guess. Correcting those errors one caption at a time is slow on a long documentary, and it is easy to miss one instance of a name that appears twenty times.

What Correct from Script does

Correct from Script compares the captions against your original script and fixes misheard words using that script as the reference. The guide describes it simply: it fixes misheard words using your script. The important point is what does not change. The caption timings were built from the actual voice, with word-level timing from the transcription, so the correction works on the wording while the timing, caption breaks, and styling you have chosen stay in place. You are not re-transcribing or starting again; you are aligning the text to the words you wrote.

A practical workflow, step by step

First, transcribe the sequence. In the Text panel, open Transcript and transcribe the voiceover on your Mac, then press Captions from Transcript and choose a caption style. VLPremiere Pro includes animated caption styles with word-by-word highlighting, and you can save your own looks as presets.

Second, run Correct from Script if you have the narration script, as the guide recommends. Use the script that was actually recorded, not an early draft, so any lines the narrator changed in the booth do not get reintroduced as errors.

Third, review the result the way a viewer will: read the captions while the video plays. Check names and terms first, then numbers and dates. The Captions tab also gives you spell check for the caption language, and Find and Replace for a term that needs one consistent form throughout the film.

Finally, set the reading rules before export: characters per line, minimum caption duration, and gaps. Those rules keep corrected captions readable as well as accurate.

When you do not have a script

Not every project starts from a written script. Interviews, observational footage, and archive material often arrive unscripted. In that case, the transcript itself is your reference: correct names and terms directly in the Captions tab, then use Find and Replace so one correction carries through the whole sequence. Spell check per caption language helps here too, especially on long sequences where a single misspelling repeated throughout would otherwise require many small edits.

Corrected once, used everywhere

Accurate captions are not only for the main film. Captions can be translated to create a caption track in another language, using on-device translation or an AI service, and that translation will inherit whatever errors were left in the source. Correcting from the script first therefore protects every language version that follows. The same captions can be burned into the picture, saved as a separate file, or embedded in the video file, and caption files can be exchanged as SRT and VTT. Right-to-left captions are supported for Arabic and Hebrew. One careful correction pass at the start saves repeated fixes in every export and every language.

What to check before you export

Before delivery, do a final read-through of proper names, technical terms, and any word your audience will search for or quote. Confirm the caption style still reads clearly at viewing size, that line lengths and minimum durations leave time to read, and that the exported form matches the destination: burned in for some platforms, a separate SRT or VTT file for others, or embedded where the player expects it. The feature guide covers the Captions and Text section in full, and the support page answers common questions about formats and workflow.

The short version

Transcription gives you timed captions quickly. Your script gives you the correct words. Correct from Script brings the two together, so viewers read exactly what the narration says, without a manual hunt through every caption in a long film. VLPremiere Pro will be available to purchase soon at vlpremier.com, and this guide series will continue with the next workflow: turning a script into narration with Voiceover Studio.