Cinemetrics – Film Analysis Meets Raw Data
The past century has seen the rise to preeminance of statistics. Aided by utopian ambitions and exponential growth in computing power, raw data became the currency of the age, heralded as the solution to society’s ills. If it couldn’t be measured, tallied, and squeezed through an algorithm, it didn’t hold a lot of weight. Issues of war production, food distribution, infrastructure development and much more were put under the guidance of Whiz Kids, who used emerging sciences like game theory and computing to crunch data and find large-scale solutions. Urban planning, architecture and medicine, formerly considered humanistic disciplines, as much art as science, were now looked on as glorified math problems, with ever more data analysis slowly sifting out a solution.
Certain academic fields of study remained impervious to the trend, however, notably in the study of the arts. Film theory fell into this category. How could you quantify esoteric qualities inherent in aesthetics, or pigeonhole an auteur’s oeuvre into a logical analytical system? No, filmmaking was outside the cold strictures of charts and graphs. Until now.
Enter Cinemetrics, a software program which aims to reduce films into easily digestible graphs. The tool, an in-development thesis project of a Dutch college student, is an ingenious new way to analyze films for the data-driven age. Breaking films into their component parts – color palette, movement, shot length – the structure of a film can be ‘seen’ with a simple glance at an animated graph (‘digital fingerprints’). Watch this video to see how it works.
The potential uses are endless. These visualizations can be surprisingly handy way to understand a film’s gestalt, and can give added insight into a director’s thought-process with his or her choices of color, motion, and pacing.
Comparing two sci-fi classics, check out the unrelenting action of Aliens vs. the static movement in 2001. Or check out the pacing, dynamic range and color distribution of The Shining:
Now, how much do these tools tell you? They won’t help you detect the Native American commentary of The Shining or the use of the Gaze in Vertigo. Close, repeated watchings are still the primary method of understanding a film. As in most things in life, balance is key. Assuming that just because a methodology solved one set of problems, it will magically work for any problem is naive. The techniques used to equip the US at breakneck speed to take on the greatest military powers during World War II, created a lot of unlivable places in the wake of that war, as communities of living individuals and marketplaces were reduced to a jumble of data points to be zoned in the most rational way: e.g. strip malls, subdivisions, office parks etc. This is why pre-industrial cities like Florence and Paris, built without the luxury of reams of data, still manage to beat postwar cities like Phoenix and Houston in that elusive category: quality of life.
But data, used rightly, can be another tool to understand the essence of something, and that certainly seems to apply here. Film theorists and filmmakers just got another powerful instrument in their toolbox.