Students need captions that are correct on the terms, timed to the speaker and available on every device, and a lecture gets them in three steps: split the recording into chapters of 10 minutes or less, run each chapter through a caption generator, and fix the terminology before you export. The research is not in dispute. Morton Ann Gernsbacher's 2015 review "Video Captions Benefit Everyone", drawing on more than 100 studies, concluded that "captioning a video improves comprehension of, attention to, and memory for the video", for hearing students as much as for deaf and hard-of-hearing ones.
This guide covers what the evidence says, why the 10-minute limit is a feature rather than a problem for lectures, when to burn captions in versus using the LMS or YouTube caption track, the terminology pass that course video needs, and how Indian-language and Hinglish lectures fit in.
What students get from captions
Three findings carry the argument.
Comprehension and memory improve for everyone. Gernsbacher's review (Policy Insights from the Behavioral and Brain Sciences, 2015) summarizes eye-movement studies showing viewers read captions without effort, and studies showing better memory for course content when captions are present. It also notes that captions in the language of the video help second-language learners more than captions in their own language: in a study of about 150 Japanese students learning English, English captions produced far better recall than Japanese ones. For a course taught in English to students whose first language is something else, that is the case for same-language captions in one sentence.
Most caption users are not disability-registered. A national study by Katie Linder of Oregon State University's Ecampus Research Unit with 3Play Media, released 31 October 2016, surveyed 2,124 students at 15 US universities. 98.6% of students who used captions said they were helpful, and about 75% used them as a learning aid, mostly to stay focused, retain information and cope with poor audio. Only 13% were registered with a disability services office. Captions are a study tool first and an accommodation second.
Captions are a baseline accessibility requirement. WCAG 2.1 Success Criterion 1.2.2, Captions (Prerecorded), is a Level A criterion, the minimum level, and requires captions for all prerecorded audio in synchronized media. The W3C's guidance says captions should identify the speaker and include meaningful sound effects, not just the dialogue. Institutions in the EU have been subject to the European Accessibility Act since 28 June 2025, and US public institutions face the Department of Justice's ADA Title II web rule from April 2024, with compliance dates in April 2026 and April 2027 depending on size. If you teach at a public institution, captions on course video are no longer optional.
The 10-minute limit, and why it suits lectures
CapsAI captions one video of up to 10 minutes at a time, as an MP4 or MOV, with a 500 MB file cap on free accounts and 10 GB on paid packs. A 50-minute lecture therefore becomes five or six uploads. That sounds like a constraint. In practice it forces the split that course designers recommend anyway.
- Students watch chapters, not hours. A 50-minute single file is skipped through; six 8-minute chapters are watched in order, and the one on the topic a student is revising is watched again. Each chapter gets its own title and its own thumbnail, which makes the course navigable.
- The edit stays manageable. Reading a caption draft for an 8-minute segment takes 15 to 20 minutes. Reading one for 50 minutes in one sitting leads to skipped lines.
- Re-recording is cheaper. When a formula changes next semester, you re-record one chapter.
- YouTube chapters work the same way. If you publish to YouTube, its help page on video chapters (read 18 October 2026) requires the first timestamp at 00:00, at least three timestamps, and chapters of at least 10 seconds. Cutting the lecture at natural breaks before you caption gives you those timestamps for free.
Cut at topic boundaries, not at a timer. A 12-minute chapter can be split at the worked example; a 4-minute recap can stand alone. Trim the "can everyone hear me" at the start of each file before you upload; it saves minutes and reads badly as a caption.
Burned-in captions or a caption track?
There are two ways to deliver captions, and course video often needs both. The full comparison is in burned-in captions vs closed captions; here is the version for lectures.
Burned-in (open) captions are drawn into the picture. They show on every player, in every LMS, in a WhatsApp forward and in a downloaded file, with the same font and position you chose. Students cannot turn them off, screen readers cannot read them, and there is no separate text to search. This is what CapsAI exports: one MP4 with the captions burned in, every word and timing edited in the browser first.
