The Afterlife Is Expensive for Digital Movies

A new study shows that storing the digital master record of a film costs much more than storing archival prints. "To store a digital master record of a movie costs about $12,514 a year, versus the $1,059 it costs to keep a conventional film master. Much worse, to keep the enormous swarm of data produced when a picture is 'born digital' -- that is, produced using all-electronic processes, rather than relying wholly or partially on film -- pushes the cost of preservation to $208,569 a year, vastly higher than the $486 it costs to toss the equivalent camera negatives, audio recordings, on-set photographs and annotated scripts of an all-film production into the cold-storage vault."

Perhaps they need to learning about DUPLICATION?

[by Anonymous Coward]

It seems Slashdot could teach them.

It's just a format refresh

[davidwr (791652)]

/. is refreshing their file-storage format to avoid obsolescence. This thread is in the new format. The one from two days ago is in the old format.

You know...

[Omeger (939765)]

DUPLICATION is a lot easier with digital forms of media. I mean, holy crap /., this is probably one of the fastest dupes in the same field of interest I've ever seen.

Re:

[Marcos Eliziario (969923)]

Look at the firehose...

Re:You know...

[sm62704 (957197)]

DUPLICATION is a lot easier with digital forms of media. I mean, holy crap /., this is probably one of the fastest dupes in the same field of interest I've ever seen.

time

[Lord Ender (156273)]

This may be true, but the cost of preserving digital content is halving every year, and can digital content can persist indefinitely; while the cost of preserving film is generally going up, and film can not be preserved forever.

Not really

[ArchieBunker (132337)]

Its all about the storage medium used. You're telling me you want to rely on a hard drive thats been sitting in storage for half a century or film? Film can be restored and if the picture degrades then you stil have something to work with. What happens when you lose bytes here or there in your digital film? Pixelation or loss of a frame all together. Then comes the problem of codecs? Will anyone be able to play a VC-1 file 50 or 100 years from now?

Re:Not really

[orclevegam (940336)]

Then comes the problem of codecs? Will anyone be able to play a VC-1 file 50 or 100 years from now?

They will if you also store the algorithm the codec uses. You can always re-write a codec in the future, so long as you know how it's algorithm and data structures work.

Re:Not really

[PopeRatzo (965947)]

It appears to have finally happened: the ever-shrinking distance between now and nostalgia has finally reached its zero-state. We are now nostalgic for our present.

Maybe we should rethink the importance of preserving popular culture indefinitely in all its pristine digital glory. Why should we spend any money storing the Dukes of Hazzard movie for 100 years, except to fuel the campy nostalgia of future wankers who probably should find something better to do with their time? It's possible that we've already wasted enough time and energy on kitsch.

I mean, it's nice that I can buy a boxed set of all the Francis the Talking Mule films, but I'm pretty sure I could live without it. It's the navel-gazing egotism of this generation that thinks every speck of its cultural exhaust is gold that needs to be protected for future generations.

I'm willing to see society put a few bucks aside to preserve culture, but I think we should wait at least a decade before deciding to go long-term with any given artifact. That would allow us to better vet the material that we're going to keep. Maybe we can have a second and third-tier of stuff that can be saved using a lossy format. I bet it wouldn't cost me more than $200k to keep a divx of the 2005 film Son of the Mask. I'm pretty sure that's plenty good enough to insure that future generations don't miss out on anything.

Re:

[ColdWetDog (752185)]

Why should we spend any money storing the Dukes of Hazzard movie for 100 years ...

Because at the rate we're going in terms of quality (vs. quantity), the "Dukes of Hazzard" may represent a pinnacle of entertainment achievement. A scary thought, but look at what's on the tube today and run that out for a couple of more decades....

Re:Not really

[orclevegam (940336)]

Coming up next "Ow! My Balls!" on the Violence Channel.

Re:

[uradu (10768)]

I assume those ridiculous costs include periodic refreshing of all the data onto new media, and not just the physical cubbyhole to store the drives in. In that case your objection is moot. The great advantage digital storage has is that given proper media maintenance and periodic replication you will have pristine copies indefinitely, something that simply cannot be said of any analog technologies. Given the right equipment, this refreshing and replication process can be automated to such a high degree that

Re:Not really

[paulatz (744216)]

The cost is really ridiculous, releasing the master on bittorrent would be so much cheaper.

