Hard Subtitles vs Soft Subtitles: Which Should You Use?

By whispertranscribe.ai Editorial Team · Last updated: November 2026 · 10 min read

Hard subtitles vs soft subtitles comes down to one question: should viewers be able to turn the text off? Hard (burned-in or hardcoded) subtitles are drawn into the video frames, so they always show. Soft subtitles sit in a separate track or file, such as SRT or VTT, so viewers can switch them on, off, or to another language.

Below: a side-by-side comparison, how both map to open vs closed captions, platform support, a decision table, three FFmpeg commands we ran ourselves, and a pre-publish checklist.

Hard subtitles vs soft subtitles at a glance

Hard subtitles win on "it shows everywhere." Soft subtitles win on everything that involves changing, translating, or switching off the text.

Question Hard subtitles (burned-in) Soft subtitles (SRT, VTT, embedded track)
Where the text lives Rendered into each video frame In a sidecar file or a subtitle track in the container
Can viewers turn them off? No Yes, from the player's CC or subtitle menu
Can viewers resize or restyle them? No Often; YouTube lets viewers change font, color, and size
More than one language One video export per language One video, one subtitle file or track per language
Fixing a typo Re-export the whole video Edit the text file and re-upload
Video quality Video must be re-encoded Video can be copied as-is
Works where the player ignores subtitle files Yes No
Equivalent caption term Open captions Closed captions

What are hard subtitles?

Hard subtitles are text rendered permanently into the video image during export. Burned in subtitles, hardcoded subtitles, and "hardsubs" all mean the same thing. Once the file is exported, the words are pixels, exactly like a logo or lower third.

So they display on any player, app, or site, because the player doesn't need to understand subtitles. But you can't undo them: there's no text to extract, no off switch, and any correction means exporting the video again.

There's also a small quality cost. Burning in text means re-encoding the video; the Aegisub manual notes that sharp-edged subtitle text typically needs around 3 to 10% more bitrate to keep the same picture quality.

What are soft subtitles?

Soft subtitles are timed text that the player overlays during playback. The video itself never changes. They come in two forms:

  • Sidecar files. A separate file uploaded or stored next to the video, most often SRT (SubRip) or VTT (WebVTT). This is what YouTube, Vimeo, and Facebook ask for when you upload captions.
  • Embedded tracks. The text is stored inside the video container as its own track. MKV and MP4 can both carry text subtitle tracks.

Soft subtitles in SRT are the most common setup. An SRT file is plain text with numbered cues and timestamps, so you can fix a name in Notepad and re-upload in a minute.

The trade-off: if an app ignores subtitle files, or the MP4 travels without its SRT, the subtitles don't appear.

Open vs closed captions: how they map

Open captions are the caption version of hard subtitles, and closed captions are the caption version of soft subtitles. The open/closed split describes whether viewers can turn the text off, not what the text contains.

The W3C's WCAG guidance on prerecorded captions defines closed captions as ones that "can be turned on and off with some players" and open captions as "any captions that cannot be turned off." It lists both as sufficient techniques for that success criterion, so burned-in captions aren't automatically less accessible.

What matters is the content. Captions identify speakers and describe meaningful sounds, such as [door slams]; subtitles usually carry dialogue only. Either can be burned in or delivered as a soft file. (Accessibility requirements vary by country and sector; this is general information, not legal advice.)

When to use hard vs soft subtitles

Use soft subtitles whenever the platform supports them, and hard subtitles when you can't control the player or the text is part of the design.

  • Hard subtitles fit: short social clips that autoplay muted, videos that get forwarded or reposted (sidecar files get lost), branded captions with a set font, color, or word-by-word highlight, and screens at events or kiosks.
  • Soft subtitles fit: YouTube, Vimeo, and Facebook uploads, multilingual releases (one SRT per language instead of one video per language), courses that get updated, HTML5 video on your own site via WebVTT, and MKV film libraries with several language tracks.

Which platforms support soft subtitles?

YouTube, Vimeo, and Facebook take uploaded subtitle files. TikTok and Instagram generate their own captions from speech, and we didn't find a caption-file upload described in either help center. The table only includes what each official help page or spec says, checked November 2026.