A caption track (closed captions) is a separate file, usually SRT or VTT, attached to the video on YouTube, Vimeo or the LMS player. Students can toggle it, change its size in the player, and in some players search it. It survives only where it was attached: a download or a repost loses it. YouTube's help page on supported subtitle files (read 18 October 2026) accepts .srt, .sbv, .vtt and several broadcast formats.
| Need | Burned-in MP4 | Caption track on the platform |
|---|---|---|
| Works in any LMS player or a shared file | Yes | Only where attached |
| Student can turn captions off or resize them | No | Yes |
| Searchable transcript for revision | No | Often, depending on the player |
| Readable by a screen reader | No | Yes, in players that expose the track |
| Survives download and reposting | Yes | No |
| Styling you control | Yes, fully | Limited to the player |
A sensible pattern for a course: burn captions into the chapter files you distribute and embed, and if your platform supports it, also publish a caption track there for students who need to resize or search. CapsAI does not export SRT or VTT; the platform's own caption tools, or a transcript you have corrected run through the TXT to SRT generator, cover the track side.
The terminology pass
Lecture audio is usually clean and single-speaker, which the recognizer likes. What it does not know is your subject. Before you open the editor, write a short list of the words that must be right in this chapter, then search the draft for each one. Typical entries:
- Technical terms and their spellings. Eigenvalue, mitochondria, amortization, Keynesian. The draft will reach for a common word that sounds similar.
- Names. Authors you cite, historical figures, the names of students who asked questions if you keep them in.
- Symbols said aloud. "x squared", "delta t", "p less than 0.05". Decide how you write them and apply it in every chapter.
- Numbers and units. Dates, equation numbers, page references, percentages. Check every digit.
- Acronyms. GDP, DNA, SQL, UPSC. The draft may write them as words or lower-case them.
- Course-specific labels. "Assignment 3", "Module 2", "Table 4.1".
Fix every occurrence, not the first one. Keep the list; the same words come back in the next chapter and in next semester's re-recording. In CapsAI every caption line is editable before export, and the pass for an 8-minute chapter takes about 15 minutes once the list exists.
Two more checks for lecture video specifically. Negations: a dropped "not" in "this does not converge" reverses the teaching point. And pauses: a lecturer thinks aloud, and the draft may merge two half-sentences into one line. Break the line where you paused.
Indian-language and Hinglish lectures
A large share of coaching and university video in India is delivered in Hindi, or in the Hinglish mix of Hindi grammar and English technical terms. CapsAI supports 11 spoken languages: English, Hinglish, Hindi, Bengali, Telugu, Marathi, Tamil, Urdu, Gujarati, Kannada and Punjabi. Pick the language the lecturer actually speaks. For a lecture that says "is equation ko solve karne ke liye we substitute x", choose Hinglish through the Hinglish caption generator: the draft comes back in Roman script with the English terms kept as English, which is how the students write their own notes. For a lecture delivered fully in Hindi, the Hindi subtitle generator gives Devanagari captions. There is no translation between languages; captions come out in the language spoken, which for a lecture is what you want.
A workflow for one lecture
- Export the recording from Zoom, OBS or your phone as MP4 or MOV.
- Cut it into chapters of 10 minutes or less at topic boundaries, and trim the dead air at the start of each.
- Upload the first chapter to the online course subtitle generator and pick the spoken language.
- Run the terminology list against the draft, fix names, numbers, symbols and negations, and check line breaks at pauses.
- Choose a plain, high-contrast caption style and keep it identical across every chapter of the course. Students should not notice the captions, only read them.
- Export the MP4 and repeat for each chapter. Projects stay in your account for 3, 7 or 15 days depending on the pack, so finish a lecture within that window.
- Upload the chapters to your LMS or to YouTube, and add a caption track there as well if the platform offers one.
For a live session you recorded rather than a planned lecture, the webinar subtitle generator page describes the same steps with the Q&A segment in mind.
What to do next
Take one chapter of an existing lecture, keep it under 10 minutes, and caption it with the online course subtitle generator. A new account has 3 free minutes to start, which is enough to see what the terminology pass involves on your subject. One-time packs start from $4 (₹49 in India, €4 in the euro area) for 20 minutes, the Pro pack covers 180 minutes, and watermark-free export comes with every paid pack.