How to cut a long lecture recording into separate episodes
- Step 1Check your LMS upload limit and attention target — Note your platform's per-file cap (many LMS instances limit to a few hundred MB to a couple of GB) and decide a target episode length — 15–25 minutes suits most lectures. Both drive the part count.
- Step 2Convert the target to a part count — For attention span:
parts = round(total minutes ÷ episode minutes). A 180-minute lecture at 20-minute episodes is180 ÷ 20 = 9parts. For an LMS size cap, also checkceil(file MB ÷ cap MB)and use whichever count is larger. - Step 3Drop the recording and set Number of parts — Accepts MP4, MOV, MKV, WEBM, AVI, M4V, TS. Enter your count (2–20). There is no 'episode at each slide change' mode — the split is by even time.
- Step 4Run the lossless split — FFmpeg.wasm extracts each episode with
-ss START -i recording -t SEGLEN -c copy -avoid_negative_ts make_zero. Slides and whiteboard text stay sharp because nothing is re-encoded. - Step 5Upload episodes to the LMS in order — Extract
name-parts.zipand uploadname-part-1…name-part-Nas separate items or a module sequence. Label them by topic where you can ('Episode 3 — Eigenvalues'). - Step 6Refine boundaries on topic changes if needed — Equal time won't align with slide transitions. If a topic must start cleanly, find the timestamp (or run scene-detector for slide-change cues) and trim at it with lossless-trimmer.
Part count for lecture length and episode target
Equal split divides total runtime by part count. Pick the count whose episode length fits attention spans and your LMS cap.
| Lecture length | ~15 min episodes | ~20 min episodes | ~25 min episodes |
|---|---|---|---|
| 60 min | 4 parts | 3 parts | 2–3 parts |
| 90 min | 6 parts | 5 parts | 4 parts |
| 120 min | 8 parts | 6 parts | 5 parts |
| 180 min | 12 parts | 9 parts | 7 parts |
Why segment a lecture
Benefits map to specific classroom problems; the splitter addresses the file side of each.
| Problem | How segmenting helps |
|---|---|
| LMS rejects the large upload | Each episode is a fraction of the size and clears the per-file cap |
| Students don't finish a 3-hour video | Shorter episodes raise completion and let students resume per topic |
| Hard to find a specific topic | Topic-titled episodes are directly navigable in the module list |
| Privacy of recorded faces / unreleased material | In-browser split means nothing uploads to a third party before the LMS |
Splitter behaviour relevant to educators
Real tool facts that affect classroom workflows.
| Property | Value |
|---|---|
| Part range | 2–20 |
| Quality | Lossless stream-copy (slides/text stay sharp) |
| Boundary | Nearest keyframe, not slide-aware |
| Output | ZIP of ordered episodes, same container |
| Privacy | 100% in-browser, no upload |
Cookbook
Cutting lectures into teachable episodes. Counts target attention spans and LMS limits.
3-hour lecture into 9 episodes of ~20 minutes
The headline case: a marathon recording into attention-span-sized episodes.
Input: cs101-week5.mp4 (180:00) Math: 180 / 20 = 9 Option: Number of parts = 9 Output: cs101-week5-parts.zip cs101-week5-part-1.mp4 ... -part-9.mp4 (~20:00 each) Upload as Module 5, items 1-9.
Fit episodes under a 500 MB LMS cap
When the LMS limit is the binding constraint, size by bytes as well as by time.
Input: seminar.mp4 (2.4 GB, 120:00)
LMS cap: 500 MB/file
Math: ceil(2400 / 500) = 5 (size)
120 / 20 = 6 (attention)
Option: Number of parts = 6 (use the larger count)
Result: 6 episodes, each ~400 MB and ~20:00.Keep slide text razor-sharp
Re-encoding softens fine text on slides. Stream-copy avoids that entirely.
Input: math-lecture.mkv (dense equations on slides) Option: Number of parts = 6 -c copy -> no re-encode -> equations stay pixel-perfect in every episode. Container stays MKV.
Align an episode to a topic change
Equal time won't hit the slide transition. Find the change, then trim.
Want Episode 4 to start at the 'Recursion' slide (1:02:15): 1. scene-detector flags the slide cut near 1:02:14 2. lossless-trimmer cut at 1:02:14 for a clean start (Routes: /video-tools/scene-detector, /video-tools/lossless-trimmer)
Protect a recording with student faces
Recorded seminars often show students. In-browser splitting avoids exposing them to any host.
Input: seminar-discussion.mp4 (student faces visible) Option: Number of parts = 5 FFmpeg.wasm processes locally - no upload. If faces must be hidden in the published version, blur them first with face-blur, then split. (Route: /video-tools/face-blur)
Edge cases and what actually happens
Episode boundary lands mid-topic
By designEqual-time splitting cuts near even marks on keyframes, not at slide or topic changes, so an episode may start in the middle of a concept. For topic-aligned cuts, find the change point (or use scene-detector for slide transitions) and trim with lossless-trimmer.
