Monday, February 22, 2010

Mad science: Cum on feel the pink noise.



Neuromathematics, the dauntingly named practice of mathematically modeling neurological processes, may have revealed the deep structure of the language of film editing. James Cutting, a psychologist at Cornell in Ithaca, New York. From a survey of 150 Hollywood movies, Cutting proposes that the pattern of editing in narrative film reflects the pattern of focus and distraction found in the human brain.

In the 1990s, a team at the University of Texas, Austin, measured the attention spans of volunteers as they performed hundreds of consecutive trials. When they turned these measurements into a series of waves using a mathematical trick called a Fourier transform, the waves increased in magnitude as their frequency decreased.

This property is known as a 1/f fluctuation, or "pink noise", [pink noise (left above) is imaged in relation to white noise (right above) - CRwM] and in this case it meant that attention spans of particular lengths were recurring at regular intervals. The pioneering chaos theorist Benoit Mandelbrot found that annual flood levels of the Nile follow this pattern; others have observed it in music and air turbulence.

To find out whether the length of camera shots in films might follow 1/f too, Cutting measured the duration of every shot in 150 high-grossing Hollywood movies in various genres released between 1935 and 2005. He then turned these into a series of waves for each film. He found that later films were more likely to obey the 1/f law than earlier ones (Psychological Science, in press). But he stresses that it isn't just fast-paced action films like Die Hard II that follow 1/f. Rather, the important thing is having shots of similar length that recur in a regular pattern throughout a film.


The text can get a little dense, but the executive summary goes something like this: When human's focus on something, they don't throw an intense laser beam of concentration on it. Rather, we look for patterns that define the object observed and groove along on those patterns. To many breaks in the pattern and the object starts to require serious expenditures in energy to puzzle through it. Too much similarity, we zone out.

Cutting's argument is that the editing techniques of Hollywood style narrative follow the generalized pattern of human attention. They fall in the sweet spot that requires our attention but doesn't overly tax the system. Cutter, in a refreshingly non-Bordwellean turn, doesn't suggest this formal element "explains" movies.

Cutting suggests that obeying 1/f may make films more gripping because they resonate with the rhythm of human attention spans, but he doubts that directors are deliberately using mathematics to make movies. Instead, he thinks films that happen to be edited in this way might be more likely to be successful, which in turn would encourage others to copy their style. This would explain why a greater number of recent films tend to follow 1/f.

Cutting, a film noir fan, is the first to admit that shot-pacing isn't everything: he found that the lengths of shots in film noir movies are typically random and not correlated with one another on any timescale. Star Wars Episode III (pictured), however, which he describes as "just dreadful", adheres rigidly to 1/f. He says that a good narrative and strong acting are probably most important.


Cutting's work does jibe with some other work in the field:

The attention theory chimes with other recent work, Tim Smith at the University of Edinburgh, UK, tracks the eye movements of movie-goers. He has shown that the editing style of modern films results in more people being focused on the same areas of the screen at the same time. He has interpreted this as a sign that audiences are more attentive to the film.

UPDATE: Another article discussing the same work by Cutting and others gives Cutting's hypothesis on just how the pink noise pattern became so prevalent in Hollywood blockbusters.

These researchers don't believe that filmmakers have deliberately crafted their movies to match this pattern in nature. Instead, they believe the relatively young art form has gone through a kind of natural selection, as the edited rhythms of shot sequences were either successful or unsuccessful in producing more coherent and gripping films. The most engaging and successful films were subsequently imitated by other filmmakers, so that over time and through cultural transmission the industry as a whole evolved toward an imitation of this natural cognitive pattern.

6 comments:

zoe said...

this is fascinating...i got stuck watching part of troy once, while i was waiting for someone, and i thought i was dying. i agree about plot and acting, but i wonder if the extremity of the (brain-leak) response had to do with "too much similarity?"
interesting...thanks!

theverysmallarray said...

"Cutting?" In a science article about film editing? Sounds like April fools came early, but we all know that even frame rates can make a difference in impact on scenes, and a studio's inter-ference with a director's vision in the name of minutes can stop a film cold. As to Star Wars-Lucas really did need Marcia's deft touch to sell a movie. It also seems time and again that in director's earlier films they use editing to dodge problems they come across in post-production and find a stronger movie for it. But any art is like that. The lean and quick, saying it the best you know how, to the middle-age, indulgent spread, to repeating yourself ad-nauseum, and nothing ever quite working out. The only directors that escape that try something different. Eastwood, maybe? Here's a more interesting question? Why don't excellent film editors make for excellent film directors if this theory is true (and we all know they do not)?

Sasquatchan said...

it's art, man, you can't make deterministic science out of art..

cattleworks said...

I'm not sure if I quite caught all that, but the subject sounds interetsing.
I'm probably misunderstanding the article, but for some reason I was thinking of the whole idea of MTV editing, where it seemed for awhile there that editing narrative was reduced to the frequency of edits to imitate the pace of music videos that became suddenly popular when MTV was originally unleashed on the cultural scene.
I hope to make some films before I die, and I'm talking homemade movies here, not Hollywood films. I've been editing some really amateurish footage I shot in 2002 recently (because only recently did I have access to an iMac and iMovie), and it's an interesting experience.
The previous comment about whether or not good editors make good directors is interesting. But it makes some sense that it doesn't necessarily follow.
Because the editors are working with materials brought to them, and they can focus on what is there to create a satisfying whole.
But that doesn't necessarily mean they can also create the material they have to edit, if that makes any sense.
Maybe that doesn't make any sense.
Because you would think that storyboarding would be a part of the storytelling process, and storyboarding would seem to be well suited to an editor.

But, getting back to my slim experience at editing. I can sort of see how sometimes it takes a little while to edit a film, because you have to absorb the process of converting raw footage into actual individual shots and seeing how they cut together, and the rhythms and associations the assemblage makes upon viewing at normal speed.
My rather half-assed, purely instinctual test is editing my footage, and then just running it and see what I feel watching it.
Theoretically, if it works, I won't feel the need to tamper with it anymore.
But, process-wise, it seems to help to do some intensive editing, then watch it. And then allow some time to go by, at least sleep on it one night, before watching the cut again to see what you think.
And that's just some organic, interior process.
Reducing the cuts to simply mathematical rhythms may offer some explanations, but I think it works best as an explanation of an audience's reaction to a film, but it can't really work in reverse, that is, coming up with a mathematical formula to simply cut footage at various speeds and hope the assemblage will achieve the same reaction with the audience, because we're not addressing the context of the subject being filmed.
Does the timing that the study refers to go to the exact frame or not? Because sometimes a single frame can make a huge difference in an audience's reaction, and that has nothing to do with rhythm but what they're watching.
You know?
I'm probably just talking out of my ass, because I'm making this up as I go along.

And, yeah, the name "Cutting" was an interesting coincidence...

CRwM said...

Cattleworks and Sassy,

I've added an update to the post that provides Cutting's explanation for how this particular pattern evolved and, as you both point out, it develops from the work to the pattern, not the pattern to the work.

Anonymous said...

About the "Cutting" thing:

Richard Wiseman (sic!) published a study in his book "Quirkology" about a significant correlation of familiy names and professions. When asked, the persons concerned answered: concidence. Wiseman assumes a complex, concealed social relation that reflects implications of the name to the bearer.