Previous  |  Next



Tim O'Reilly

Tim O'Reilly

Movie Shot Lengths and Attention Deficit Disorder?

Paul Kedrosky has a fascinating post up entitled Open Source Attention Deficit Disorder Measurement. He writes:

There are stories making the rounds about people becoming nauseated while watching the popular new movie The Bourne Ultimatum. Not because they movie is so bad, but because the shot lengths are so short, averaging (apparently) something like two seconds. In some people that sort of thing -- alongside fast camera moves -- seemingly induces vomiting. Fascinating.

That, however, got me thinking. Many people, myself included, think movie shot lengths are getting shorter and edits closer together. It is, to one way of thinking, a reflection of our collective attention deficit disorder....

So, is it true? Have shots gotten shorter over the years? Until recently that wasn't something on which you could readily find data, but now you can at least begin to, courtesy of a growing database of public movie shot-length data at Cinemetrics.

This is indeed fascinating, and indicates how collective intelligence is spreading into ever more obscure areas. People are cooperating to measure shot lengths in movies, through the bionic software combination of a computer tracking program that watches the movie with you, and the volunteer clicking each time a scene changes.

This is fascinating. You learn something every day. From the cinemetrics site:

In verse studies, scholars count syllables, feet and stresses; in film studies, we time shots. "If I use one word, I would have to say timing," Chuck Norris said in a recent interview to ABC’s Nightline answering what attribute won him six karate world titles. "Timing I think was my key thing. I was able to figure out the timing to close the gap between my opponent and myself and move back, and that was I think the key." Much like martial arts, or like poetry and music, cinema is the art of timing. This explains why, early on, filmmakers as Abel Gance or Dziga Vertov in the 1920s, or as Peter Kubelka or Kurt Kren in the 1960s not only counted frames when editing, but also drew elaborate diagrams and color charts in order to visualize the rhythm of their future film. This also explains why a number of scholars interested in the history of film style (as Barry Salt in England, David Bordwell and Kristin Thompson in the US or Charles O’Brien in Canada) count shots and time film lengths to calculate the average shot lengths of the films and/or use these data in their study.

But I digress. My real topic is collective intelligence. There are many different kinds:

  • Perhaps the most important, the discovery of meaning in activity that people are already doing -- for example, in Google's discovery of PageRank, the idea that link activity can be useful in providing better search results.

  • Explicit volunteering to build a collective work, as in open source software, or more radically, in the collective editing environment of Wikipedia.

  • Automatic monitoring of user activity, as when watches your music listening habits to make recommendations for music you might like, or when Amazon watches your buying habits to recommend other purchases you might like to make.

  • Some hybrid of the two, as at Cinematics.

And obviously, all these areas bleed into one another. When your camera starts automatically geocoding, it's moving the geotagging of photos from a conscious user activity to an autonomic, computer mediated one.

We'd love to hear more examples of hidden data pools being created or brought to light. We're interested in this because of the Money:Tech conference (of which Paul is the program chair), but even in areas like movie scene timing, which is obviously not related to the subject of that conference, we're fascinated by the phenomenon.

Where else do you see new forms of meaning being discovered in data that is already being collected, or new, meaningful collections of data being created, either by new devices, or by people collaborating with each other or with their machines? What has made you sit up and say, "Wow! We really are on the threshhold of something very new and different?"

tags:   | comments: 13   | Sphere It


0 TrackBacks

TrackBack URL for this entry:

Comments: 13

Christopher Palmer   [08.28.07 09:42 AM]

On a slightly related topic, the book Everything Bad is Good For You discusses the cognitive effects of fast edits in movie and TV. It also discusses user created datasets and statistics gathered for various TV shows, such as number of movie references in Simpson's episodes, cross referencing of "in-jokes" in Seinfeld episodes, character association networks in 24.

