"Can I learn a language just by watching TV?" is one of the most-asked questions on every language-learning forum, and the answers range from confident yes to indignant no. The actual data is more interesting than either extreme. There are now decades of input-hypothesis research, a sharp government-funded estimate from the US Foreign Service Institute, and a new wave of large-scale community datasets from immersion-based methods like Dreaming Spanish and Refold. Put together, they let us answer the question with real numbers instead of opinions.
I've pulled the cleanest sources I could find — the FSI School of Language Studies, the Refold methodology documentation, Dreaming Spanish's published roadmap, the 2025 study in the British Journal of Educational Psychology, and Stephen Krashen's input hypothesis literature. Here is what they actually say about how many hours of TV it takes to reach a useful level — and the specific conditions that make those hours count.
The FSI baseline: 600 to 2200 hours
The cleanest hard number comes from the US Foreign Service Institute, which has trained American diplomats in foreign languages for over 70 years. The FSI publishes hour estimates for English speakers to reach "Professional Working Proficiency" (roughly CEFR B2 / ILR Level 3) in around 70 languages, grouped by difficulty.
These numbers are real, classroom-measured outcomes — not marketing — and they have been remarkably stable for decades. Category I languages (Spanish, French, Italian, Portuguese, Dutch, Norwegian, Swedish) take roughly 600 to 750 hours. Category II (German) takes around 900 hours. Category III (Indonesian, Swahili, Malaysian) takes about 1100 hours. Category IV — the largest group, including Russian, Polish, Turkish, Hindi, Hebrew, Vietnamese, Greek, Czech, Thai, Finnish, Hungarian — takes around 1800 hours. And Category V "super-hard" languages — Arabic, Mandarin, Cantonese, Japanese and Korean — take around 2200 hours.
English-speaking adults, classroom hours to ILR Level 3
Source: US Foreign Service Institute, School of Language Studies. Estimates assume an adult English-speaking learner with full-time classroom instruction.
The FSI explicitly notes these estimates are floors, not ceilings, and they assume intensive classroom training (typically 25 hours per week plus 3–4 hours of homework per day). They do not include the additional immersion most learners actually need to retain the language.
The immersion-only number: roughly 1500 hours to upper-intermediate
The other useful benchmark comes from the comprehensible-input community — specifically from Dreaming Spanish, which has tracked thousands of self-reported learner hours since 2018. Their published roadmap suggests 1500 hours of comprehensible input to reach what they call "advanced beginner to intermediate" — comfortable understanding of native-level Spanish in most contexts. That number lines up with Stephen Krashen's input hypothesis (i+1: input must be slightly above current level) and with the Refold method's estimates for the immersion-heavy phase.
Importantly, "comprehensible input" is not the same as "TV in the background." For hours to count, you have to understand most of what you are hearing — Krashen's original benchmark is that the learner should understand the gist, with new vocabulary acquired from context. Watching K-drama with zero Korean knowledge is not comprehensible input; it is decorative noise. The same K-drama with English subtitles (then later Korean subtitles, then bilingual) can be highly comprehensible — that is the difference between hours that compound and hours that evaporate.
Translating those hours into TV episodes
So how much TV does this actually mean? Streaming dramas average around 45–55 minutes per episode. Anime is typically 22 minutes. K-dramas frequently run 60–80 minutes per episode for the prestige format. Working with a midpoint of 50 minutes:
600 hours of input (entry-level FSI for Spanish/French) ≈ 720 episodes ≈ roughly 90 seasons of a standard 8-episode series.
1500 hours (immersion-only benchmark) ≈ 1800 episodes ≈ roughly 225 seasons.
2200 hours (FSI for Japanese/Korean) ≈ 2640 episodes ≈ roughly 330 seasons.
To reach B2 / upper-intermediate by language difficulty
Calculation: FSI hour estimates ÷ 50-minute average episode length. Assumes active, subtitle-assisted viewing — not background TV.
The episode counts are sobering, but they are also achievable. The average American Netflix subscriber watches roughly 3.2 hours per day, according to Nielsen 2024 streaming-share data — about 1170 hours per year. A K-drama fan who watches 1.5 hours per day in Korean (a single episode) would hit 547 hours in one year, 1095 hours in two, and the lower FSI estimate for Korean in around four years of consistent viewing.
For a Category I language like Spanish, the math is far gentler: 1.5 hours per day of Spanish TV plus 30 minutes of vocabulary work reaches the lower FSI estimate in around 12 months — verified by hundreds of Dreaming Spanish community members posting timestamped progress reports.
