Wednesday, January 25, 2012

Digital Film Preservation

In the January 20 issue of Daily Variety, David S. Cohen discusses the growing problem of preserving digital films.  


(Above: Ironman2) While the film Hugo was a clarion call for preserving old analog films, it turns out that hanging onto digital film files may be an even bigger and more pressing problem.

The Academy of Motion Picture Arts and Sciences sounded the alarm in a new report called "The Digital Dilemma." The problem stems from the cost of preserving files, the fate of independent productions that haven't yet found distribution, the rapid changeover of file formats and software applications, and the sheer volume of data. According to the Academy, the total content associated with a single digital movie is well above three petabytes. (Each petabyte is a thousand terabytes or a million gigabytes).

Cohen writes: "Digital storage, be it on hard drives, DVDs or solid state memory, simply isn't on a par for anything close to the 100-plus-year lifespan of film. The life of digital media is measured in years, not decades, and file formats can go obsolete in months." 

"The best archiving solution today," says Cohen, "is to print out to film, ideally with a three-color separation printed onto black and white archival film. That's a very expensive solution."

The content at risk is not just the final film and the outtakes, but also the test footage, commentaries, auditions, and concept art. They all might suffer the same fate as the lost plays of Aeschylus, the paintings of Apelles, or the sand paintings of the Navajo.
------

18 comments:

Gregory Lee said...

3 petabytes sounds like a lot, and it is 1500 times as much as I can keep on my little home DVR, which only holds 2 terabytes, so it would be a lot for me to keep. But, gee, films in national distribution cost like 10s of millions or more, don't they? Give me a small fraction of that kind of money, and I think I could give that data a nice safe home.

I'm not convinced there's a problem.

James Gurney said...

Well, I don't know about you, but if I use my personal case as a microcosm of Hollywood's problem, there is cause for concern.

I just chucked a computer that was the only way to view tens of thousands of my old, old documents that I would have had to convert manually (it would have taken me a month to do it). I've got dozens of hours of audio on minidisks that depend on one cruddy old MD player to keep working. I don't have the time to properly conserve all my own digital stuff or the money to farm it out, so I don't even know what I'm losing.

I don't worry as much about all my old papers and paintings.

Aaron said...

HA HAHAHA Take THAT, all you CGI-everything, digital, 3-d moneybags. Sorry had to get that out of my system. I am all for technology, but in the last 5 years especially I have felt that instead of technology slowly testing and working it's way into our entertainment/popular culture, that it has been a full on bombardment to raise the revenue line. Movie make-up, modeling and miniatures and virtually a thing of the past, replaced by digital manipulations and CGI. Let me again say I love how technology has made movies more dynamic, most of the time. Back to the topic and point though, with Skyrocketing costs and NOW massive archival costs, we all know that means even higher ticket prices. The funny thing is Hollywood talks about issues like piracy as a reason for the supertech push, but piracy isn't going down, and the written quality of most movies isn't getting better. The only legit reason I can think of is simply the giddiness of Directors to use any and every new gadget or tool that comes out. I know it sounds goofy, but really what other reason for the rush to supertech, it is killing the cost of movies, and killing everyone's wallet. Now they have to worry about the bombastic cost of storing their wundertech infused movies. Before anyone writes me off as a 29 y/o curmudgeon, think about this in 2003 Lord of the Rings ROTR was released, at a cost of 94 mil., it grossed more than 400mil. in the U.S. alone. Contrast that to Cowboys and Aliens which made 175mil, disappointing I'm sure, but it probably wouldn't have been as big a flop had it not cost almost almost 170mil. to make (and it would have been digital had Jon Favreau not insisted on 35mm). Green Lantern cost north of 200mil. and had 1300 CGI/effects shots spread across four separate effects studios (BTW it broke even). All the Digital, CGI in the world doesn't make a movie MORE enjoyable if it's not well acted or well written, yet some of the best movies (even blockbusters) put the script ahead of the tech. I'll shut up now :)

James Gurney said...

Wiliam, I understand your points, but keep in mind that digital film preservation problems are not punishments visited upon the big studios who overindulge in expensive VFX. They're issues for all films, including the little independent documentaries shot on digital. It's not just a matter of finding the cash to store huge files, it's also a matter of assigning people to labor-intensive cataloguing and file transfers. And what happens five years in the future to digital concept art for stalled projects who never found a home?

zishen said...

The article makes a really good case for analog archiving, but people seem to gloss over one thing that it wrote: Film still does have a lifespan of 100 odd years. What happens after that 100 years?

I`m not a film buff or film archivist, but I have an inkling that replication of said film print wouldn`t be a perfect copy. There would still be some form of loss or degradation. But even if the transfer is perfect, the film master itself might have undergone some form of degradation, causing the new copy to inherit those.

