Many universities have begun exploring the use of the Internet for sharing academic coursework, either via “flipped classrooms” or Massive Open On-Line Courses (MOOCs). Over the last year, I have uploaded approximately 50 videos to my YouTube Channel, most of them academic lectures. I hope that I have learned something in this process that will you to publish your work more broadly, as well!
I would start by explaining that my lectures come from multiple purposes and even multiple university campuses. My longest-running series of lectures came from a weekly seminar on topics of my own choosing called “the Useful Hour.” I produced fourteen of these sessions (with help from Brigitte Glanzmann when I had to be away for a week), though I only started recording them on video for the last twelve. I recorded the eight-session bioinformatics module from our division’s B.Sc. Honours program as a trial run for creating a “flipped classroom” in future years (a model where students watch lectures outside of class and spend in-class time working exercises). More recently, I collaborated with the H3Africa BioNet to produce a four-lecture module on Gene Expression. From time to time, I help the Tygerberg Postgraduate Student Council by recording a lecturer. Each of these experiences has had its own lessons to convey.
The technical aspects of recording a video are generally easy enough that even a Ph.D. can do it! Today’s budget camcorders capture more detail with better sound under lousier conditions than did cameras that cost five times as much even five years ago. Best of all, one no longer needs to wrestle with tapes and analog-to-digital transfer loss. Today we simply pull the Secure Digital card out of the camcorder and plug it into the socket on a laptop, where the video files are instantly accessible. Of course, many people record video using digital cameras or cell phones. Preparing videos for upload to a public server, however, is frequently more difficult than the initial capture. I’ll talk about these aspects below.
Focus on the speaker
We must start with video that is worth watching. Far too frequently, I see that people recording lectures focus on the slides rather than the person who is delivering the lecture. Reading text from video is generally unpleasant, and the reality is that looking at people fires circuits in our brains that academic content does not. Video is a format designed to capture motion; it is a notoriously inefficient method for capturing still images, though! Keeping the camera on the speaker, then, makes more sense. This comes with some caveats:
- Viewers still need to be able to see the slides. My answer has been to produce a PDF from the PowerPoint or other presentation software, since almost everyone has the ability to view PDFs on any platform. I post the PDF to a shared directory on Google Drive, and I include the URL leading to the PDF in the YouTube description.
- From time to time a researcher will point to a particular part of a slide. This is probably problematic on video; if he or she has used a laser pointer, the spot of light will either be too bright (green) or too dim (red) to appear well on video. A moving mouse pointer might be better. If the speaker is old-school (like me), he or she may use a stick to point at the slide instead. This can create a problem of the lecturer “blooming” as he or she moves away from the bright field created by the projector into the relative dark outside the projector’s light.
- How will a person watching the video know to advance to the next slide? Hopefully the speaker says “next slide” out loud. When my parents recorded my brother’s and my first efforts to read aloud, they told us to bang a spoon against a mug to produce an audible chime with each page turn. That was even more fun than reading!
- Software is publicly available to integrate the slowly-changing slide video with the quickly-moving speaker video. Screencast-O-Matic will produce videos of up to fifteen minutes in its free version. This approach will guarantee that your viewers are seeing the same slide the lecturer is seeing as the talk progresses.
Light and detail go hand in hand
As I alluded above, lighting is frequently a problem in academic lecture videos. We frequently keep our lecture halls very dim in order to make the slides stand out as much as possible. In a large venue, you may have a spotlight on the speaker, which will help. In a medium venue, you may have a light in the ceiling directly above the speaker, which can make him or her appear somewhat ghoulish. The more you rely upon zoom, the less light will reach your camera! Keep that camera close. If you can open the blinds on a window so that your speaker is lit, you will have a more interesting video. Try to find ways to position your camera between the light and the subject (without casting a shadow, of course). Never forget that the projected slides are much brighter than the subject you are trying to record. If even the corner of the projected image appears in-shot, expect the speaker to become a flat silhouette.
Today’s cameras can record in very high resolutions, such as 1080p (the same as your HD television). If lighting is truly problematic, you may want to consider forcing your high-resolution camera to a lower resolution, such as 720p; this may allow it to combine intensities across multiple transistors for each pixel. Similarly, you should expect that a camera with a larger “retina” will outperform one with a tiny CCD in low light. To put this in plain terms, do not expect a cell phone to produce quality video in semi-darkness, no matter the name on the label. That said, I have observed that my “mirrorless” Canon EOS-M2 is inferior to my much cheaper Canon VIXIA HF R62 for video. The lenses and electronics of the EOS-M2 are optimized for photos, not video.
Privacy issues are a big deal
Ensure that your audience knows that the lecture is being recorded. Bad things can happen when a person does not want his or her image to be on-line and somebody else decides that they shall be. Imagine how much worse this becomes when that member of the audience is a minor! Nobody should be forced into public view because he or she attends a talk.
