Hardsubs vs softsubs: why it matters for language learners.

Quick answer

Softsubs are separate text tracks that the video player draws on top of the picture. Hardsubs are pixels burned permanently into the video frames. Softsubs work with language-learning tools like 2Subs because the extension can read the text. Hardsubs do not, because there is no text to read — only an image.

If you searched for "hardsub vs softsub," "what are hardsubs," or "softsub translation," you are probably trying to figure out why your favorite dual subtitle extension works on Netflix but fails on a downloaded anime episode. The short version: streaming platforms ship softsubs, classic fansub releases often ship hardsubs, and study tools depend entirely on the difference.

The technical difference

Softsubs are text. A subtitle file lists each caption alongside a start time, an end time and sometimes styling instructions. The video player loads the file, watches the timeline and draws the captions on top of the picture as it plays. You can turn them off, switch language tracks, change font size, or copy a line to a flashcard. The captions live alongside the video, not inside it.

Hardsubs are part of the picture. A video editor takes the subtitle text, renders it as graphics, and bakes those graphics into every affected frame during encoding. From the moment the file is exported, the captions are pixels — indistinguishable from a character's shirt or a building in the background. There is no track to disable, no language to switch and no way for software to know "this group of pixels is the word arigatou."

Common softsub formats

• SRT (SubRip): the lowest common denominator. Plain text, timecodes, line breaks, almost no styling. Universal across players.
• WebVTT (.vtt): the web standard. Used by HTML5 video, supported by every major streaming platform and most browsers natively.
• ASS / SSA (Advanced SubStation Alpha): the fansub favorite. Supports fonts, colors, positioning, fades, karaoke timing and overlays — perfect for anime sign translations.
• TTML / DFXP: XML-based, common in broadcast workflows and used by some streaming services internally before they convert to WebVTT for delivery.

Common hardsub formats

There is no such thing. Hardsubs do not have a format because they are not a separate file — they are part of whatever video container holds the encoded picture (MP4, MKV, AVI, WebM). Once burned in, the captions follow the video everywhere it goes, exactly as they were rendered.

How streaming platforms deliver subtitles

Every major Western streaming service uses softsubs. That is what makes dual subtitle extensions possible at all.

• Netflix: WebVTT delivered as a separate text track per language. The Netflix player composites the captions on top of the video at playback time. This is also why Netflix dual subtitles can be added by a browser extension that reads the track.
• YouTube: WebVTT for both creator-uploaded captions and auto-generated captions. The player exposes the track to extensions, which is why YouTube subtitle translation is possible without touching the source video.
• Crunchyroll: mostly softsubs in modern releases, delivered as text tracks the web player renders client-side. Some older sign translations and effects use overlay tricks but the dialogue lines are softsub text.
• Disney+: softsubs across the catalog, with multiple language tracks available depending on title and region.
• Prime Video: softsubs, with the player handling the rendering. Track availability varies more by title than on Netflix.
• Apple TV+: softsubs with strong styling and high-quality language coverage on original productions.

The exceptions show up outside the streaming world. Fansub releases from the 2000s and early 2010s, ripped anime DVDs, encoded torrents and a lot of "scene" releases used hardsubs — sometimes because the encoders wanted to guarantee styling, sometimes because the source itself was already burned in. Older karaoke videos, news broadcasts and some user-uploaded clips on the open web are also hardsubbed by default.

Why dual subtitles only work with softsubs

A dual subtitle extension does one thing: it reads the subtitle track the player has loaded, then renders a second line — either a different language track or a machine-translated version — on top of the video. That entire workflow assumes the captions are accessible as text.

With softsubs, the browser already has the subtitle data in memory. 2Subs reads it, sends it through translation if needed, and draws the second line using your styling preferences. With hardsubs, the captions are part of the picture. The browser has no idea they exist; it sees a frame of video, not a sentence of Japanese. There is nothing to read and nothing to translate.

This is also why dual subtitles depend on the same mechanism across every platform: text track in, translated text out. The extension does not care whether you are watching a K-drama or a Spanish thriller, only that the captions are still text by the time they reach your browser. 2Subs' feature set — pronunciation hover, read-aloud, custom overlay styling — all build on top of having that text available.

The OCR workaround (and why it usually fails)

In theory, you can use optical character recognition to extract hardsub text from frames in real time. The pipeline looks like this: grab a frame, run OCR on the lower portion of the picture, send the recognized text to a translation engine, then render the result as a second line. A few experimental tools have tried it.

In practice the approach falls apart for three reasons.

• Accuracy on stylized fonts. Fansub typography is gorgeous and inconsistent — drop shadows, gradients, outline colors, italic mid-sentence, sign translations rotated to match a poster on the wall. OCR engines trained on document scans struggle with all of it. A 5% character error rate is the difference between "I love you" and "I love yu," which is fine; "He was assassinated" becoming "He was Assistantinated" is not.
• Performance hit. Running OCR on every frame of a 24fps video, then translation on every recognized line, then rendering — all in a browser tab — eats CPU. Battery life on laptops gets ugly fast and weaker machines drop frames.
• No real-time sync. Even if OCR is fast, it is not instant. Captions appear on screen, the engine starts processing, the translation arrives a beat late, and now your second line is desynced from the first. For language learning, where you are matching sound to text to meaning, that mismatch is worse than no second line at all.

