How to turn a lecture recording into notes

Why the first 24 hours matter

Two findings from learning research shape this routine. The first is spacing: short sessions spread over days beat one long session. The UNC Learning Center advises a short piece of work for each class every day, and warns that rereading notes leads to quick forgetting. The second is retrieval practice: pulling an idea out of memory strengthens it more than reading it again. The Washington University Center for Teaching and Learning cites studies in which recall from memory beat repeated studying for long-term retention.

One day is a practical deadline, not a magic number. You still remember the room and the board, and the next lecture has not yet built on this one. The first pass makes the notes. The later passes, spaced over days, are self-tests from them.

The 24-hour routine: record lectures and turn them into notes

Budget about 30 minutes the same day and ten minutes two days later. It fits a 9 a.m. lecture you review that evening at the kitchen table.

1. Record the lecture openly. Ask the professor first, then put the iPhone flat on the desk or record on an Apple Watch. See how to record lectures on iPhone. You get one audio file per lecture.
2. Transcribe it the same day. A searchable transcript lets you find specific passages without replaying the entire lecture.
3. Skim the transcript with the slides open. Ten minutes. Find where the professor says "the main point", "two reasons", or "this matters for the exam".
4. Write notes under the four headings. Fifteen minutes, one page, definitions in your own words.
5. Compare your notes with the summary. Five minutes, after you wrote your own page. Add what you missed. Fix what it got wrong.
6. Test yourself two days later. Close the notes and write everything you remember on a blank page. Then mark the gaps. Repeat before the exam.

A simple notes structure: four headings

Four headings force you to decide what matters. One page per lecture is enough.

• Main claim. One or two sentences. If you cannot write it, read the last five minutes of the transcript.
• Key terms. Five to ten terms, each with a one-line definition in your own words and the slide number.
• What the professor stressed. Anything repeated, written on the board, or flagged with "the common mistake is" or "this will be on the exam".
• Questions. What you did not follow, and the questions you expect on a test. The Learning Scientists recommend writing your own questions as retrieval practice. They are your self-test prompts.

How to use a transcript without re-listening

A transcript lets you treat a 50-minute lecture like a document.

• Search for signposts: "important", "remember", "for example", "in summary", "exam". Each hit marks a place the professor cared about.
• Read the last five minutes first. Some lecturers sum up at the end.
• Jump to the gaps in your class notes. Fill them from the text.
• Play audio only for a passage the transcript garbled: a formula, a name, a number. Use the timestamp and listen for 30 seconds.
• Confirm technical terms against the slides. Transcription mishears names and jargon.

What a lecture summarizer does well, and where it fails

An AI summary of a lecture recording is useful for structure: the order of topics, a first list of terms, and a check of what your notes missed.

It fails in predictable ways. It does not know what matters for your exam. It can flatten a careful "usually" into a wrong "always", reverse a definition, or drop the aside where the professor said what will be tested. And reading a summary is passive: the Association for Psychological Science reports that a 2013 review rated practice testing and spaced practice high-utility and rereading low-utility. Treat the summary as a draft, not a page to reread.

Worked example: a fictional 50-minute lecture

Everything in this example is invented to show the method. The lecture is week 6 of an introductory microeconomics course: 50 minutes on price elasticity of demand. The student recorded on an iPhone, transcribed that evening, and wrote this page.

• Main claim. Elasticity measures how much quantity demanded changes when price changes. It decides whether a price rise raises or lowers total revenue.
• Key terms. Price elasticity of demand: percent change in quantity divided by percent change in price (slide 4). Elastic: more than 1. Inelastic: less than 1 (slide 7). Total revenue: price times quantity (slide 9). Determinants: substitutes, necessity or luxury, share of income, time horizon (slide 12).
• What the professor stressed. "The sign is negative; we report the absolute value." Said three times. Use the midpoint method on the problem set. The exam will give a revenue change and ask which way elasticity goes.
• Questions. Why does a longer time horizon raise elasticity? Transcript at 31:40: more substitutes appear over time. For office hours: midpoint or point elasticity on the exam?

Then came the comparison with the summary. It listed the four determinants, which the class notes had missed. It also wrote "elastic goods have an elasticity below 1", a reversal; slide 7 showed the opposite, so the student fixed it. It dropped the "absolute value" line the professor repeated three times. Two days later, the blank-page test recovered the main claim and three determinants. "Share of income" was missing, so it became a question for the next test.

Questions people ask