This is all written in the interest of archiving the film for posterity, of course

crowline said...

seems like there is a disconnect between the "3 petabyte" figure and the best solution being to back up to film... if you are transferring to film you are just going to get the "flattened" final frames, right? All the data they are referring to in the massive "3 petabyte" figure has to include all the development files etc. right? If you are throwing that out and just keeping the final frameset, it seems like you could just store that frame sequence in a lossless open format (say openexr?, even png and flac for audio)

something like this (but obviously way higher res than 1080):
http://media.xiph.org/BBB/BBB-1080-png/

It'll be kinda big still, but manageable, and you can at least integrity check the files, and store in a redundant environment. Transferring to analog and back as the "best" solution just makes me cringe. I must be missing something, cause obviously these guys would know what they are talking about...

Also a cautionary tale on the importance of well documented open formats...

Keith Parker said...

So do you think that part of the reason it takes so much space is because of them cataloging photo's of every note, prop, costume, etc? Or because of higher def. pictures and sound quality? Maybe they just need to streamline preservation to what is important...

You are always posting very cool stuff on your blog. Mostly art stuff, although not always directly related. I don't know what you will think of this, but someone posted a link to an article about introverts on my facebook page. I am an artist, and an introvert at heart as well. I figure most artists are introverts, well at least the good ones, because it requires so much time to really learn to draw and paint well. What are your thoughts on this?

Btw here is a link to the article I read: http://www.scientificamerican.com/article.cfm?id=the-power-of-introverts

Wonkey the Monkey said...

crowline beat me to part of what I was going to say, and I totally agree with his(?) proposal that the finished version of the film be given priority, saved in lossless, super-high-definition image sequences.

The problem with legacy formats and such mentioned in the article is mostly related to preserving all of the raw information that went into creating the film. I'm not terribly sympathetic to that concern because the primary reason for keeping, for example, the Final Cut project file for a film is so you can go back and fiddle with it later. I wince when I think about the George Lucas-style re-edits and re-re-edits of films that will inevitably emerge in coming decades because all movies remain infinitely editable.

If a studio or independent creator wants to hold onto that information for future use, that is their responsibility. Archivists and historians should focus on preserving the films themselves in as high quality as possible, and then maintaining those files by performing redundant backups on a regular basis.

Concept art and the like have always been and always will be challenging to preserve because of the myriad methods used to create it. I expect that cataloging of pre-production resources is much better now than in the past because of digital storage; nobody is going to come across a drawer of concept drawings and decide they need the space more than they need those doodles.

James Gurney said...

Great points -- We should probably add that the physical makings of film are often even more expendable -- stuff like latex prosthetics, foam miniatures, sets, and the like. Another great area of losses right now are Saturday morning animation backgrounds.

We can't and shouldn't keep all of the stuff used in making movies -- most of it probably is useless. But apart from the artistic failing of directors who can't stop tinkering with their old movies, there's a value to film historians and to technicians who need to go back under the hood of an earlier film to convert it to worthy new formats.

Elisabeth said...

Sandpaintings aren't meant to last, in fact I think their transience is an important aspect of them. Perhaps not the right analogy here? Otherwise, a very important issue. I spent much of an art student internship in the 1980s trying to preserve and restore 3/4" videotapes that were already demagnetizing. The problems have been there almost from the beginning.

Janet Oliver said...

This comment isn't directly related to the issue, but it does involve conservation. Have you been or will you go to see the New American Wing at the Met? If so, perhaps a review for your readers? I'm saving up my pennies to make the trek sometime this year . . .

youngstudios said...

nothing lasts forever.

fortunatly hollywood hasn't made a good movie in fifteen years so not much will be lost.

Kirk Witmer said...

I had a thought before reading all the comments above, but william may have summed it up best when he referred to "the giddiness of Directors to use any and every new gadget or tool that comes out."
My thought ... all of this is yet another good example of humans outsmarting themselves.

Andrés Carrandi said...

Thanks for posting this. It's really counter-intuituve and I would have never found out in any other way. I'm sharing this with every film follower I know.

Best!

Joel Fletcher said...

We all need to keep in mind that digital preservation is not just a problem for the movie industry. Everyone's personal photos and videos are in jeopardy as well. A lot of people simply do not back up their data, and all hard drives will eventually fail. Due to the rise and ubiquitous nature of smart phones and digital cameras, there has never been a more documented era than the current one. Ironically, some experts predict that ultimately the digital photographic record will be mostly lost due to neglect. Back up your files regularly folks!

Unknown said...

Thank goodness for the good people of Archive.org, as well as our good friends the pirates https://www.pcworld.com/article/248571-2/why_history_needs_software_piracy.html

K. W. Broad said...

This is an issue I had wondered about back when studios first started switching to digital.

It's still a bit of a long shot, but I'd say let's hope that Eugen Pavel delivers: http://www.storextechnologies.com/index.html

Ruby said...

Yeah, that's the other sticking point now, and what is often taken for granted. We tend to assume that digital technology is this huge, monolithic thing that is multi-pronged and omniscient, that the solutions have been generated by it before we've given a moment's thought. Where data storage is such a given with the cameras and hardware that it isn't much of a concern. Yet that is where it's all gonna come down to: whether we'll still have the information around at then end of years and decades. Better start stacking up on storage medium that will not only be adept but that will actually last.

Ruby Badcoe