Video editing OLD


This is old information, these days we use ffmpeg-editlist and ensure that no learners are in the videos in the first place.

See also

Video checking OLD for how to check a video and give an edit-list to someone else to do the editing.

The purpose of this page is to give video processing volunteers a starting point. (It also has some hints for workshop organizers).

For some of our online lessons, we release videos on YouTube. This is not necessarily for brand new people to watch and learn the material (though they may), but especially for people who attended the workshop to review what they saw. As such, it’s more important to get them published fast, than make them perfect.

What we want

Our workshops consist of lectures, demo, and exercises in breakout rooms. We record the main Zoom room, and also livestream the main room via Twitch. We would like the video of the workshop to be processed so that it can be released on YouTube. This should not be a major production: it is more useful to those who want to review what they saw in person, rather than a new person watching.

We will provide the following:

  • Raw video files (probably two copies one recorded from Zoom and one from Twitch - so there is a backup.)

  • List of lessons (= final videos) and which raw files contain them and when.

  • List of instructors

We want out:

  • One processed video file per lesson.

  • With irrelevant breaks removed.

  • Without any video from learners. We use Zoom so that learners should not appear in the stream, but we can’t be sure it works so this needs to be checked.

The rough process is:

  • Load up the right video files in the editor.

  • Find the start of the lesson (hint: look for a change of instructor - ask us if you need help!) and cut off the stuff before.

  • Watch through the videos. Most of the lecture parts are fairly standard and can be fast-forwarded through (it’s rare for a learner’s picture to appear here).

  • Cut out the idle time during breaks.

  • In exercise sessions, learners go to breakout rooms, which are not recorded. This part can be cut out, but sometimes the instructor stays in the main room to do the exercises for the stream.

  • Don’t be too precise. We aren’t trying to make a masterpiece to end all masterpieces, but a something for those who were at the workshop to refer back to. So:

    • Imprecise start/stop/break times are fine

    • Other random off-topic chat is fine

    • Voices of learners is fine and expected

    • Video of learners is not ok (really, this is the only thing that needs care).

Before/during/immediately after the workshop

  • From day 1, advertise that “the workshop may be recorded and put on YouTube. We will prevent any pictures and names from going there, but your voice may be. Please don’t include your name in hackmd unless you accept it may be published. We support your right to be anonymous in this workshop.”

  • Same announcement at the start of the workshop.

  • Record in zoom. Note: when you start the recording, make sure that someone is currently sharing the screen, and the screen is a good size (e.g. normal Full HD, as opposed to some vertical shape). The dimensions when the sharing first starts determines the dimensions for the entire course.

  • Immediately after workshop, go to Twitch and download the raw streamed version. You have to be logged in as the channel, then the option is naturally provided to you.

  • Choose some standard, shared place and immediately upload videos there. Recommended naming scheme:



Processing principles:

  • Remove any participant videos, if they accidentally make it into the video file. This is really the only serious rule in the processing, if we didn’t have to check this we could just upload the raw ones and it would be good enough.

  • Create one final video per lesson in the workshop

  • Work incrementally, upload processed ones when you can, get quick feedback.

If it’s not clear, the course organizers will provide a list of the lessons (final outputs) and the respective inputs (which source files go into it).

You can generally:

  • Use some video editor

    • iMovie on Mac

    • OpenShot is a simple cross-platform editor (tutorial)

    • (please give more ideas here)

  • (so far, this is not a general “how to edit video” guide… you will need to find one for your editing program)

  • Create a new project for the output (e.g. the Jupyter lesson)

  • Import the raw video files which contain Jupyter (e.g. day 4). If one lesson is split over multiple files, combine them.

  • Cut off the part before and after the lesson itself (saving frequently). You’ll have to figure out the start and end times, this may be hard when there are several files.

  • Begin watching the lesson. Look for the following things:

    • Break time? Exercise session with irrelevant stuff in the video? Cut the time out.

    • Any non-instructors pictures in the stream? Cut it out. Sometimes you might need to blank the picture while

    • Don’t be too perfectionist - the goal is to get something done, not maek the perfect videos.

  • Export the videos with a high quality, e.g. jupyter.mp4. It will go to YouTube which will render lower resolutions, so you don’t need to worry about this so much.

  • Upload the videos to the processed subdirectory of the google drive. Do this immediately, video by video. It’s better to get continuous feedback on this. You are done!


We upload them to YouTube (not that we agree with all the ethics of YouTube, but it seems like the least bad and most useful of the options).

  • Preview the processed videos, do a quick check for any issues.

  • Upload to the channel. For one workshop, put all related videos into a playlist. CC-BY license.

  • This is a prototype channel description you can copy:

    git-intro 1/2, CodeRefinery 25.may-4.jun 2020 day 1
    Day 1: git-intro: LINK-TO-LESSON
    Part of a series of video recordings of the CodeRefinery workshop,
    25.may-4.june.  CodeRefinery teaches intermediate software
    development skills to researchers in the Nordics.
    Workshop page: LINK-TO-WORKSHOP
    Q&A for day 1: LINK-TO-HACKMD
    (table of contents below)
  • Create a table of contents (can be done later, after uploading). This divides the videos into chapters, with clickable links in the description and labels in the video’s time slider. In the bottom of the description, put this text and it is automatically parsed:

    00:15 Introduction
    02:30 Motivation -
    17:17 Basics -
    38:29 Staging -

    You may want to ask someone for help with this, since it can take some time to go through the videos.

    Example with table of contents: