Rewatching Shows You Already Know, Dubbed: Good Advice With a Vocabulary Ceiling
Ask in any language forum how to get past understanding half of what you hear, and one answer keeps coming up: find something you already know by heart, watch it dubbed into the language you're learning, and let the plot carry the meaning while your ears do the work. It costs nothing and it works. The catch is the ceiling: which words the script happens to contain, and how often it happens to repeat them, not how hard you try.
Published · updated
The short answer
Rewatching something you know well, dubbed, works: you already carry the meaning, so your attention is free for the language. It raises comprehension faster than it grows vocabulary, because a fixed script repeats what it already repeats, and the words you still need are rare. Closing that gap takes something that decides when a word returns.
1 of 6
Rewatch something you already know, dubbed
The community's top answer. The mechanism behind it is real.
Pick something whose plot you could recite from memory: a childhood cartoon, a sitcom you've seen four times, a film you quote at people. Watch it dubbed into the language you're learning. You already carry the meaning, so none of your attention goes on working out what's happening, and all of it is free for how the language says it. Comprehensible-input researchers built a theory around this same mechanism; learners found it first, on forums, with no name for it.
It costs nothing, it needs no account, and unlike most study plans, people stick with it. Start here before anything else on this page.
The ceiling shows up around the third or fourth rewatch. What improves fastest is your comprehension of that episode, then that show, then that actor's register. What grows slowly is the vocabulary you carry out of it: the script decides that. The words a script repeats are the ones a writer needed often, the common ones you already had by B1. The word that stopped you is the one the writer needed once.
That shape has a name. Word frequency in natural language follows Zipf's law: a small set of words accounts for most of everything anyone says, and frequency falls away steeply behind it. Nation (2006) put numbers on what that costs a learner. Following unscripted speech with 98% of the words known takes something in the region of 6,000 to 7,000 word families, and reading unsimplified text takes 8,000 to 9,000. Someone sitting at 3,000 word families is thousands short, and each of those thousands is, by construction, rarer than the 3,000 they already hold. Rarer means the distance between one appearance and the next is measured in hours of listening.
A rewatch hands you one encounter with the word that stopped you, right when you paused to look it up, and then the episode moves on. When that word next appears in front of you isn't your decision.
Best forAnyone stuck around half comprehension who wants the cheapest real gain available, and who has not tried it yet.
2 of 6
Turn on subtitles in the target language
It makes tonight's episode understandable, and it quietly converts listening practice into reading practice.
This is the most common next move, and the most criticised one. The criticism holds up: if the words are on screen, your eyes take them, and you finish the episode having read the script rather than understood the speech. Learners know this and do it anyway, because not understanding is unpleasant enough to end the habit.
Used well, subtitles are narrower than how most people use them: good for confirming a word you half-heard, or for a second pass over a scene you missed completely. Left on as the default for everything you watch, they substitute for the skill you're trying to build.
Either way, subtitles do nothing about the frequency problem. They tell you what the rare word was; they don't decide when you'll meet it again.
Best forA second pass over a scene you did not follow, rather than as the permanent setting for everything you watch.
3 of 6
Comprehensible-input video pitched at your level (Dreaming Spanish and its equivalents)
Hundreds of hours of speech built to be understood at your level, with no exercises attached.
A rewatch borrows its comprehensibility from a plot you already know. Comprehensible-input channels build it into the speech itself: slower, more repetitive, supported by gesture and context, and graded from near-beginner up to native pace. For Spanish, Dreaming Spanish has the deepest library of any single source, thousands of hours, free on YouTube with a paid tier adding tracking and downloads. Most major languages have a thinner version of the same idea.
This is the strongest answer here for the comprehension half of the problem. If native speech feels like a wall, start here, and expect it to work slowly with no score to track it by.
Two limits worth naming: it's listening only, so it does nothing for reading, writing, or speaking, and which words come back, and how often, is still decided by what the speaker chose to talk about, a better-chosen script than a dubbed cartoon, but a script all the same.
Best forLearners whose problem is understanding real speech at all, rather than retaining the specific words they keep losing.
Visit Dreaming Spanish4 of 6
Look words up as you watch (Language Reactor, Migaku, Readlang)
They put the word within reach the moment it stops you. What happens afterwards is still the script's decision.
Language Reactor sits on top of Netflix and YouTube and puts both subtitle tracks in reach, so an unknown word is a hover away. Migaku goes further and assembles the sentence, the audio and a screenshot into a review card while you watch. Readlang does the same for text in a browser. All three remove the part of sentence mining that is pure clerical work, and for someone already watching native content they are the least disruptive tools here.
These are per-word click workflows, so you're mining while you watch, not just watching, and the subscriptions cost more than the free options above. Migaku and Readlang both build a card as you go, which means a review queue comes with it, and the review queue is usually what people quit.
What they solve is meeting a word and understanding it on the spot, and they solve it well. Whether that word comes back to you again is still up to the show, not to anything you clicked.
