A common idea in modern journalism is that for a documentary to be “fair,” its creator must have no biases about the subject in question. Even if this were remotely possible from an ideological standpoint, the filmmaker would still have to know nothing and care nothing about the subject or message of his film until its completion. It is a ridiculous concept.
As many readers know, the Botkin family is extremely pro-home-education, and has been since the early ‘80s, and so when we began work on Homeschool Dropouts in August, we were already familiar with the history and current state of home-education movement.
We had already discussed many of the issues that we wanted to focus on in the film with several homeschool graduates and homeschool leaders, so choosing our interviews was very straightforward. Since we were also familiar with the positions that our interviewees would be representing, we had a pretty clear outline for the script before we began shooting.
Act one makes the case that a problem exists within the home education community – the second generation is not fully supporting their parents’ vision – and provides historical context for this development. The second act describes this problem and its causes in more depth, and the third act explains how and why the problem must be solved. Simple, obvious, and hopefully a clear and concise way to present the material.
After shooting all of our interviews, we transcribed the strongest segments (Adobe Premiere CS4’s new auto transcription feature almost worked, but not quite), and began to organize the clips. Despite already having a clear direction for the film, the material we got from the interviews suggested and provided many new points, and so the script was reordered and adjusted.
Documentary filmmaking blurs the lines between production and pre-production, since lots of the rewriting can occur during and after most of the shooting. Once we had selected the best material from the almost seven hours of interview footage, we began writing interstitial segments to tie everything together, and filmed all the Botkin children delivering his or her lines. In future, we may try to have fewer hosts.
Since the film was basically edited at the script level, it was a simple matter to just drop the appropriately trimmed files into the Premiere timeline. The only complicating factor was maintaining sync between the A and B cameras and the externally recorded audio of the interviews. The standups, shot on a single camera with onboard audio, was easier to manage.
A lot of our outdoor shoots had background noise that we needed to remove. Rather than clean and tweak the finished audio of the final edit, we decided to process all the audio we had captured. This way, we would be editing with the final audio, and when we locked the edit, the bulk of our audio mastering would already be done.
I set up a number of scripts in Adobe Audition and we basically batch-processed everything overnight – noise reduction, normalizing, multi-band compression, and a little EQ, all customized for each setup, but not each individual file. At the beginning of a project, there’s much more time for rendering things than at the end, so even though I ended up processing several hours of audio instead of 50 minutes, it was a more efficient workflow.
We did the same with color correction. Since nearly all of the footage was beautiful right off of the camera, we went for a very minor grade, with very few adjustments which rendered quickly. I decided to render corrections to all the footage, rather than just the selected files. This meant that any last minute adjustments to the edit wouldn’t require additional grading and rendering.
That bought me time to work on the cover design and DVD menus right up until the deadline even as we continued to polish the edit. This was our first project to use Adobe CS4, and we used all the pieces; Premiere for the edit, After Effects for the graphics and color correction, Photoshop for the cover art, and Encore to build the DVD.
Even though Cineform’s real-time engine wasn’t functional in CS4 yet, the codec is so efficient that we got real-time performance anyway. Since we had color corrected anything that needed it in advance, we never needed to render anything in the timeline until final output. Also, Premiere CS4 was rock solid stable no matter how many clips were in the bin or how many sequences were in the project.
I’m already looking forward to our next project. The quality of the 5D’s image, the speed of tapeless workflow, and the stability of the whole Adobe package made this a very fast film to complete, with just over two months from shooting the first interview to mailing off the master. It never felt like we had to fight to make things work in a technical sense (although the script was a different story).
The only things that could have drastically improved the process would be the Cineform real-time engine, which comes out November 6th, and more accurate voice recognition in CS4, which is being upgraded. It really is an excellent production package for documentary filmmaking.