I've been doing quite a bit of related reading about the pros and cons of changing media and the automatic collection of data and writing up some reviews and notes. Some other good books are Everything is Miscellaneous and The Cult of the Amateur, which take the same basic facts and come to radically different conclusions about their impact.

adm   [08.28.07 10:08 AM]

this is a good post.

one other item that falls into this category, i think, is that thing with the Notre Dame cathedral a couple months ago, where the researchers automatically stitched all those photos together from a public image repository. the possibilities there are pretty great.

Mikel   [08.28.07 12:52 PM]

One of my favorite examples of surprise in the collective is a couple years old .. krazydad's time graph flickr photos tagged with sunset...

Taking the time of day sunset-tagged photos were captured, and the time of year, a graph clearly shows a sinusoidal curve of changing sunset hours in the northern hemisphere. There's several interesting secondary effects revealed too.

Thomas Lord   [08.28.07 10:51 PM]


Shots are getting shorter because the makers of moving pictures of all sorts have, over the past century or so, begun to amass a rich, iconographic vocabulary. This is evident on the surface or in any gloss, even over the craft (c.f. the "On the Lot" T.V. show and Scott McCloud's classic "Understanding Comics"). This is self-evident on the surface. The high art has little to do with this -- the best directors play with parameters like tempo but generally don't use the common iconography except ironically or with a broad and occaisional formality or in order to up-end it by finding something genuine within it. But, the bulk of stuff that gets cranked out of the meat grinder....

But, per web 2.0 thinking, the main story here is that with the right web site you can get N+1 people to throw on a DVD and sit there with a stop watch and a pad of paper.



Tom   [08.29.07 12:59 AM]

The reason for decreasing shot times may be partly technological too, for at least a couple of reasons I can think of:

1. It's technically feasible with digital editing to edit thousands of shots into a movie. That would have been a nightmare with scissors and sticky tape, not even considering the effort required to catalog and keep track of the source footage.

2. CGI special effects may actually require that we don't look at them too long, otherwise the illusion is broken. The "uncanny valley" may become more obvious if you get the opportunity to stare at a scene too long, if you're only given a quick 2 second impression you might not notice.

That and what Thomas said above about building a vocabulary or short-hand to communicate with the audience efficiently. Once the audience knows the jargon you can compress more into a movie (though in my experience that just can't be happening!)

Love that the data exists, whatever the cause of the trends it exposes.

Victoria Squires   [08.29.07 05:03 AM]

Well I'm one of the ones who feels sick. Some movie have flickering lights - yuurk! Some have that terrible panning or a camera on somebody's shoulder - makes me sea sick. And cutting from one scene to the next too quick makes it unwatchable.

Come on film makers, just get good scripts, good actors, and put the camera on a tripod and leave it there for a bit!

Roman   [08.29.07 07:51 AM]

Regarding "...hidden data pools being created or brought to light":

Somewhat similar to Mikel's comment:
There is a web page that visualizes the last 10 months of U.S. Stock Market actvity(by using screenshots from

Can the human eye spot patterns more easily than do algorithms? I don't know (yet).

Thomas Lord   [08.29.07 12:33 PM]

Tom, thanks for pointing out the role of technology in shot times. Yup, that's obviously part of it, too.

It seems to me like we're entering into a really exciting era in film based on technology. First, as you pointed out, we got fancier effects (Star Wars IV models, for example) then CGI. The new breakthrough may not have started with but was first loudly demonstrated in the first Matrix film: revisiting the relationship between film and photography. A moving picture is a sequence of stills and the famous flying-kick or dodged bullet scenes in the Matrix were rooted in the "aha!" realization that the sequence of stills dont' all have to come from the same camera to form a smooth pan. Rotoscoping reminded us that in C_I you can fill in the blank with something other than G. I hear the new "Speed Racer" movie will explore yet another technique for combining frames from multiple cameras into a single continuous shot. Technology is turning the design of cameras used to shoot a film into an axis of directorial creativity -- directors with a strong understanding of photography are getting all kinds of new degrees of freedom. It's an exciting time!

