Bill Cheswick's Thumbnail Experiments

Introduction

This year I bought a new HP DesignJet 6100ps wide-format printer for the Infolab. I wanted to run it through its paces, so I decided to try printing every frame in a movie, and other collections of images. I want the highest reasonable resolution for each image: I don't want to see half tone images.

My goals: to see what it looks like, of course. For movies, can this new way to see a movie give insights into color palette and staging, and scene pacing and transitions? Can film editors, directors and students learn something using this visualization?

Film makers take great pains to get the artist effects they want on film, and I have taken some pains to try to reproduce there efforts on paper. This is no easy: the inks are not the exact same colors as the films, and the printing process has a much different dynamic range and gamma.

The results look pretty good to me, but I haven't received any professional feedback yet from a pro.

Split Ends

The movie shown here is Split Ends, an independent film that appeared in the Hoboken film festival. I am grateful for permission from Bob Pusateri, Elizabeth Christensen, and Gila Zalon, the producers, for permission to display a visualization of their film.

You can see that the printout is quite long: about 10 yards for this 90 minute movie. It is all printed on a single sheet of paper, at close to the highest resolution the printer can manage. There are about six seconds of movie per column.

If you are familiar with this movie, the various scenes stand out: the golf course scene, the blown fuse scene, and the striking lighting of a night scene shot while walking past various storefronts. The opening credits are done in quite a different style, and it shows.

Sita Sings the Blues

Nina Paley's movie is quite striking, as most animations seem to be. Here she is shown with the smaller 30-seconds/column version on glossy paper. This is definitely frameable.

You can read more about it in her blog. She has provided a number of insights and suggestions that have improved this project.

We hope to display her movie at MOMA, perhaps.

Other Movies

Of course, I would like to try this on many other movies. The problem is one of copyright violation. While some argue this is fair use, it is clearly debatable, and I am not trying to blaze new legal ground here.

Some of the movies I'd like to see are The Princess Bride (one of my favorites), Fantasia, Star Wars, Monsters Inc., the Matrix, and many others.

In particular, I am interested in the reactions and suggestions of artists and teachers that can improve the visualization, or suggest new uses.

There are movies in the public domain, but almost all of them are in monochrome, or simply not very interesting. There are a couple of Warner Brothers cartoons that slipped out of copyright, and they look great.

Seeking Permission

So I am seeking permission from film owners to print out a single copy of the film of their choice, and give it to them. I'd love to hear commentary about the fidelity of the visualization, and whether it is useful in any new way to a film editor or director.

In addition, I'd love permission to simply display the given movie on one of the many long, vacant walls here at AT&T's Shannon Labs here in Florham Park, NJ. This is not a public display, but is visible to researchers and cleared visitors to the building.

With further permission, I could show off one or more movies, perhaps at a conference like TED. Other uses might require further work. I'd love to see a movie shown in this form at MOMA, a project I suspect they would love. I know people who know people there, but haven't made any contacts about this yet.

Making Money

I believe the owner of a popular movie could make a fair amount of money with these visualizations. I have a few ideas that seem to me to be money-makers. I am not looking to get into the business myself.

Other visualizations

I have tried other visualizations. For example, all 43,000 Cheswick family photographs fit on a couple square yards of paper, usefully. A have also tried visualizing some time lapse movies I've made, and my own stop-action film Gorgo Eats Miami. I've also played with some sequences captured on TV, to learn more about telecine and the varying frame rates found there.

Poster length computations for Sita are here.