Automated subtitles and captions

Why every video needs subtitles

Subtitles stopped being optional the moment social feeds started autoplaying video on mute.

Viewers scrolling in public, at work or in bed read your video before they ever hear it. No captions usually means no watch time.

The audience math goes further than muted playback. Captions serve deaf and hard-of-hearing viewers, non-native speakers who follow written text more easily than fast speech, and anyone watching in a noisy environment. Search engines and platform algorithms cannot listen to your audio either, but they can read a transcript and a subtitle file.

The only real argument against subtitling everything has been the labor. Typing out dialogue and timing it by hand takes several times the runtime of the video itself. Automated subtitles remove that cost: the transcription and timing happen in minutes, and your team only reviews and exports.

Which subtitle format to use

Subtitle files carry the same three things, the text, when it appears and when it disappears; the format decides where you can use them.

SRT (SubRip) is the plain-text standard almost everything accepts. If you only export one file, make it an SRT. New to the format? Start with our guide to SRT files.

VTT (WebVTT) is the web-native format used by HTML5 video players. It behaves like SRT with extra support for text styling and positioning cues, a WEBVTT header line, and periods instead of commas in its timestamps. Converting between SRT and VTT keeps the text and timing intact in both directions, though VTT's styling and positioning cues are dropped on the way to SRT. Sonix exports both from the same transcript, so "SRT vs VTT" is rarely a hard choice: use VTT for web players, SRT everywhere else. Already holding the wrong file type? Follow our VTT to SRT conversion guide.

TTML is an XML-based format common in broadcast and OTT streaming delivery specs. Sonix exports it alongside SRT and VTT.

SUP is different: it stores subtitles as images rather than text, and it is the subtitle format found on Blu-ray discs. Because there is no text inside a SUP file, platforms like YouTube cannot read it and the wording cannot be edited directly.

Burned-in (hardcoded) captions are not a file at all. The text is rendered permanently into the video's pixels, so it plays everywhere and viewers cannot turn it off. This is the workhorse for social feeds, where videos autoplay muted. Sonix can burn subtitles into video for you, with your choice of font, size, color and position.

Converting SUP to SRT

SUP files contain images, not text, so getting them onto the web means running OCR or regenerating them from the audio track.

If you have Blu-ray subtitles in SUP format and need them on the web, they must be converted to a text format first. Since SUP files contain images, the conversion requires OCR to read the text out of each subtitle frame, which is error-prone with stylized fonts.

In many cases the faster, cleaner route is to skip conversion entirely: run the video through Sonix and generate a fresh, accurately timed SRT from the audio itself. Our step-by-step SUP to SRT conversion guide walks through both options.

Captions, subtitles and SDH: what is the difference?

The three terms get used interchangeably, but they solve different problems.

Subtitles

Subtitles are a translation or transcription of the dialogue, written for viewers who can hear the audio but do not understand the language or want reading support. They carry speech only.

Captions

Captions are written for viewers who cannot hear the audio. Closed captions (CC) can be toggled on and off by the viewer; open captions are burned into the picture and always visible. In US usage, "captions" implies accessibility: they include speaker changes and meaningful sounds, not just dialogue.

SDH

SDH (Subtitles for the Deaf and Hard of Hearing) combines both. SDH looks like subtitles but adds the non-speech information captions carry: speaker identification, sound effects and music cues. SDH is the standard deliverable for streaming platforms and the safest choice when accessibility is a requirement.

When is each required?

If your audience simply watches with sound off, standard subtitles or burned-in captions are enough. If you are meeting accessibility obligations, publishing to streaming services or serving deaf and hard-of-hearing viewers, you need captions or SDH. Sonix labels speakers automatically and gives you the editor to add sound effects and music cues before export; see our SDH subtitle creation page for the full workflow.

How do you add automated subtitles on each platform?

The generation step is identical everywhere, upload to Sonix, review, export; what changes is the file you export and where it goes.

YouTube

Export your subtitles as SRT or VTT, open YouTube Studio, and upload the file under Subtitles for your video. Sonix's timing is precise enough for direct upload, and unlike YouTube's auto-captions, you have already corrected names, terminology and punctuation before the video goes public. Full walkthrough: how to add subtitles to a YouTube video.

Adobe Premiere Pro

Export an SRT from Sonix and import it into your Premiere Pro project, where it lands on the captions track ready for styling, or use Sonix's XML export to bring the transcript directly into your edit. From there you can restyle captions or burn them into the export. Step by step: how to add subtitles in Premiere Pro.

Final Cut Pro

Sonix exports FCPXML, Final Cut Pro's native interchange format, so your subtitles arrive as proper caption roles with timing intact rather than as pasted text. Import the file, assign the caption role, and adjust appearance inside Final Cut Pro. Guide: how to add captions in Final Cut Pro. Editing in iMovie instead? Use our iMovie captions and subtitles guide.

TikTok, Reels and Shorts

Vertical video changes the rules: most viewers watch muted, sidecar subtitle files are unreliable across social apps, and the safe zone is narrow. Burn the captions in. Generate subtitles in Sonix, pick the 9:16 story canvas, style them with a large font and a high-contrast color, keep them centered above the caption and UI overlays, and download the finished video. One upload works identically on TikTok, Instagram Reels and YouTube Shorts. See also: adding captions to Reels.

Web players, course platforms and Vimeo

For video embedded on your own site, an LMS or Vimeo, export VTT, the format HTML5 players read natively. Upload the VTT alongside the video and the player handles the caption toggle, which keeps the text selectable, searchable and accessible to screen readers in a way burned-in captions are not. Most course platforms and enterprise video hosts accept SRT as well, so exporting both formats from the same Sonix transcript covers every embed.

What do WCAG and the ADA require for captions?

Public-facing video usually has to be captioned.

The Web Content Accessibility Guidelines (WCAG) are the reference standard for digital accessibility. WCAG 2.1 Success Criterion 1.2.2 (Level A, the minimum conformance level) requires captions for all prerecorded audio content in synchronized media. Success Criterion 1.2.4 (Level AA, the level most organizations target and most regulations reference) extends the requirement to live audio content.

In the United States, the Americans with Disabilities Act (ADA) has repeatedly been applied to digital content, and organizations have faced legal complaints over uncaptioned video. Government agencies, universities and broadcasters face additional obligations under Section 508 and FCC captioning rules. This page is not legal advice, but the practical takeaway is consistent: accurate captions on published video are the baseline expectation, not a nice-to-have.

Compliance-grade captions must be accurate, synchronized and complete, including speaker identification and relevant sound information, which is exactly what SDH provides. Automated generation plus a human review pass in the Sonix editor is the fastest way to get there. For the broader picture of accessible media, read our guide to making web content accessible.

Subtitles in other languages

Yes, and this is where automation pays off twice: Sonix transcribes in 54+ languages and translates finished transcripts into 54+.

Generate native subtitles for content recorded in Hindi, Spanish, Japanese, Arabic and dozens more, not just English. The transcript comes back in the language's own script, timed word by word, ready to split into subtitle blocks of any length.

To reach new audiences, translate your subtitles instead of re-recording. Sonix's automated translation converts finished transcripts into 54+ languages with a side-by-side view for review, and the subtitle timing carries over, so a translated SRT is minutes away from the original.

We keep dedicated workflow guides for the languages creators ask about most, including Bengali subtitles, Urdu subtitles, Tamil subtitles and Hindi subtitles. Each covers the same core pipeline with language-specific notes on scripts and fonts.

One practical tip for multilingual output: keep one corrected master transcript, then translate from it. Fix an error once and every language version inherits the fix, instead of correcting the same mistake six times.