I can now see the top of the mountain I am climbing towards. My first steps in that direction occurred with the start of principal photography on The Unruly Mystic: John Muir in May 2015 and it took 18 months to film all my interviews, traveling around the US and Canada. Now after three months of editing, I can see the outline of the film emerge above the tree line.
- As posted to Facebook.
There were multiple steps of preparation taken in order to make this edit milestone, so I want to use this post to explain my process and approach to encourage others in their own film-making as well as document my specific efforts in more detail. I am sure there are other articles and books on this process, but this is what worked for me.
I am only going to focus on the steps that occur after filming has been completed, so this post is more about the editing process rather than production. It also requires a good understanding of how Adobe Premiere already works, and might make you think about it more as a database tool then purely a visual means to edit video.
There are other articles that do a nice job of explaining how best to capture great picture and sound with a two-camera setup. That said, it is always important when shooting for the edit, to conduct the interview repeating questions to all your subjects for a variety of response.
Defining Your Topic of Interest
For instance, I asked everyone to share a favorite story of John Muir, and suggested one in particular of John Muir climbing a tree during a thunderstorm, as an example. The end result is that I can retell that story in the film through different voices with how I edit it together. That approach can also work with other lines of questioning, so you need to know what the focus of your film is going to be, before you start filming. So created a hand-written document outlining the topics that I thought would be relevant throughout the filming and editing process.
Since I work in Adobe Premiere Pro (PrPro), I know from working on my previous documentary, that I don’t want to be editing these interview sequences with their original MXF* files. I had several instances where the original file got corrupted or “lost” which ruin an entire sequence when it became time to export it. This step also allowed me to safely keep the “original” files off-site as I was working with a facsimile of my original footage. That facsimile was actually a new MXF by exporting a new sequence after I imported all the original files. *MXF is the output 1.9 GB video file created when shooting on the Canon XF 100.
Prepping the footage
I re-import these new facsimile MXF files, create a multi-camera edit, and then add a time-code to the sequence. It isn’t important to worry about cutting back and forth between the cameras at this point of the edit. You just need one camera angle (preferable is frontal) and the best audio recording for the next step of exporting again but in a lower resolution format that you can upload online with burnt-in time code as a H.264 file.
That burnt-in time code H.264 video becomes the resource for the next step which I didn’t have the luxury of referring to in my previous documentary, creating a transcription of the interview in a Word Doc based upon the time code. There are multiple services that can do this aspect of the work for you, and I was happy with the person that I found available to me. While it can cost $100 for one hour of footage, billed by the minute, the benefits as you will see are well-worth that cost in how many different ways you will use that transcription.
I created the following process that worked for me. Once I had the transcription back, I would go through the document with several different color markers. I had previously written out my topics and assigned them colors.
For instance, using the idea of hearing John Muir’s story retold of climbing the tree, I would highlight that section. There would be multiple other themes that I might be following, and I would do the same for those two.
Marking the Interview Sequence
After I had finished going through the 15 page document, I would open the multi-cam sequence in PrPro and put a marker at each beginning point for that topic in the timeline.
I would also copy and paste from the Word Doc, that specific transcription sentence or paragraph into the marker’s description itself. I would also color code the marker to match the highlighter for future reference back on the sequence. Then I would put a cut at the marker in-point and another at the end of the edit. That clip would be then nested with the same descriptive of the transcription.
So the Matt Fox timeline looked something like the following:
Descriptive nested sequence and themed topic bin
After I have a descriptive nested sequence, I could put it into a bin with similarly themed topics. Ideally you want a manageable number of bins to be putting these nested sequences into. Once in their bin, I made sure that all the nested sequences had a uniform color i.e. green (forest) so that I could know if I had already used the clip or not after I built a new sequence out of them for a rough cut. Once I used the clip in the new sequence, I would reveal it back in the project folder from the edited sequence with a short cut key I made (Alt+D), I would change the color to red (rose). While PrPro can reveal clips that are being reused in a sequence, it doesn’t do that for nested sequences.
While I didn’t do this task of creating descriptive nested sequences put into specific categorical bins until I had completed all my interviews, you could do this after each interview, but I think it is important to be open to the entire spectrum of the interview footage and refresh your memory by rereading the transcriptions and actively highlighting the best dialog as it relates to your chosen topics. This is especially important if you are interviewing a lot of people over a long period of time. If the interviews are short, there are few of them and take place in a short amount of time, you can start earlier. I had an hour of footage for each of the 15 interviews over 18 months.
Now that I have all my nested descriptive sequences into their respective bins, I can easily select those sequences and make a new sequence from them for editing. Because I like to know who is speaking (thumbnails are not available with nested sequences), I would go through all the clips and assign a color to each speaker and I could also randomly place the clips so that no two clips followed in the same color.
Because the topic was similar in each of the clips, there was no harm in doing this step as I could easily move the clips around when I started editing. Sometimes of the randomness of this action would lead to new discoveries in the editing process.
For example, I noticed that each bin I had consisted of 10 minutes of material that came from 20 clips, so by thinking of each sequence as its own movie, I could easily watch the entire sequence, move the clips around to make the flow better and clean up the dialog in and outs back at the original nested sequence. Ideally you should be able to do this task directly in the edited sequence but I didn’t always have access to the visual cue of the wave form on that sequence without re-rendering it again.
With a new sequence created based upon this process, I could now start the edit on that themed topic bin, knowing that I had all the relevant descriptive nested clips in one bin, I could always find where I took that clip from in the original interview by reviewing the descriptive markers on the sequence, I had a means to note what clip was used or not, and I could double click any nested clip to open the descriptive nested sequence where I could access each individual multi-cam edit which proves helpful in cutting back and forth between two cameras to improve dialog edits after I decided on my final edit.
Now the magic of editing could happen!
POST NOTE: I am reminded of how creative people sometimes create systems and processes in order to do their actual creative work. In 1991, I was fortunate to be working at Paramount Studios as secretarial temp (basically I had computer skills, and the older staff was still pounding on typewriters), and found my self work as a director’s assistant to Bruce Robinson, who was in the final stages of his script entitled JENNIFER EIGHT. The movie had yet to be green-lighted by the studio, so he was dealing with required changes and rewrites to the script. Since he was one of those guys that pounded the script out on a typewriter, I was hired because I knew how to use the computer and script-writing software like Final Draft that had just become available the year before. In case you don’t know Bruce’s amazing writing, he is arguably most famous for writing and directing the cult classic Withnail and I (1987), a film with comic and tragic elements set in London in the 1960s, which drew on his experiences as “a chronic alcoholic and resting actor, living in squalor”.
Bruce’s personalized format of the script is what challenged me when copying his work into the newer format. Besides justifying the left margin with a simple carriage return, he would also justify the right margin with misspellings, extra spacing and anything else to have a clean line down that side. Later he told me that he liked to repair antique Swiss watches for fun. There was no apparent need to justify the right margin of the script, but he told me that it helped him to focus on the creative tasks better. Perhaps my own system and process that I outlined above is similar in that regards. If so, then find what works for you personally and let that become your muse.