Recording shows identifiable students
Handle privacyThe split is private (in-browser), but the content still shows whoever was recorded. For FERPA/GDPR-sensitive seminars where the published version shouldn't reveal faces, blur them first with face-blur, then split the redacted file.
LMS rejects the file even after splitting
Check capIf an episode still exceeds the LMS per-file limit, you used too few parts for the size. Recompute ceil(file MB ÷ cap MB) and use the larger of the size-based and attention-based counts. High-bitrate captures may also benefit from compression before splitting.
Need more than 20 episodes
ClampedThe count caps at 20. A very long lecture series exceeding 20 episodes can be split into 20 then a part re-split, but consider whether 20+ tiny episodes actually help students versus a sensible 15–20.
Slides re-encode soft if you transcode
AvoidedBecause the split is stream-copy, dense slide text and equations stay pixel-perfect. Only transcode (with video-transcoder) if the LMS can't play your codec — and prefer a high-quality setting to keep text legible.
Episode starts with a negative timestamp in a player
Preserved-avoid_negative_ts make_zero rebases each episode's timestamps to zero, so LMS players and embedded viewers show clean starts and keep audio in sync rather than displaying odd negative offsets.
Variable frame rate from screen-recorded slides
PreservedScreen-recorded lectures are often VFR. Stream-copy preserves the timing, so episode durations reflect the real timeline. Boundaries still follow keyframes regardless of frame-rate variation.
Recording exceeds your tier's file cap
RejectedFree caps a single file at 1 GB; a multi-hour HD seminar often needs Pro (10 GB) or higher. The limit is on file bytes, not on lecture duration.
Source duration can't be read
ErrorIf the capture was stopped abruptly and lacks duration metadata, the tool raises 'Could not determine video duration.' Remux the file (or re-export from the recording software) so the container reports a valid duration, then split.
Frequently asked questions
How do I cut a long lecture into episodes?
Decide a target episode length (15–25 minutes works for most lectures), convert it to a part count (total minutes ÷ episode minutes), and split. The tool divides the recording into that many equal parts losslessly, and you upload each as a separate LMS item.
How long should lecture episodes be?
15–25 minutes generally matches student attention spans and natural topic chunks. Shorter (more parts) raises completion rates; longer (fewer parts) keeps related material together. Choose the part count that lands episodes in your preferred range.
Will splitting blur my slides or equations?
No. The split uses stream-copy (-c copy) with no re-encode, so dense slide text and equations stay pixel-perfect in every episode. Re-encoding is what softens fine text, and this tool doesn't re-encode.
Can it cut at slide or topic changes?
No — it's an equal-time split, so boundaries land near even marks on keyframes, not on slide transitions. To align cuts to topics, find the change points (try scene-detector for slide cuts) and trim there with lossless-trimmer.
Is the recording uploaded to a server?
No. The split runs entirely in your browser via FFmpeg.wasm, so recordings showing student faces or unreleased course material never reach a third party — important for FERPA/GDPR compliance. It only reaches your LMS when you upload it.
How do I make episodes fit my LMS upload limit?
Compute ceil(file MB ÷ LMS cap MB) for the minimum part count, then use the larger of that and your attention-span count. Each resulting episode is a fraction of the original size. If episodes are still too big, compress before splitting.
How many episodes can I create?
Between 2 and 20. For most lecture series that's plenty; if you genuinely need more than 20, split into 20 then re-split a part, but longer episodes are usually better pedagogy than 20+ fragments.
What format are the episodes?
The same container as your recording — an MP4 export from Zoom or Panopto produces MP4 episodes. If your LMS can't play the codec, convert with video-transcoder, preferring a high-quality setting to keep text legible.
How do students follow the order?
Episodes are named name-part-1 … name-part-N in sequence. Upload them in order and title them by topic ('Episode 4 — Recursion') so students can navigate the module logically.
Can I hide student faces before publishing?
The splitter doesn't redact, but you can blur faces first with face-blur, then split the redacted recording. Both run in your browser, so the original never leaves your machine.
How big a lecture recording can I split?
Free allows one file up to 1 GB; Pro up to 10 GB; Pro + Media / Developer up to 100 GB. A multi-hour HD seminar typically needs Pro or higher. There is no duration cap — only file size.
Will every episode be exactly the same length?
They target equal length, but each cut snaps to the nearest keyframe, so durations vary by up to one keyframe interval. If an episode must begin at an exact point, trim that boundary with lossless-trimmer.
Privacy first
Every JAD Video tool runs entirely in your browser via WebCodecs and FFmpeg (WebAssembly). Your video files never leave your device — verified by zero outbound network requests during processing.