there's a formula for the length of each shot in a movie

[Florin Andrei]

Psychologist Professor James Cutting and his team from Cornell University in Ithaca, New York, analyzed 150 high-grossing Hollywood films released from 1935 to 2005 and discovered the shot lengths in the more recent movies followed the same mathematical pattern that describes the human attention span. The pattern was derived by scientists at the University of Texas in Austin in the 1990s who studied the attention spans of subjects performing hundreds of trials. The team then converted the measurements of their attention spans into wave forms using a mathematical technique known as the Fourier transform.

Article:

http://www.physorg.com/news185781475.html

Link to original paper:

http://people.psych.cornell.edu/~jec7/pubs/cuttingetalpsychsci10.pdf

 

[Christian Tanner]

try to watch a few lumière brothers flicks and the like - then, right after, skip to an action scene in "end of days"...

 

[Keith Putnam]

One assumes that Cutting (how appropriate) is familiar with Sergei Eisenstein's theories and practices of montage... though I'm pretty sure Eisenstein's mathematical methods were less complicated than Fourier transforms.

 

[George Butterfield]

Psychologist Professor James Cutting and his team from Cornell University in Ithaca, New York, analyzed 150 high-grossing Hollywood films released from 1935 to 2005

So right away we have a pre-determined subset of all films. In one sense I suppose the high grossing films are the "best" since making money is the first goal of most producers. Making quality films is not the first goal. Is there shot length best suited for the making of "quality"films?

 

[David Rasberry]

So right away we have a pre-determined subset of all films. In one sense I suppose the high grossing films are the "best" since making money is the first goal of most producers. Making quality films is not the first goal. Is there shot length best suited for the making of "quality"films?

I don't think narrative or visual quality per se is the issue here, just that human attention to organized visual stimulus bears a relationship to well known mathematical structures in physics (1/f) that span many different phenomena. Film makers have intuitively taken better advantage of this as their craft has evolved.

 

[sander kamp]

LOL written by professor Cutting!

 

[Subhadip Sen]

That article is overly scientific. Cinemetrics is the best shot regarding shot lengths. There's no doubt that the average shot length has reduced considerably over the years. Yes, it may have to do something with decreasing human attention span. On the other hand, we have directors like Bela Tarr or Theo Angelopoulos who have continued pushing the boundaries of the shot length.

 

[Andrew Rieger]

Bela Tarr is the best example of the opposite direction but his films are an acquired taste. I love them but other people that I have tried to show them to call his films almost unwatchable and very boring. Even some critics have found his long takes to be overly long. His longest film, which is like 5+ hours has significantly less shots than a typical 2 hour movie so you can see why people would be annoyed. Still a master of the art IMO.

 

[Ryan Sims]

English please

English please

In plain English, I get out of this paper that:
1. for slow scenes, all the edits are slow and close to the same length
2. for action scenes, all the edits are fast and also close to the same length.

They just applied math to compare shot lengths of films over the past century of film making.

Anybody else get any more out of it?

 

[David Rasberry]

In plain English, I get out of this paper that:
1. for slow scenes, all the edits are slow and close to the same length
2. for action scenes, all the edits are fast and also close to the same length.

They just applied math to compare shot lengths of films over the past century of film making.

Anybody else get any more out of it?

I got a connection to Fourier wave theory out of it. The concept in a simple sense that a lot of small ripples can add up to a large wave front if the timing is right. Thus the pace defined by length of shots in a scene can be used to build to an effective climax that has maximum visual/emotional impact. The article did not deal with the dramatic nature of film, just the mathematical relationship of shot length with wave theory.
There is a cultural shift too in the faster pace of recent films compared with older ones. I experienced a little of this first hand producing educational shorts for an art museum to introduce gallery shows. The audiences ran from kindergarten age to seniors. Watching people respond to these shorts, I learned that a pace that worked for people who had not grown up watching TV vs kids that had watched from infancy was very different. The kids could pick up visual cues much faster and get more out of the visual structure while older adults relied more on narrative and couldn't follow fast visual sequences as well.

 

[Noah Kadner]

Um it's art not science, flavin flavin...

Noah

 

[Dan Hudgins]

Editing

Editing

Yes, movies that are good and are edited well do have a pattern.

I analyzed the Alfred Hitchcock and others like INDIANA JONES and there are very distinct patterns that are useful in planning shooting.

Do this:

1) Purchase a hand counter of the type used for counting people or things where you push the button and it counts up.

2) Get a good digital stop watch.

3) Get some lined paper and graph paper.

4) Put a DVD into the player and start it.

5) Count the start of the fade from black or title start as the start.

6) Push the counter each time there is a film CUT.

7) Every 5 minutes write down the total on the hand counter.

8) When the film is over, subtract the total for each 5 minute period from the next one to get the number of shots in each 5 minutes.

9) Plot the shots in each 5 minutes on graph paper.

You also get the average shot length by dividing the running time by the number of cuts.

If you do it right it you will see clear patterns on some films, so much so that the editor must have had some kind of plan, on other films it would seem the filmaker and editor were not working on a "macro scopic" time frame.

I made a computer program to automate this and studied many films, with the computer data you can get the lengths for every shot. A 5 minute sample time seems to make the best graph most times.

The movie Them! (1954) makes a nice graph and might be a good first try.

Its easy to skip a cut if you have the sound on and get watching the film, so turn the sound off and watch just the cuts.

Also try Toy Story (1995). In many films the cutting rate peaks about every 25 to 30 minutes and the same for the slow parts. The rate of climb vs. fall in the most interesting detail, see the last part of the US Edit of the The Third Man (1949) (not the original, the film was re-cut for the US market by someone in Hollywood that seems to have known about cutting rates).

