Technical epilogue: Video production for, ‘Current eye-tracking research on pedagogy’
I’ve given talks before. Let’s just say that.
This week, I am due to contribute as Discussant at a symposium at AERA 2022. As I will not be attending the conference in-person, I have contributed a video recording of my commentary.
The crafting process was intricate. The entire process felt qualitatively different from my norm for preparing such things, including particular challenges that come with home ‘studio’ recordings via webcam-baked footage.
I thought I’d share what I did, how it turned out, and what I might do differently in future. There are the things, too, that were not ideal or most effective 一 but that I think I may keep anyway, in future, because few feasible alternatives are available that I know of.
Fully scripted
Normally, my preparations would be to produce a purely verbal ー and unscripted ー commentary. I’d rehearse that ahead of time, until it becomes inner speech. This involves no technical, or even written, preparation. Not especially in contrast to what I’ve done this time.
Rather than starting and ending with the PowerPoint slides, I started by writing a full script within PowerPoint Notes this time. Only after I had drafted the script for each slide did I populate the slide itself with succinct numbered points.
The script would then appear as Presenter Notes during the Presentation. Each numbered point within the slide would correspond with the numbered point in the Notes.
I wrote the script mindfully as such: as narrative to be spoken rather than written, or read. I first wrote as a draft, expecting to edit, trim down, and refine at the next stage.
Lecture capturing
I recorded my talk using Panopto, capturing a headshot via the integrated webcam, plus only the second screen. I used Snap Camera for my virtual background with one of the university’s corporate backgrounds: this was necessary because Panopto (still) does not enable virtual backgrounds natively.
I set out to record my talk as if I was giving a live presentation online. Eventually, I learned to force through a recording of the full presentation, re-doing parts while the camera was ‘still rolling’. The hesitations and the mistakes would have to be edited out.
Throughout, I kept my eye on the clock in PowerPoint’s Presenter View, which I used as a guide as to how long I have left to speak. This served to pace me more than anything: the repeated subsections that I, in a sense, left in meant that the actual recording was longer than it would finally be.
I had my right hand on the mouse for button pressing: I started the video recording and moved between windows this way. I had my left hand on a wireless pointer for slide changes.
Video curation
I edited the talk within Panopto, removing hesitations and mistakes as (finally) planned.
Trouble-shooting slides
Some slides contained wrong content that I only spotted at the stage of video curation. Or, content that I could not ultimately keep in the video due to time.
So, I uploaded the PowerPoint slides and mapped these onto the right time points in the video commentary.
At my time of Panopto use, I was unable to select the slides as ‘secondary stream’ instead of the originally recorded second screen. So, I downloaded the video of my headshot, uploaded this onto a new video project in Panopto, then re-mapped the PowerPoint slides onto the relevant timepoints of my commentary in the new video project.
Video output
I chose the side-by-side video output style, since the picture-in-picture resulted in slide content behind obscured at times.
Captions
I found some of the audio recording to have dipped out, probably due to the audio enhancement tool on my laptop (Dell Precision). So, I set out to compensate communicative clarity by making use of the closed captions, which Panopto can automatically generate. I then edited the captions, correcting where automated capture was misled.
Moreover, I opportunistically planted words in that I should have used but had not actually spoken. Thus, I was hopeful that the spoken and written words would, together, convey the optimal commentary.
The captions did not burn into the output video by default, nor was this possible by preference within Panopto. So, I used HandBrake.io to burn in the captions that I had carefully edited in Panopto. To do this, I downloaded the caption file (not the full unedited file) from Panopto, then converted this .txt file to .srt. I then uploaded the .srt file to HandBrake (full tutorial here).
Takeaways
The final outcome was professional enough. I think it certainly pointed to a thoughtfulness during development, with priority given to clear communication of the content. Technical courage all served to support this goal, which I think will be noted.
Technically, though.
I would like to find a video editor that is more versatile than Panopto in future: the slide corrections, the process of overriding the original second stream with these, and the caption burn-in all felt unduly protracted.
Snap Camera gave jumpy edges to my headshot, which undermined the professional demeanour a little: before the audience registers a technical fault, they may have formed a first impression that I am overly informal and friendly given the nature of the event.
Finally, I really should have implemented the editing phase of the script writing process: without this, my speech-based writing was not speech-based enough by the time I started recording. And I should have rehearsed my talk before attempting to record. By starting to record prematurely, I lost the time that I thought I was saving 一 and spent it in a much less tranquil way. There really is no shortcut to a smooth recording experience.
Originally published at http://drnoramcintyre.wordpress.com on April 18, 2022.