We frequently expect a period of questions and answers at the end of a lecture (and sometimes in the middle). A novice camera operator may automatically swing to capture the questioner in action. Depending on the situation, this part of the video may need to be truncated outright due to privacy issues.
Video is big and hard to handle well
When I upgraded to my Canon VIXIA HF R62 from a JVC Everio (GZ-HM30AU), I had a rude shock. My old camera had captured 720p video in very manageable MTS files, but the new camera captured 1080p video in massive MP4 volumes. I used a 16 GB SDHC card for videos. The cameras assumed that no file should be allowed to be larger than 4 GB (linked to 32-bit computing). With the new camera, I consume 4 GB every 33 minutes! At a couple of long events I recorded, I found that I needed more storage than the 16 GB card could provide. I solved that problem by upgrading to a 64 GB card.
Naturally, keeping the raw footage of every event I video is not practical. If each of the 50 videos I posted to YouTube over the last year produced 66 minutes of raw footage, I would need to archive 400 GB for just this period! Similarly, posting these videos to YouTube would be a problem. Each hour would span two files, which would require my viewers to watch part ‘A’ and then queue up part ‘B’ immediately afterwards; many would just skip watching the end, humans being humans. To compound the problem, I live in South Africa, which means my upload speeds to network servers are dreadfully slow. My home DSL line, for example, achieves 0.3 Mbps. I have uploaded one GB before, but it takes hours. In any case, I will probably need to truncate a bit of time off the front and the back of the video. In short, I need to do video editing.
While semi-professionals might opt for Adobe Premiere and those who “think different” will break out iMovie, I am a bioinformaticist, and I like software that lets me master high-quality videos with a minimum of fuss and bother. I use ffmpeg, a very powerful suite of tools that one can use directly on the command line. Most of the time, I am (a) concatenating my source video files into one movie, (b) including only a middle section, and (c) writing a more compact movie from the source materials. To use a recent example, I have two input files; I write their names into a file called list.txt:
file mvi_0031.mp4 file mvi_0032.mp4
Next, I run a command line that looks like this:
ffmpeg.exe -ss 00:00:15 -f concat -safe 0 -i list.txt -t 00:50:00 -c:v libx264 -preset slow -c:a copy output.mp4
In order, the options do the following:
- -ss specifies where in the combined files ffmpeg will start the output video (in this example, after the first fifteen seconds).
- -ff concat -safe 0 -i list.txt specifies that the files listed in list.txt should be combined into one video and that they are formatted the same way.
- -t specifies the total duration of the video to be encoded (in this example, exactly fifty minutes).
- -c:v libx264 -preset slow specifies that my output video will be MPEG 4 pt 10, a very common format for storing video (and one that YouTube knows how to read).
- -c:a copy directs ffmpeg not to re-compress the audio, making it sound just as nice in the output as it did in the original.
The ffmpeg software is very good at reducing the size of videos without compromising its quality. I find that I can represent an hour-long lecture in a two GB 1080p video, rather than the nearly 8 GB of source footage. If I am filled with caffeine for my lecture, the video size increases a bit (more motion requires more bits for accurate representation).
These smaller videos can then be uploaded to my YouTube account. Happily, if you have a Gmail account (or if you use a different email address to log into Google Services), you can simply use that login for YouTube. One clicks the arrow pointing up, and a screen will appear to which you drag your video file. All done, right?
No job is finished until the paperwork is through!
Meta-data is key to your video reaching an audience, and too few people spend adequate time on this step. I would call your attention to both the “Basic Info” and “Advanced Settings” pages that video authors can complete. Of course, you should enter a paragraph of information in the basic description blank. Ask yourself what web searches should find your video, and be sure you include those key terms in the text. For good measure, add them again in the keywords section! I like to include the university name where the recording took place. Hopefully the social media minders for these schools will highlight your video to their large audiences. YouTube will sniff the video for still frames that might be representative for the video. I always try to pick the one in which I do not look like I’m suffering a fit of some sort.
Advanced Settings has more options to help users find your video. Pick a category; generally my lectures fall in the “Science and Technology” category. Be sure to enter a video location. Google will translate your information to GPS coordinates so people can find videos shot near particular locations. Enter a recording date, and select the language of your video (especially if you are not using English).
In many cases, you will have several videos that belong together as a set. When I produced a short biography and four videos on Gene Expression for H3A BioNet, I also created a “playlist” that contained all five videos in the correct order. Remember, if you can hook a viewer into watching one of your videos, you might be able to retain their interest for a few more! Ideally, people will like your stuff enough that they subscribe to your YouTube channel, receiving a notification every time you post a new video. You will be launched on your next career as a YouTube star!