Best forLearners who already watch or read native content daily and want the lookup handled without leaving it.
Visit Migaku5 of 6
Put the word on a schedule (Anki and spaced repetition)
The direct answer to the frequency problem, and the one with the highest drop-out rate.
Everything above leaves the timing of your next encounter to the content. Spaced repetition takes that decision away from the content and hands it to a scheduler: the word comes back a little before you'd have forgotten it, then later, then later again, until it's yours. Anki is the tool most people use for this, running FSRS-6, an open-source scheduler that targets a constant 90% probability of recall, built for exactly this problem.
This is why it sits at number five, not number one. Every card is built by hand: find the sentence, translate it, format it. Sentence-mining guides put the admin at roughly an hour for every twenty words mined, and that's the part most people abandon in the first week, not the reviewing. There's a quieter second limit too: a word drilled on a card is met in isolation, away from wherever you found it or will need it.
If the card-making is the part that stops you, that has its own answers and we have written them up separately.
Best forLearners who want the timing of every review decided for them and do not mind building the deck by hand.
Visit Anki6 of 6
TraceSpark, for putting the words you met yourself back into something worth reading
The rewatch mechanism with a schedule attached: content you can follow, built from words you chose.
A rewatch works because you already carry the meaning, so your attention is free for the language. TraceSpark reaches the same condition from the other end. Instead of borrowing comprehensibility from a plot you know, it writes at the level you set, and the words in it are ones you met yourself. Capture a word, phrase or expression from any app on your phone, a news article, a message, or a screenshot with the word circled, using Android's share or text-selection menu. It goes into your next chapter or podcast episode, and a second pass checks which of your captured items appeared, so you can see the exact words woven into each chapter.
The scheduling is the part that answers this page. Your captured words and phrases are tracked by FSRS-6, the same open-source scheduler Anki ships, aimed at the same 90% probability of recall, and the review happens by meeting the word again inside a chapter rather than on a card. You grade it there, from a console at the bottom of the reading screen, without leaving the story. Series run to ten chapters with the arc planned before chapter one, so it continues rather than resetting. Audio is written to the device, so a commute with no signal plays fine.
Before you click anything: TraceSpark is Android only, no iOS, no web reader. It's free to install on Google Play and nothing is charged today; paid plans start later. There's no speaking practice and no correction of anything you say or write, so it does nothing for the production half of the plateau; a tutor or a speaking app is the better spend for that. It assumes you can already follow a story at a strong B1 level, and if you already keep an Anki deck, nothing here asks you to give it up.
None of this is a reason to stop rewatching things you love. That habit still costs nothing and still works; this page is about the one thing a show's script can't do for you. You can check the writing before installing anything: we publish one full generated chapter at each of three levels, printed exactly as the app produced it, nothing rewritten.
Best forLearners past strong B1 on Android who already meet words in their own reading and listening, and want those particular words to come back without building a deck.
Frequently Asked Questions
Does rewatching dubbed shows actually help you learn a language?+
It does, and more reliably than most study plans, because people finish watching. Something whose plot you already know, dubbed, frees your attention for the language instead of the story: the same principle comprehensible input rests on. Comprehension improves fastest. Vocabulary grows more slowly, because a fixed script repeats the words it already repeats, and the words you still need are the rare ones it used once.
Why do I understand more and more but still not learn new words?+
The two run on different fuel. Comprehension improves from repeated exposure to the common words and structures a script uses constantly. Vocabulary past the intermediate plateau depends on words that are, by definition, rare: Nation (2006) puts the vocabulary needed to follow unscripted speech at roughly 6,000 to 7,000 word families, and every one you are missing occurs less often than the ones you already have. Meeting a rare word once is not learning it, and one rewatch gives you one meeting.
How many times do I need to see a word before it sticks?+
There's no single number; anyone who gives you one is rounding. Studies of incidental vocabulary learning span a wide range of estimates, but two findings hold across them: one encounter rarely produces durable knowledge, and encounters spread over time beat the same number crammed together. That's the basis of spaced repetition. In practice, the interval matters as much as the count, and a script controls neither.
Is TraceSpark telling me to stop watching things in my target language?+
No. That would be poor advice: watching what you enjoy costs nothing and is the reason most people stick with a language long enough to get good at it. TraceSpark handles one thing a show's script doesn't, putting the words you captured back in front of you on a schedule instead of leaving it to whatever the script says next.
Does this apply to phrases and grammar, or only to single words?+
Words and phrases, yes. The same frequency skew shows up in multi-word sequences too, and it gets steeper as the phrase gets longer, so a set expression you need is rarer in ordinary content than any single word inside it. Grammar is a different case: spaced re-encounter does help it stick, but frequency isn't the lever there that it is for vocabulary.
Ready to see what your own words sound like?
Turn the words you capture yourself into a personalized audiobook chapter you can listen to on your commute.
