In January of 2010, I was asked to film and deliver edited versions of all the presentations of a three-day conference entitled “the Cassini-Huygens Project: Huygens legacy and future Titan exploration” held in Barcelona. I accepted the challenge and started planning.
I had not filmed presentations before myself, but I had seen filmed presentations on the internet and in general I was not very impressed. I noticed that often the sound was not great, the speaker was filmed from far away and/or together with the big screen where the powerpoint slides were projected and, when filmed with multiple camera’s, I was in general not too impressed with the editing.
With the Barcelona conference I saw a chance to take a different approach that I’d hope would give a more satisfying result. I decided that the filming and editing of these presentations had to be much closer to filmmaking and less to a mechanical recording of an event. I placed myself in the viewer’s seat with the thought: “How to duplicate the sensation of sitting in this meeting room, listening and watching the speaker deliver her/his talk?”. Since I am also a scientist, I know exactly what it is to be in a meeting and all I had to do was to find a way to translate that into video in the best possible way.
Below I describe the concept I came up with and that I have fine tuned over the past few years. I have recorded and edited almost 60 presentations to date. At the end of the article you will find links to some of my work.
The sound has to be as good as possible and there is just no excuse. If the sound is no good, the viewer will simply not connect to the speaker, gets bored after a short time and looses interest. Many meeting facilities do have excellent sound systems which can be tapped straight into the camera with the help of the local technician. If the signal comes from the mixing table, then it is likely to include the signal of all microphones in the room, including the ones that are passed around for questions of the audience. Yet it is always best to have one’s own good radio-mic set at hand. It is worth the investment!
The camera sits on the first row. This is the best place, because only from there can one get a good image of the speaker, using the zoom capabilities of the camera, ranging from Close Up to Full Shot. In general, I do not care filming what happens on the screen where the supporting material is projected, as I add this later during the editing stage. The speaker, and in particular her/his eyes, needs to be seen clearly in order for the viewer to really connect, keeping in mind that the size of internet videos in general is small. In most cases the speaker will be constantly turning from looking at the screen, pointing at the slides, to looking at the audience. I therefore always place the camera on the first row and somewhat towards the side of the podium such that as the speaker turns from the screen to the audience and back, she/he is always turning into the camera. During filming I constantly track the speaker and keep the image manually in focus (which can be quite challenging with speakers who move about a lot). As the presentation goes on, I do listen to the content and keep track of what happens on the screen visually, so that I can decide re-adjustments of the framing when I am sure that in the edit I will be showing the slide and not the speaker. I find it important to keep listening to the speaker at all times: when the speaker makes an important statement at some point, I will want to chance to a Close Shot framing of the speaker, which tells the viewer that this is important and needs to be listened to carefully.
Here is where most of the work comes in. To start with I almost never show the speaker and the screen at the same time. I believe this adds nothing: when I am sitting in the audience attending the lecture, I also either watch the slide or the speaker, never both at the same time. I obtain the slideshow (and other material she/he might show, such as animations) from the speaker before or after the talk and use these slides as still images in the edit. I always cut hard between the image of the speaker and a slide, that is, I do not use cross dissolves for these edits. In my view this is the most natural way and mimics best what we actually see ourselves when we move our eyes from the speaker to the screen and back. Finding the right moments to cut is sometimes tricky and it does at times take some effort to find the most natural cut. Key moments to look out for are when the speaker looks up at the screen, or when does she/he clicks on the button to change a slide.
I avoid showing the slide and the speaker at the same time. However, having just the slide standing and hearing the speaker can quickly become boring. In real life the speaker will be literally pointing at the slide with a pointer to emphasize what she/he is saying and indicate any important elements on the slide supporting that. In the edit, I solved this with a simple trick: highlighting. In the figure below an example is shown.
You can download additional instructions here, where in the last two pages I explain how to do this in Final Cut Pro 7. But any good editing software will have a way to do something similar. Note that I do keep the slide within the Title Save area, in case a DVD needs to be made.
As the speaker advances, I highlight words, phrases, formulas, graphs, etc. that are important at that moment, but without loosing the rest of the slide. The viewer’s eyes will be directed to the highlighted area, similar to what happens when the speaker points with a laser pointer. I like to fade in / fade out these highlights, which I find is less distracting.
Sometimes slides can be really too full, or a graph can be very complex or small on the slide. If the speaker discusses a detail about a such graph, I simply take that graph separately, retrieving it from the slide show, and show it over the slide in whole or zoomed-in. I sometimes do take the liberty to redesign the slide somewhat, to better fit in the video.
The goal is always to be as sound and as clear as possible. In all cases I try to edit as natural as possible, keeping in mind at all times that I want the viewer to feel “in the room” with the speaker.
Verification and Result
Once I have a first edit, I send it to the speaker for verification to make sure I did not commit any errors in the highlighting or elsewhere. As soon as the verification is done and final adjustments are made, I produce the final video for online publication.
An advantage of this way of producing it that the final result is a high-end result. A disadvantage is that this method is rather time-consuming, specifically the editing part: on average it takes me about eight times the length of the presentation, so for a one hour presentation, I’ll be editing for a day. Most of that time goes in the highlighting and its adjustments. I am not aware this can be sped up by an automated process, since the editor is reacting to the spoken and written content, which, it seems to me, requires a human being to make it exact.
In an academic teaching setting, where many lectures are recorded for later consultation and reviewing, one can ask the following: how much does the content of a lecture change from year to year? For most lectures, probably not much. Once recorded and edited in the way described here, the result will be useful and watchable for quite some time.
(this article was published in the Media and Learning New Letter, 14 February 2013)