Platform or container Soft subtitles Hard subtitles What the official source says
YouTube Yes: SRT, WebVTT, and other formats Yes, as part of the video Lists supported caption files including SubRip and WebVTT (YouTube Help); viewers can turn captions on or off and change font, color, opacity, and size (caption settings)
Vimeo Yes: caption and subtitle files from the Languages panel Yes, as part of the video Upload a file, set Language and Type (Subtitle or Caption), then toggle it on (Vimeo Help)
Facebook Yes: SRT Yes, as part of the video SRT files need a locale in the name, such as clip.en_US.srt (Facebook Help)
TikTok Platform captions only: auto-generated and creator captions you can edit Yes, as part of the video Captions are generated from speech after you select the video language; you can edit or remove them (TikTok Support)
Instagram Platform captions only: automatic captions from speech recognition Yes, as part of the video Captions are created automatically, and viewers can choose to always or never show them (Instagram Help)
MP4 file Yes, as a text track (mov_text / 3GPP Timed Text) Yes Player support varies; HandBrake's docs say soft tracks need "the correct playback software" (HandBrake docs)
MKV file Yes: multiple tracks, including SRT, SSA/ASS, WebVTT, and image-based PGS or VobSub Yes Matroska defines how each subtitle codec is stored (Matroska spec)
WebM file Not in HandBrake Yes HandBrake: "Soft subtitles are currently unsupported in the WebM container"

For TikTok and Instagram: if you need your own reviewed wording or styling, burn it in. If the automatic captions are good enough, edit them in the app.

Decision table: hard or soft, by scenario

Scenario Use Why
Uploading to YouTube Soft (SRT or VTT) Viewers can toggle and restyle captions; you can add more languages later
Facebook post Soft (SRT) Use the locale filename rule; burn in only if the clip will also be reposted elsewhere
TikTok or Instagram Reels Hard, or the app's own captions No caption-file upload described in their help centers
Clip sent by chat app or embedded in slides Hard Sidecar files get lost in forwarding
Course on your own site Soft (VTT) HTML5 players read WebVTT tracks; easy to update
Same video in five languages Soft Five small text files instead of five video exports
Screen at an event or kiosk Hard You can't count on the device's subtitle settings
A few foreign-language lines in an otherwise native video Hard for those lines, soft for the rest Everyone needs those lines, whatever their settings

How to burn in or add soft subtitles with FFmpeg

FFmpeg is a free, third-party command-line tool. It isn't part of whispertranscribe.ai. The commands below are minimal examples built from FFmpeg's official documentation. We ran each one with FFmpeg 7.1.5 on Linux against a 10-second test clip we generated, then checked the output with ffprobe. File names are placeholders.

1. Burn subtitles into the picture (hard subtitles):

ffmpeg -i input.mp4 -vf subtitles=subs.srt -c:a copy output-burned.mp4

The subtitles filter draws the SRT onto every frame, so the video is re-encoded; -c:a copy leaves the audio alone. It needs an FFmpeg build with libass. A force_style option changes font and color.

2. Add a soft subtitle track to an MP4 (no re-encoding):

ffmpeg -i input.mp4 -i subs.srt -map 0 -map 1 -c copy -c:s mov_text -metadata:s:s:0 language=eng output-soft.mp4

-c:s mov_text converts the SRT to MP4's timed-text format. Video and audio are copied as-is; the language tag labels the track.

3. Put two languages in one MKV (no re-encoding):

ffmpeg -i input.mp4 -i subs.en.srt -i subs.es.srt -map 0 -map 1 -map 2 -c copy -metadata:s:s:0 language=eng -metadata:s:s:1 language=spa output-soft.mkv

MKV keeps the SRT tracks as they are. In our test, ffprobe listed two subrip subtitle streams tagged eng and spa.

Prefer a GUI? HandBrake, another free third-party app, can import SRT files from its Subtitles tab and output them either as a "Hard Burn" or as soft, selectable tracks. Its docs note it can burn only one subtitle track into a file. Most video editors can also import an SRT and render it into the export; check your editor's help pages, since menus differ by version.

Whichever tool you use, keep the clean master video and the subtitle file so you can fix a typo or add a language later.