I still get creeped out, though, at the reduction of all this just to an opportunity for crowd-sourcing data collection for the moral equivalent of a degree thesis. That take imposes something between the audient and the film that wasn't there before. Audient is reduced to watching the film to satisfy somebody else's goal. The meaning of the film is how it contributes to the economic project of the guy collecting the data. I understand the desire to teach people to be able to look at phenom. in the world through "web 2.0" glasses -- go, Tim, go -- but:

Here, in this case, in this gentile, largely inconsequential case, we can begin to see the converse side of the "web 2.0" coin: that it is "the giant sucking marketing machine" imposing itself in new places within people's lives and distancing those people from the case of their individual lives and experience. When I start appreciating film not for itself but because I can blog about it or because it gives me an excuse to participate in an on-line data-collection forum, what am I really gaining? Is my world enriched? Or is my distraction from the thing before me an impoverishment?


Thomas Lord   [08.29.07 12:43 PM]

Oh, and duh:

If you want to find the intersection of web 2.0 and cinema, start with the TV show "On the Lot".

How about instead of crowdsourcing the collection of cold, dead, pointless data we crowdsource the search for talent and the development of human-scale culture?

Most "reality" shows crowdsource the search for the bitchiest or most manipulative people or the biggest drama queens, etc. That one, though, is like applying a fixpoint combinator to the meme of a reality show: an entertaining production constructed by crowdsourcing a search for innovation in the field of creating entertaining productions. The producers of that show are just too damn smart, thank goodness.


Ryan Miller   [08.29.07 10:47 PM]

An interesting take on collective intelligence is Stuart Skorman's project Cinemasaurus. The project is taking the collective intelligence of experienced video store clerks to create a gene map of movies, tagging their attributes relative to one another. As I understand it, it will be a discovery tool like Pandora was for music.

It is still in development, but a podcast interview is here:

Christopher Palmer   [08.30.07 09:12 AM]

@Thomas Lord,

I second the recommendation for Scott McCloud's Understanding Comics. For a book that has no stated agenda about being about cognitive psychology, it has some of the best insights on cognition, pattern recognition, concept formation, and emotional concepts of symbols that I have ever read.

The section of "Understanding Comics" regarding the different perceptions of time via panels (and in the interstitial white space between panels is particularly fascinating in regards to the different ways movies handle it. For example, a panel can show a indeterminate period of time (as when several people are shown talking), a single frame of time (as during a sequence of action shots), or two panels can be perceived as happening simultaneously. Between frames, a millisecond, an hour, or a thousand years can pass and readers can perceive this through visual clues even without the "Meanwhile, across town..." or "Present Day" text blocks over the images. Film, of course, has a different visual vocabulary of editing, but the fact remains that viewers "learn" these languages by experience. A reader of "Little Nemo" from the early 20th century might have quite a bit of difficulty with modern comics until they picked up the nuances just as an average filmgoer from the 1940s might be perplexed by the rapid fire editing of the Bourne movies or the non-linear presentation of the story in Pulp Fiection.

The book I recommended, Everything Bad is Good For You, says that in general, this is a good thing. Even escapist movies like the Bourne films and modern television series (even sitcoms like Seinfeld and The Simpsons) require much higher congitive skills than, say Dallas or The Flintstones or Casablanca. In other words, they may not be giving us all ADD, they may be training us to tune our perceptions and make quicker evaluations.

DERRICK   [10.17.07 01:28 AM]

I'm read to work with you

Amie Stilo   [01.10.08 02:48 AM]

I recently watched The Bourne Ultimatum and thought it was actually not as bad as they said it would be. The short shots are definitely short but what I find more nauseating is the camera shake, in The Bourne Supremacy the camera shake was really bad.

Post A Comment:

 (please be patient, comments may take awhile to post)

Remember Me?

Subscribe to this Site

Radar RSS feed