As I've talked about a lot recently, we've been providing our software development and programming training as online workshops. It's been our plan all along to try to record these session to both have a corpus of material to help people we are consulting for but also for those times we are running a session and someone can't attend. We wanted to do this even when we weren't running these online but the classroom style we used didn't really lend itself to recording as: a) I tend to move around the room a lot and b) we would easily miss my answers to students' questions.
Last Friday (the 3rd of April) we ran our Introduction to Data Analysis in Python workshop (which happens to be the first course I wrote when I joined Bristol). I recorded the session and have spent some time editing the video to make it watchable on YouTube.
I'm quite new to making videos so I'm still learning. I have in the past made a few development videos for a voxel engine I was working on with my brother but they had minimal editing done to them. This time, however, I wanted to spend the time to make it maximally useful to the viewer so that they could actually be used to learn the topics from.
Here's the video in its final form:
To create the video I used Open Broadcaster Software to record my screen (the same screen that was being shared on Microsoft Teams) from about 10 minutes before the session started until after it had finished. This created a ~400 MB video file spanning the three hours which it was then my task to reduce it down.
Of the three hours, I spoke for about two of them and the rest of the time was taken up with the students working on the exercises. I took the time to remove any long quiet sections like this so that it would work as a teaching tool after the fact.
Another facet of the online teaching which has evolved as we have done more of it is the teacher answering peoples questions "in front of" the whole class, rather than sending them off into a breakout room to get help. Since we had these recorded and we've often found that if one student has a question then probably more are wondering the same thing, I wanted to keep these in the video. As I wasn't always very good in the session at repeating people's questions "for the video" I added annotations in the final video whenever I was answering a question which was not covered in the core material.
For the job of doing the video editing, I used Kdenlive which is a free and open-source video editor which started on the KDE platform but is available on all the major operating systems. It had all the features I needed like title cards, blurring, cross-fading and track slicing.
All-in-all, the editing of the video took me about 4 hours and the rendering took about 3 hours. In the future I may do less editing and annotating to cut this time down and/or create shorter topical videos which will reduce the render time per video. I was loathed to spend longer editing this video into chunks as I'm planning rewriting the flow of the course this year to take a more top-down approach.