Burned-In vs Soft Captions for Podcast Clips
For podcast clips you post and re-post across feeds, burn the captions in. Burned-in (open, "hardcoded") captions are baked into the video pixels, so they survive a download, a re-share, and a re-upload to the next platform; soft (closed) captions live on a separate text track that drops off the moment the file leaves the app it was made in (3Play Media). For a clip whose whole job is to travel, the version that travels with it wins.
There's one real exception, and it's the part most "open vs closed" explainers skip: on YouTube, you want both, burned-in text for the mute majority and an uploaded soft track the search index can read. Below is what each type actually is, a platform-by-platform table of which one survives the trip, the single scenario that needs both, and the mistakes that quietly kill captions in transit.
What's the difference between burned-in and soft captions?
Burned-in captions are part of the picture; soft captions are a file attached to the picture. Burned-in (also called open or hardcoded) text is rendered into the video frames during export, so it's always on and can't be toggled. Soft (closed) captions sit in a separate sidecar track, an SRT or VTT, or a platform's own auto-caption layer, that a player loads at playback and a viewer can switch off (Rev).
That single structural difference decides everything downstream. Because burned-in text is the image, it goes wherever the pixels go, every device, every re-share, every repost, but no algorithm can read it, since to a machine it's just shapes (Subly). Soft captions are the opposite: a search engine can index the words and a viewer can resize them, but the track is fragile in transit and disappears when someone downloads and re-uploads the file without re-attaching it.
Why this matters more for clips than for the full episode
Your long episode usually lives in one place. A clip's entire purpose is to get downloaded, cross-posted, duetted, stitched, and saved, and most of those moves strip a soft caption track. The same clip often gets uploaded across several platforms to chase a viral hit, and every one of those re-uploads is another chance for a separate caption track to fall off along the way.
The other half of the stakes is that captions aren't optional in the first place. Most social video is watched on mute, a directional figure put as high as 85% by Digiday back in 2016, with later studies landing roughly between 69% and 85%, so treat it as a range, not a law. Either way, a clip that loses its captions during a repost loses the silent majority on that platform. Burned-in text is the only version that can't get lost.
The platform-by-platform survival table
Here's the part the generic explainers don't give you: whether each caption type actually survives on the platforms a podcast clip travels through. "Survives re-upload" means the captions still show after you download the clip and post it elsewhere. "Survives compression" asks whether the platform's re-encode degrades the text.
| Platform / move | Burned-in (open) | Soft (closed) track |
|---|---|---|
| TikTok → download → repost | Survives (in pixels) | Auto-caption renders in-app only; gone in the saved MP4 |
| Instagram Reels (no SRT upload supported) | Survives | No reliable native track to begin with |
| YouTube Shorts / long-form | Survives | Survives only if you upload/keep an SRT; indexable for search |
| Re-share / duet / stitch by another account | Survives | Almost always lost |
| Platform re-compression | Survives, but soft at low-res export | Unaffected, it's text, not pixels |
| Archiving / repurposing months later | Survives in the file | Survives only if you saved the SRT separately |
Sources: TikTok's auto-captions render inside the app and aren't baked into a downloaded file, so they vanish on repost (BibiGPT, "Extract TikTok captions," 2026); a separate caption file does not auto-travel with the video on re-upload (3Play Media); Instagram Reels and TikTok don't reliably support uploading your own caption file, so burned-in is the practical default there (University of Maryland accessibility guidance). The one honest caveat on burned-in: because the text is part of the image, an aggressive re-encode can blur it, so export at full resolution and high bitrate, or your captions pixelate where soft text never would (Subly).
The pattern across that whole table: for anything that gets downloaded, re-shared, or reposted, burned-in is the only caption type you can count on. Soft tracks are reliable in exactly one direction, staying put on the original platform where you uploaded them.
The one scenario where you need both
YouTube is the exception that earns a second caption type. Burn in your styled captions for the mute-watching majority, and also keep an accurate soft track (an uploaded SRT), because YouTube reads caption text as searchable content, so a clean transcript helps your clip surface for spoken keywords that a pixels-only clip can't rank for (3Play Media, "7 ways captions improve SEO," 2026). Burned-in text alone is invisible to that index; the soft track is what the algorithm can actually parse.
Two honest caveats here. First, most caption-SEO sources are vendors who sell captioning, and exact ranking percentages in that material are marketing-flavored, the defensible claim is the mechanism (search engines read text, not pixels), not a specific lift number. Second, accessibility compliance is the other reason to keep a soft track where you can: for formal ADA/WCAG/FCC contexts, a separate, toggleable caption track is the safe standard, and burned-in text doesn't give users the ability to resize, restyle, or turn captions off (University of Maryland). On platforms that won't accept an SRT, Reels, TikTok, open captions are the accepted fallback. On YouTube and your own site, do both.
Common mistakes (and the fix)
- Trusting a platform's auto-captions as your visible captions. TikTok's native track renders only inside the app and disappears from a downloaded file, so a reposted clip arrives blank (BibiGPT). Fix: burn in your own reviewed captions before export; treat any native track as a supplement.
- Exporting burned-in captions at low resolution. The text is part of the image, so an underpowered export plus the platform's re-encode leaves it blurry (Subly). Fix: export at full 1080×1920 and a high bitrate.
- Skipping the SRT on YouTube. Burning in captions and uploading no transcript throws away the one platform that indexes caption text (3Play Media). Fix: burn in and upload a corrected SRT on Shorts and long-form.
- Posting auto-generated captions without reading them. Speech-to-text mishears names, jargon, and homophones, and once burned in you can't fix a word without re-exporting. Fix: review the transcript first, the clip-detection step and the captioning step are separate jobs, and a sloppy caption shows on every repost. The full setup lives in how to add captions to podcast clips.
- Assuming "soft" means "safe." A soft track feels reversible and tidy, but it's the version most likely to vanish in transit. Fix: default to burned-in for distribution and keep the SRT as a saved file you can re-attach, not your only copy.
FAQ
Are burned-in captions the same as hardcoded subtitles?
Yes. Burned-in, open, and hardcoded all describe captions rendered permanently into the video pixels during export (Rev). They're always visible, can't be toggled off, and travel with the file. "Soft" or "closed" captions are the opposite, a separate, toggleable track that a player loads at playback.
Do soft captions transfer when I download and repost a clip?
Usually not. A soft caption track is a separate file that doesn't automatically travel with the video on download or re-upload, so the reposted clip shows up with no captions unless you re-attach the track (3Play Media). TikTok's auto-captions are worse on this front, they render only inside the app and vanish from a saved MP4 (BibiGPT).
Will burning in captions hurt my YouTube search ranking?
It won't hurt, it just doesn't help search by itself, because YouTube can't read text that's part of the image. The fix is to do both: burn in your styled captions for mute viewers and upload an accurate SRT so the words are indexable (3Play Media). Treat caption-SEO lift figures from tool vendors as directional, not exact.
Can compression ruin burned-in captions?
It can if you export poorly. Because the text is baked into the image, a low-resolution or low-bitrate export plus the platform's re-encode can leave captions soft or pixelated (Subly). Export at full resolution and a high bitrate and the text holds. Soft captions don't have this problem because they're text, not pixels.
Which should I use for podcast clips, burned-in or soft?
Burned-in for almost everything you distribute. Clips get downloaded, re-shared, and reposted, and burned-in text is the only version that survives those moves. Add a soft SRT on YouTube for search and on any compliance-bound content for accessibility. For the editing decisions around caption look and accuracy, see auto vs manual captions, the best caption fonts for clips, and word-by-word animated captions.
Once captions are burned in cleanly, the next lever is which moments you caption in the first place, see how to pick the best AI-suggested clips.