Post-production: The Rough Cut

There is an adage that you make three films during any production: the one you plan, the one you shoot and the one you edit. Once production was done, it was time to start the third phase of the film, called post-production.

Step One: Importing the Footage

To begin, I needed to download all of the footage from the cameras and audio capture devices. It is possible to download these files to your computer and start editing the raw files, but this time, I decided to pursue a best practice in the film industry.

Transcoding Video

Video is usually downloaded directly from a camera as an MP4 file. While this filetype is nice because it is pretty compressed for file transfer, it puts more strain on the computer while you are editing because it needs to be decompressed and managed to play back as you edit. This is normally not too much of a concern for me because I have a Mac with a fast processor, but with the tap dance performance, I needed the sync to be tight with a low risk of lag on the playback. The answer for this was to transcode the video before editing.

To do that, essentially you open all of the files in a media encoder and tell it to process the file and convert it to another filetype, also known as a codec. I chose Apple ProRes, because it is efficient for my particular system. It takes a while to convert the files, and I was surprised at how it the conversion made all of the files way bigger – I had assumed that because ProRes is supposed to make editing more efficient, that files would be smaller. That is not the case. Let’s just say that when you are editing a film, you need lots of storage space to house all of the files as well as a fast processor for good playback while editing.

Sorting the Footage

I prefer to name my files before I start importing into the editor so that I know what I am working with. After the video files were transcoded, I sorted them into folders of the types of shots (A camera vs. B camera moving shots) and who each performer was. I also added information about which take each file represented.

Step Two: Syncing the Footage

This step is crucial before you start assembling a cut of the film. Because I had a separate set of audio files from video files, I needed to begin by syncing them together.

Syncing Two Different Audio Sources

The camera collects audio as you shoot, but it is not high quality – it is far away from the action and really intended to give you something to sync higher quality audio files to. As explained in the article about pre-production planning, I had hired Daniel to collect quality audio through separate microphones during production. His tracks were recorded to a separate device and I needed to download that footage and pair it with the footage collected on camera.

I work in Adobe Premiere Pro, and one super nifty thing about that application is that it has a feature to auto-sync footage. You can highlight two tracks and tell it to analyze and find the similarities, and it will align the two tracks in your editing timeline. Voila! Problem solved.

However, it failed in this circumstance. I find that this feature is much more geared towards syncing voices in dialogue, and struggles to sync subtle sounds like the ones I was working with: movements in a space and taps from tap shoes. So, that meant I had to manually sync the two tracks.

To accomplish this, I had to analyze the waveform of each audio track and look for where they matched. While it was time-intensive, it isn’t that hard to do, because it’s easy to see where the sounds match up from the visual cue of the waveform.

Syncing the Music With the Dancer Footage

After having synced the two audio sources together, the next step was to align the dance takes with the original music that Stephanie Henry composed early in the process. During production, I intentionally chose not to have the music playing live in the space to keep the audio track clean (it was instead delivered to the dancers via an ear monitor). I wanted the audio track to only contain the sounds of the movement. This worked well, but it did cause some level of difficulty for when I wanted to sync the music in post-production.

The process I described above of analyzing the waveforms to manually sync footage didn’t work in this case. Because there was no music playing live in the space, I didn’t have any data to base my guess on where the music aligned. What I did know was the following:

1. The music constantly played in a loop

2. It contained some moments of silence between iterations

Knowing this, I watched the footage and looked for moments when the dancers paused or changed their movements due to a lack of rhythm or inspiration from the music. It was very subtle and not a perfect process, but I worked hard to get it right to make sure the dancers appeared to be reacting with a sense of rhythm based on what they were hearing in the music. In retrospect, it would have been better to somehow feed the music audio track to one of the recordings, but I couldn’t figure out a good way to do that. If anyone out there has a suggestion, I’m open to hearing it!

Step Three: Picking Highlights

With the difficult work of syncing out of the way, now came the fun part: reviewing the footage for my favorite moments of production.

Focusing on Each Dancer

I created a separate file for each dancer’s highlight reel. I chose the moments that I found most appealing to watch despite some of the challenges of production: at some points, the performer got tired, got cold (since we filmed during a blizzard), or made a misstep while experimenting. I flagged each moment that I thought represented the highest points of creativity the dancer experienced during the shoot.

Considering Mistakes

Because the nature of the piece was to focus on making art in the moment, I considered including mistakes in the cut of the film. Mistakes are a part of the process and in a final product, we don’t often see the journey. In the end, I chose to focus on the high points of the work rather than the low points so that the audience could be blown away by the performers’ creativity like I was during the shoot.

Blending Footage

Our two camera setup involved shooting the dancer live head-on (A camera), and inviting in a B camera to enter the shot and circle the dancer. That meant that oftentimes Heri, the B camera operator, was directly in the shot collected by the A camera. I considered what it would be like to share the camera operator in the cut of the film – could it show that this is a process with different people involved in the production beyond the dancer? Is it authentic to “pretend” that there is no crew collecting the footage and leave them invisible to the viewer?

I strongly considered showing Heri in the footage, but in the end, I felt that it could distract the viewer from the dancer’s performance. That meant I had to find ways to work around when Heri was in the shot. Although the dances were all performed live, I cut them apart to make them seem like one continuous take that avoided the disruption of a camera operator entering the shot.

Step Four: Compiling the Rough Cut

Once highlights were selected, I married them together in short segments, one representing each dancer. I wanted to give each one equal time, so ended up allotting about two minutes per dancer in the cut.

Color Correction Pass

We worked with three different cameras in this production. Two were the same type of camera, the Sony A7sii. But, they were collecting footage in different light settings and didn’t always look like they were in the same environment. The third camera was a GoPro that Heri attached to the 360º tripod, and had much different color settings. The trick was to make sure that they all looked related.

Adobe Premiere Pro has color correction tools to help with this process, but it is an art more than a science. I analyze footage from each camera to make sure colors appeared consistent, looking specifically at skin tone and costume colors – was skin a bit more yellowish in one camera? Was the blue of Shari’s dress more purple in another camera? Were the reds of Kaleena’s costume looking consistent between shots? These elements were important to me because each performer had color providing a different perspective on their character.

Audio Pass

The last step before wrapping the rough cut was to analyze the audio to look for issues that could be distracting to the viewer. A key issue we experienced during shooting was that occasionally a refrigeration unit in the space would power up, introducing some additional background noise. Kevin Krein took a pass at the audio footage to minimize this noise, but if it did appear in the track, it was important to add transitions between each cut so that it didn’t sound wildly different from one shot to the next.

Also, if the dancer was closer to a microphone during a moment in the piece, the background sound would swell louder, and I worked to make sure the levels were as consistent as possible between shots. Again, these efforts all work towards the final goal of making the audio a helpful element of the experience and not a distracting one for the viewer.

Ready for Screening

With all of these considerations in place, we were ready to export a version of the rough cut to provide for additional musical performers during live recording sessions. This process would add the final layers to the film’s soundscape.