This information is useful in using your "shot budget" if you movie has 800 shots, then you need to block the scenes in some places for a one minute shot and in other places for maybe 30 shots in that scene, which you do depends on where along the movie "time line" that scene is, so you cannot move a scene around without re-cutting if you want the "graph" of the finished film to look right, and the audience to "feel" the cutting rate change in the right places.

 

[Elsie N]

I got a connection to Fourier wave theory out of it. The concept in a simple sense that a lot of small ripples can add up to a large wave front if the timing is right. Thus the pace defined by length of shots in a scene can be used to build to an effective climax that has maximum visual/emotional impact. The article did not deal with the dramatic nature of film, just the mathematical relationship of shot length with wave theory.
There is a cultural shift too in the faster pace of recent films compared with older ones. I experienced a little of this first hand producing educational shorts for an art museum to introduce gallery shows. The audiences ran from kindergarten age to seniors. Watching people respond to these shorts, I learned that a pace that worked for people who had not grown up watching TV vs kids that had watched from infancy was very different. The kids could pick up visual cues much faster and get more out of the visual structure while older adults relied more on narrative and couldn't follow fast visual sequences as well.

I've seen something like this even in home movie situations. I remember one video in particular where some very young family members (pre-teens) would actually stop watching a video about themselves and go do something else, while their parents would continue watching attentively. On the other hand, some other videos with long takes, when accompanied by relevant music, seemed to hold everyone's attention.

 

[Michael Romano]

A mathematic formula might actually apply to good storytelling but would you ever use it in practice? A good editor just knows when to cut, end of story. He/She use their embedded talent and in the end it fits the formula, without ever having to know how it works. It is very interesting info, I love that sort of stuff but I won't think about applying it to the creation phase of a film. It really shows more about how attention spans have dropped sharply over time in this country than how to formulate a hit theatrical release.

Thanks for the read!

 

[Dan Hudgins]

coverage

coverage

A good editor just knows when to cut, end of story.

Hitchcock then was not a good "editor" because he worked from a plan? (more so in the later Hollywood films like Birds).

The editor can only cut from the footage shot, if there is not enough coverage shot the shot lengths will not be there to get the distribution of lengths by plan or instinct.

It cannot hurt any filmmaker to look at the macro scopic editing in films to know those that do well from those that don't. Unless you have looked at the cutting rate graphs you cannot understand the things about editing that are not perceptible on simple viewing of shot to shot connections.

The larger the scale of the trends the more relevant they are to your own projects, things such as average shot counts and lengths, slope angles on the change in the rate of cutting, etc. are all useful in planning the shooting of the movie. The editor can trim the shots to length, but the director must block the shooting so that the shots come out the right length to start with.

If you don't shoot the needed close-ups and point of view shots along with the coverage the editor will not have those shots to cut with.

ILM and George Lucas set up a film research unit and their films show distinct graphs. I did my research about 25-30 years ago and talked to Coppola about and he told me to call one of his people that was associated with American Zoetrope as a Dramaturgist or something like that, when I talked to him on the phone he said that he was familiar with such concepts and was interested in meeting to talk about it. That was in the late 1970's I think.

Since the paper/html link says that, quote "By contrast The Empire Strikes Back (1980) and the 2005 blockbuster movie Star Wars Episode III (which Cutting considers to be “just dreadful”) both follow 1/f rigidly" one can think that someone at ILM et al. was working from some kind of plan at some point in the process.

At a minimum doing analysis of your end result lets you know how much it deviates from the norm for profitable films, and knowing how to make adjustments to bring it closer to the norm could improve the return on investment. How can it be better to be oblivious to patterns and trends that are common in successful films, or at least the average shot length etc. for planning the story board?

The editor does not have full control over the editing order or length of the shots in a story boarded film when the director is not giving coverage for every possable cutting of each scene.

 

[gameboykk1]

Bela Tarr is the best example of the opposite direction but his films are an acquired taste. I love them but other people that I have tried to show them to call his films almost unwatchable and very boring. Even some critics have found his long takes to be overly long. His longest film, which is like 5+ hours has significantly less shots than a typical 2 hour movie so you can see why people would be annoyed. Still a master of the art IMO.

 

[David Rasberry]

A mathematic formula might actually apply to good storytelling but would you ever use it in practice? A good editor just knows when to cut, end of story. He/She use their embedded talent and in the end it fits the formula, without ever having to know how it works. It is very interesting info, I love that sort of stuff but I won't think about applying it to the creation phase of a film. It really shows more about how attention spans have dropped sharply over time in this country than how to formulate a hit theatrical release.

Thanks for the read!

L. Francis Herreschoff, a famous American yacht designer said something to the effect that the artist creates intuitively. The mathemetician figures out how he did it after the fact. The engineer then reproduces it.

My experiences producing for kids is not that they necessarily have short attention spans, they can process and retain visual information so much faster than us old farts that it takes a lot more effective content within a given time frame to keep them interested.

 

[Marc Wielage]

A mathematic formula might actually apply to good storytelling but would you ever use it in practice? A good editor just knows when to cut, end of story.

Amen to that.

I think it's impossible to reduce the art of editing to math, science, and numbers. To me, it's all about emotion, mood, and feeling. There is no magic formula that will work, either for editing, for filmmaking, or anything else. Walter Murch's In the Blink of an Eye goes into this far better than I ever could.

To me, the only thing that works for editing is gut instinct and experience, plus patience and a lot of trial & error. Shot length means nothing, because you have to balance in the issue of camera movement, long dolly shots, quick handheld scenes, and so on, which also dictate pacing and timing.