Create the subtitle file first with whispertranscribe.ai

Hard or soft, you start with an accurate, timed subtitle file. whispertranscribe.ai, a browser-based transcription tool powered by OpenAI's Whisper model, creates that file. It doesn't burn subtitles into video. It exports SRT or VTT files that you upload as soft subtitles or burn in with a video editor or player tool such as the ones above.

  1. Open the AI subtitle generator, or the MP4 to SRT converter for video files.
  2. Choose the spoken language from 130+ languages. Many languages have a Noise / Low-Quality Audio/Video option for rough recordings.
  3. Upload a file up to 5GB, or paste a YouTube, Vimeo, Google Drive, or Dropbox link into Transcribe from URL.
  4. Review names, numbers, and jargon. Accuracy reaches up to 99% on clear audio, but background noise and strong accents lower it, and burned-in mistakes are permanent.
  5. Export SRT or VTT, then upload it or burn it in.

Other starting points: MP3 to SRT for audio, translate audio to English for English text from other languages, the video to text converter for plain transcripts, and the full features list. You can start for free with no account required; paid plans start from $9.99/month billed annually (see pricing).

Pre-publish checklist for hard and soft subtitles

Accessibility

  • Captions name speakers when it isn't obvious who's talking
  • Meaningful sounds are described, such as [phone rings]
  • No more than two lines on screen at once
  • Burned-in text has strong contrast (an outline or background box) and doesn't cover faces or on-screen text

Multilingual

  • The source-language file is final before anyone translates it
  • One file per language, clearly named, such as video.es.srt
  • Translated lines checked for length, since they often run longer

Platform compatibility

  • Social clips for TikTok or Instagram: hard subtitles inside the safe area, clear of on-screen buttons
  • YouTube and Vimeo: SRT or VTT uploaded and switched on
  • Facebook: locale in the filename, such as .en_US.srt
  • MP4 with soft tracks: tested in the player your audience actually uses

FAQs

Are hard subtitles or soft subtitles better?

Soft subtitles are better wherever the platform supports them, because viewers can turn them off, pick a language, and resize them, and you can fix mistakes without re-exporting the video. Hard subtitles are better when the player ignores subtitle files, when the video will be forwarded or reposted, or when the text is part of the design. Many creators keep a soft SRT for YouTube and burn in a separate copy for short social clips.

What is the difference between hardcoded and burned-in subtitles?

There's no difference. Hardcoded, burned-in, burnt-in, and hard subtitles all describe text rendered permanently into the video frames during export. The same idea is called open captions when the text includes captions for deaf and hard-of-hearing viewers, such as speaker names and sound descriptions. In every case the viewer can't hide the text, and changing it means exporting the video again from the original project or master file.

Are SRT files soft subtitles?

Yes. An SRT file is a plain-text sidecar file that holds numbered captions and their start and end times, and the player overlays the text while the video plays. That makes it a soft subtitle format. You can upload an SRT to YouTube, Vimeo, or Facebook, add it to an MKV or MP4 as a selectable track, or use it as the source for burning hard subtitles with a tool such as FFmpeg or HandBrake.

Can you remove hard subtitles from a video?

Not cleanly. Burned-in text was merged into every frame at export, which leaves no separate layer to delete. Blurring or cropping the text area damages the image, and pixels behind the words can't be recovered. The practical fix is to find the original video without subtitles, or the editing project, and export again. If you need new subtitles, transcribe the audio and create a fresh SRT rather than trying to extract the burned-in text.

Can whispertranscribe.ai burn subtitles into my video?

No. whispertranscribe.ai creates the subtitle file, not the finished video. It transcribes your audio or video in 130+ languages and exports SRT or VTT, which you can upload to platforms as soft subtitles. To get hard subtitles, open that file in a video editor or a third-party tool such as FFmpeg or HandBrake and export a new video with the text burned in. Keep the SRT so you can fix lines and burn in again later.


Need a subtitle file for your next video? Upload a file or paste a link into the whispertranscribe.ai subtitle generator and export SRT or VTT in 130+ languages.

References

Ready to turn your recordings into text?

Transcribe audio and video with whispertranscribe.ai.

Start transcribing ↗