When I started 20 years ago, post production was still very much a film medium, characterized by flatbed editing machines, film bins and racks of workprint. Then came the transition to videotape as a film replacement, and, as world events and the influence of file-based news cameras brought data capture into acceptance, tape was superseded by file-based post.
As a post-production engineer, I’ve been lucky to be able to experience these three successive generations of production media first-hand. And my close colleague and frequent collaborator Mike Nuget, has lived the front line of these changes as one of PostWorks’ senior finishing editor and colorists.
The industry’s adoption of file-based workflows stands in stark contrast to film-based methods. Film as a medium is more than 100 years old; it is (or was) reliable, stable, mature. In great contrast, file-based workflows are in a constant state of reinvention; there is nothing stable about them. If film was a rugged old man, then files are impulsive teens, full of energy and possibility, but never quite the same from day to day.
To best understand what daily life is like in an all file-based world, consider the most basic question asked about every post production process: “How long will this take?”
It’s a very reasonable question! Film- and videotape-based workflows yielded mostly obvious answers, as those processes were more linear, more predictable: “How long will it take to capture a one-hour videotape?” Well, an hour, of course.
But with today’s file-based workflows, these answers are not so simple. “How long will it take to transcode this file?” you may ask. “Well,” I would answer, “that depends on the codec, frame size and bit rate of the source, and on the target codec, frame size, etc. of the media generated. And it also depends on the speed and age of the computer system doing the work. By the way, that software version you’re running does not yet support this new camera codec ….”
It’s not a satisfying answer.
The upside here is that processing files can be, and often is, faster than videotape — and often much faster than real time. The downside is that process time is often difficult to predict. This is an ambiguity that applies to nearly all aspects of modern post production. (Again: impulsive teens.)
For most productions, daily life begins largely at night. For scripted television and feature films, the “dailies” process begins to unfold the moment the camera media arrives in post: The color pipeline developed on-set is applied to the footage in the form of Color Decision Lists (CDLs) and Look Up Tables (LUTs), with additional manual grading as needed. This is followed by syncing the audio and encoding it, often with four or more deliverables for editorial, for marketing and for web-based and mobile review systems. And all of this is done overnight, executed by teams of yeoman colorists, engineers, media technicians and project coordinators. It’s “time to make the donuts,” taken to an exceptional extreme.
For the nonfiction production (reality TV, science and factual), this process is similar but often at incredible scale — shooting ratios have been known to push past 700:1.
Imagine this: 4,300 hours of footage, for only eight 44-minute finished episodes (these are real numbers, from a client at PostWorks.) This represented “only” 103 terabytes of raw camera data. And that’s before editorial proxies were created, generating another 21+ Terabytes. Once editorial proxies were ready, all 4,300 hours of multicam footage had to be grouped together, tagged for content and transcribed, so that story editors and producers could develop their narrative.
Ultimately, next-generation computing systems and software may offer more power to process files faster, but there is always more raw material to process!