Why subtitles change the equation
Here is where the research becomes practically actionable. A 2025 study published in the British Journal of Educational Psychology by Yuan et al. compared three subtitle conditions: target-language only (L2), native-language only (L1), and bilingual (both shown simultaneously). The key finding: sequential exposure — first L1, then bilingual — produced the strongest vocabulary gains, outperforming either condition alone.
An earlier eye-tracking study in Studies in Second Language Acquisition (2024) demonstrated that bilingual subtitles do not slow comprehension; viewers efficiently switch attention between tracks. The cognitive cost is lower than intuition suggests, and the vocabulary acquisition rate is consistently higher than single-track viewing for intermediate learners.
The practical implication: the same 1500 hours of TV can yield very different results depending on the subtitle setup. Pure L1 subtitles (English subtitles on a Korean show) are a comfortable on-ramp but become a crutch — your brain optimizes for reading English and stops processing Korean audio. Pure L2 subtitles (Korean subtitles on a Korean show) are best for advanced learners but quickly become discouraging at A2 or below. Bilingual — both tracks shown together — is the configuration most consistently supported by current research for intermediate learners.
What "passive" watching actually delivers
The honest answer: very little, if your level is low. A 2021 meta-analysis on incidental vocabulary acquisition from TV found that learners at A1–A2 acquire only 1–3 new words per hour of subtitled viewing, and almost none from un-subtitled foreign content. At B1+ the rate rises to 5–10 words per hour, and at B2 it climbs further as more of the input becomes comprehensible.
The implication for low-level learners is that the first 100 hours of TV are inefficient unless paired with active study, well-chosen subtitle configuration, or a graded-content source (children's shows, comprehensible-input YouTube). Once you cross the comprehensibility threshold — typically around 1500-word receptive vocabulary, or roughly an upper-A2 level — the hours start to compound, and TV becomes one of the most effective input sources available.
How much time do streaming viewers actually have?
For context, here is what the average viewer is already spending on streaming. Nielsen's 2024 streaming-share data shows US streaming households watching around 3.2 hours per day on average, and Netflix subscribers reportedly average around 2 hours per day. That is between 730 and 1170 hours per year of streaming viewing already happening — most of it currently spent on English-language content.
If even one-third of that time were redirected to target-language content with the right subtitle setup, most adults could reach B2 in a Category I language in 2–3 years without ever sitting down to formal study. For Japanese or Korean, the runway is closer to 6–8 years on a one-third allocation, or 3–4 years for a serious learner who dedicates most streaming time to the target language.
The practical takeaways
One: the FSI numbers are the floor. If you are not in an intensive classroom program, expect to need at least the FSI estimate, and probably more. The immersion-only path tends to land between 1× and 2× the FSI number — and that's only with deliberate, comprehensible input.
Two: subtitle configuration is the lever that turns hours into results. Single-track L1 (English on a Korean show) is the comfortable default, but the research consistently shows that intermediate learners gain more from bilingual subtitles. For beginners, sequential exposure — first L1 to build context, then bilingual — is the strongest supported pattern.
Three: the comprehensibility threshold matters more than total hours. Below B1, raw TV time is inefficient. Above B1 (~1500-word receptive vocabulary), each hour compounds. Most learners would be better off front-loading vocabulary apps or graded-content viewing for the first 100–200 hours, then transitioning to native streaming with assistive subtitle tools.
Four: most adults already have the hours. Nielsen data says the average US streaming household watches 1100+ hours per year. The math problem is not finding time — it is shifting that time toward target-language content while keeping the viewing experience enjoyable enough to sustain for years.
Tools that let you watch the shows you already enjoy with dual subtitles in your target language — without changing platforms or losing access to the original — are what make the second and third points operationally possible. Sublo translates the subtitles already on screen in Netflix, Disney+, HBO Max, Amazon Prime, Apple TV+, Crunchyroll, YouTube and the rest, using Gemini AI, into 40+ languages — including dual-subtitle mode for the configuration the research supports. It exists because the gap between "I want to learn from TV" and "the platform actually supports it" is wider than streamers will close on their own.
The short version
The FSI says 600–2200 hours, depending on language. The immersion-only literature says around 1500 hours of comprehensible input for upper-intermediate. At one hour of TV per day, that is two to six years. For most adults the time is already there in their streaming habits — what is missing is the configuration that turns watching into acquisition.
If you want to watch your favorite shows on Netflix, Disney+, HBO Max, YouTube and 8 other platforms with dual subtitles in 40+ target languages — Sublo translates the subtitles in real time, with the bilingual setup the research supports.
Install Sublo — free