OCR has its place — archival projects, batch translation of old DVDs — but it is not how a daily study tool works. Every reliable dual subtitle workflow assumes softsubs.

What to do when you hit hardsubs

If the video you actually want to study has hardsubs, you have a few practical options.

• Find the same title on a softsub platform. Most popular anime is now licensed to Crunchyroll, Netflix or Prime Video in major markets. The same episode there will have softsubs and probably multiple language tracks. Compare with the best anime for learning Japanese list to pick a title that is also good study material.
• Search for an SRT for the same release. Subtitle archives exist for almost every popular show. If you legally own the video (you bought the disc, you have the streaming rights) you can pair it with an external SRT file in a local player like VLC. Legality varies by jurisdiction — stick to subtitle files for content you legally have access to.
• Use the original audio with a separate transcript. If the show has an official transcript or a fan-made one, open it on a second screen and follow along with the audio. You lose the timed sync but keep the studyable text.
• Pick a different show. Sometimes the cheapest option is the right one. Plenty of softsub content covers every level and genre — best Netflix shows to learn Spanish and best Netflix shows to learn Korean both run entirely on softsub catalogs.

Why softsubs are simply better for learners

Even if hardsubs weren't a dead end for dual subtitle tools, softsubs would still win on every metric that matters for studying.

• Styleable. Want the captions bigger? Want a black background behind them so they are readable on bright scenes? Want one line in monospace so kanji is clearer? Softsubs let you change all of that without re-encoding anything.
• Searchable. A softsub file is just text. You can grep it for a phrase, build a frequency list of vocabulary, or jump to every scene where a specific word is said.
• Copyable. Drag a line into Anki, paste it into a translator, send it to a friend. With hardsubs you would have to retype it from the screen.
• Tool-friendly. Dual subtitle extensions, pronunciation hover, read-aloud, dictionary popups, sentence mining — every modern study tool reads softsub text. None of them can do anything with pixels.
• Replaceable. Bad translation? Swap the file. Outdated regional dialect? Use a fresh one. Hardsubs lock you into whatever the encoder decided.

How to check if a video has softsubs or hardsubs

The fastest test takes three seconds.

1. Toggle the CC button or open the player's subtitle menu.
2. If you can turn captions off and they disappear while the video keeps playing, they are softsubs.
3. If turning captions off does nothing — the text stays burned on the screen — they are hardsubs.
4. Bonus check: try changing the subtitle language. If a language picker exists and switches the visible text, definitely softsubs.

On streaming platforms this is almost always softsubs. On downloaded files in VLC or mpv, look at the player's subtitle track menu: if multiple tracks are listed, those are softsubs you can switch between. If the menu is empty but you still see captions on the picture, hardsubs.

Quick rules of thumb

• Watching on Netflix, YouTube, Crunchyroll, Disney+, Prime Video or Apple TV+? Softsubs. Dual subtitle tools work.
• Watching a fansub release from a torrent? Often hardsubs. Dual subtitle tools won't.
• Watching a clip on TikTok, Instagram or a news site? Almost always hardsubs (the captions are baked into the video file as it was uploaded).
• Watching a DVD rip? Depends on the source — disc subtitles are technically bitmap subpictures, which behave like hardsubs for OCR purposes but as switchable tracks for player purposes.

For the rest of this guide's worth of practical setup, see the 2Subs pricing page for the free install and Pro upgrade options, or jump to the related guides below.

FAQ

Are Netflix subtitles hardcoded or softcoded?

Netflix subtitles are softcoded. Netflix streams captions as a separate WebVTT-style text track that your player renders on top of the video, which is why language pickers can switch tracks instantly and why extensions like 2Subs can read them to render a second subtitle line.

Can 2Subs work on hardsubbed anime?

No. 2Subs reads the subtitle track that the streaming player exposes to the browser. Hardsubs are pixels burned into the video frames, so there is no text data to read, no track to switch and no way to add a translated second line. If you want dual subtitles for anime, watch on Crunchyroll, Netflix or another softsub platform.

What file format is a softsub?

The most common softsub formats are SRT (SubRip, plain text with timecodes), WebVTT (web-friendly variant used by streaming platforms), ASS/SSA (Advanced SubStation Alpha, supports styling and positioning) and TTML/DFXP (XML-based, used in broadcast and by some streamers).

Why do fansubs use hardsubs?

Fansubbers historically burned subtitles into the video to preserve typesetting, karaoke effects and sign translations that softsub formats could not always carry. Hardsubs also guaranteed that anyone playing the file saw the captions, regardless of player or codec support. Modern releases lean softsub because tooling has caught up and viewers expect to be able to turn captions off.