Video essay subtitles with clear quotation and source checks.

Make the argument readable without changing who said it

A video essay can move from your narration to an excerpt, then to your interpretation of that excerpt. Polyglot can create timed subtitles from the finished recording, translate reviewed text locally on an Apple Silicon Mac, and export SRT or WebVTT. The editorial task is to keep the boundaries clear: which words belong to the source, which are your paraphrase, and where a viewer can find the original.

This workflow is for essays you own, using recordings and other material you are authorized to process and publish. It adds a short manual excerpt record beside the subtitle project. Polyglot does not identify source works, check quotations against them, generate citations or decide whether reuse is permitted.

Resolve permission and attribution before the subtitle pass

Confirm that the planned use covers the footage, spoken excerpts, music, still images and any translated version. Keep the relevant permission or license record with the production notes. A credit line is not a substitute for permission, and access to a public video does not by itself settle reuse. Hold material with unresolved rights instead of assuming that an educational or critical purpose clears it.

For Creative Commons material, its attribution guide recommends recording the title, credited creator, original source and specific license, with relevant changes identified. Follow the applicable terms and supplied credit information. This page does not determine whether a particular license permits your edit or translation. Store the full reference details in a readable source list or other suitable credits location; do not try to fit the whole bibliography into fast-moving dialogue captions.

Record two times for every borrowed passage

Lock the essay cut before generating the release subtitles. In a separate document, give each excerpt a source key, its original source locator, its position in the essay, the kind of use and the matching credit entry. An original source time helps a reviewer find the passage; the essay time helps them check its new context. The exported subtitle must follow the essay timeline, not the source timeline.

A paraphrase deserves a source entry too, but should not acquire quotation marks that make it look verbatim. When an excerpt ends, check the transition into your commentary. A voice label or other necessary caption context may help where the speaker is otherwise unclear; Polyglot does not identify speakers automatically.

The following newly written fictional example illustrates the review record. All sentences and filenames are invented. It is a manual production note, not an app feature or a tested transcription result.

Fictional excerpt review
Essay: colour-v05.mp4

Source A / own demo-v02
Original: 00:42–00:46
Essay:    01:18–01:22
Use: direct quotation
Heard: “Blue is not colder.”
Draft: “Blue is colder.”
Action: restore “not”.
Credit key: A

Essay: 01:23–01:27
Use: creator paraphrase
Heard: “The light changes
how we read the colour.”
No quotation marks.
Reference key: A

Picture says: demo-v01
HOLD: correct source card
before locking the cut.

Check the audio, quotation and source card separately

Listen to the excerpt in context, then compare the draft with the actual words. Check names, omitted negatives, qualifications and where quoted speech starts and ends. W3C warns that an automatic caption can reverse meaning by missing a single word. Correct a recognition mistake, but do not quietly rewrite a speaker’s inaccurate statement to match your preferred argument.

Next compare the on-screen source card and the separate reference entry with the material actually used. A correct subtitle does not repair a wrong edition, version or source link in the picture. If the narration misquotes a text, resolve that in the essay itself before the final subtitle pass. If the video edit changes, revisit the excerpt positions and timing instead of assuming the old track still fits.

Keep the cue text focused on audible content and necessary caption context. Preserve spoken attribution such as “in the creator’s words”. Avoid inserting an unspoken URL into the speaker’s sentence. If a source card carries essential information that is never voiced, plan suitable description or a clearly labeled companion explanation. W3C treats visual description separately from captions; speech recognition cannot read the card for you.

Translate the quotation without borrowing authority

Have a reviewer competent in the target language compare the translated passage with the reviewed source and its surrounding commentary. Preserve negation, uncertainty and the shift from quotation to paraphrase. Check the spelling of titles and creator names against the approved reference list. A published translation of a source may use different wording; do not silently present a new machine translation as that edition’s exact quotation.

Polyglot preserves source timing and checks translated cue structure, but structural checks do not establish faithful meaning or correct attribution. If the target text is too dense, refine it with the reviewer or finish individual timing and positioning in suitable tools. Keep a separately reviewed file for each delivery language, and recheck affected translations after source edits.

Accept the subtitle track and source list as one delivery

Export the format requested by your destination and attach it to the exact final essay. YouTube documents language-specific subtitle uploads with timing; confirm the requirements for your own player. Polyglot produces separate subtitle files, not burned-in credits, a finished cinema package or an automatically published video.

Check the opening, every quoted-to-narrated transition, a dense citation scene and the ending. On a small screen, make sure captions do not obscure the source card or visual comparison. If repositioning is needed, use the destination’s supported caption tools; Polyglot’s WebVTT import does not retain styling or positioning. Include meaningful non-speech audio in an accessibility caption review rather than treating speech transcription as a complete track.

Open each public source link and confirm that it leads to the work and version named in the essay. Keep credit keys consistent across the picture, reference list and any companion transcript. Deliver the final video identifier, approved language files and public references together; keep private permission records out of the public bundle. Release when both the caption playback and source-reference check pass.

Keep exploring.