Re:

[MightyYar (622222)]

Couldn't you set up a giant datacenter that does nothing but store data in a safe way? The copying would come automatically as you upgrade the datacenter with more storage, and I doubt that the cost would increase too much over time since data constantly becomes more inexpensive to store.

I don't know what kind of data volume we are talking about, but for the $1059/year that it costs to store a film print, Amazon's S3 will store over 588 GB worth of data. For the $12,514 quoted in the article, they could sto

Re:

[MightyYar (622222)]

I think that you could probably safely compress all of the "extra" footage that you don't know whether it is worth keeping or not. Maybe hundreds of hours of the crew talking to one another while re-doing scenes will be interesting to someone down the line, but probably not. And I doubt that they will need the original quality to be intact in any event.

Note that in the days of real film, you wouldn't have that extra footage at all, since film is expensive and they couldn't afford to just keep the cameras ro

Re:

[Captain Splendid (673276)]

What happens when a company goes bust and their codecs are not updated?

That's where OSS comes in. Give it time.

Re:

[zippthorne (748122)]

"What happens when you lose bytes here or there in your digital film?"

Your error-correcting codes do their job and correct the error. They also gives you a tangible warning sign for when it's time to refresh the media: when you no longer get 100% reads (or when the error% exceeds some acceptable threshold that happens to be well below the ecc's max error rate), you move to new media.

And you do ridiculous amounts of parity bits, like O(size of the data) amounts of parity.

If you're really concerned about fut

Re:

[encoderer (1060616)]

exactly.

This "study" is probably from a manager barking off orders to a bean counter:

1. determine how much HD space we need per movie
2. figure out the cost
3. multiply that by a format refresh every 2 years
4. come up with an absurd guess on how expensive it will be to maintain codecs and compatible systems
5. act like this system will have no business utility other than storing archived movies
6. add it all up
7. divide by number of movies sold so we can figure out how much to raise prices, then multiply that n

Re:Not really

[Score Whore (32328)]

You keep saying "store the codecs" which means you're not thinking about this problem in a sane fashion. You don't compress. At all. That's the point of archival. It's not a matter of some geek boy lossfully reencoding his porn collection to fit on CD. It's a matter of keeping the original source material forever. There's no codec here. Just store the data flat with as many bits of precision as you have in your source material. End of story. The only real question is do you store this on spinning disks or stopped disks. Put it on a bunch of hard drives with some parity and error correction codes. Then shut them all down along with the entire infrastructure needed to read the data. Periodically fire it back up to verify that any bit rot that has come along can be corrected and then shut it down again. Every ten years or so, migrate the whole thing to whatever is new in storage. But don't ever compress this shit.

Re:Not really

[CyberLord Seven (525173)]

It's not that simple.

The article is very clear that digital film production creates much more data!

Directors no longer need to husband expensive film stock so they often leave camera rolling while they work out scenes. This is not necessarily garbage footage that can be discarded. Some of this material will be valuable to film historians and also financially valuable as it can be filler for the "extras" that are now included on DVDs.

Digital production creates a much larger set of data that needs to be preserved and updated.

dupe

[Lars T. (470328)]

Dupe [slashdot.org]

Not dupe, rerun.

[JamesTRexx (675890)]

Looks like /. has also been affected by the screenwriters strike. o.O;

Re:

[cgenman (325138)]

It's just their way of preserving the story.

Expensive Duplicates

[mosel-saar-ruwer (732341)]

Yeah, it costs a ton of money in disk space, mirroring, bandwidth, and power bills to maintain all those duplicates of the original [slashdot.org].

Re:Expensive Duplicates

[Falladir (1026636)]

Because nobody wants to expend more than a few dozen gigabytes (at the MOST) on a movie for personal viewing purposes. The task here is to preserve the "originals," the full-resolution, lossless cuts that were filmed on the set. I think I read that the footage that actually appears in Spiderman 3 constitutes 4 TB of information. Consider that a bunch of un-used footage also needs to be saved, and you'll agree that only a few insane enthusiasts would ever be willing to download and preserve that amount of information (at least with technology as it is now).

Re:Expensive Duplicates

[smitty_one_each (243267)]

The obvious way to capitalize on the 4TB of data is to sell a new version every year with slightly lengthened scenes, out-takes, alternate endings, tie-ins, and boxed sets containing the previous and subsequent installments.
By 2015, you'll have "Deluxe Duke Spiderman 3 Power Gold Director's Cut Nukem Forever".
And you will like it.

Re:Expensive Duplicates

[TooMuchToDo (882796)]

By 2015, you'll have "Deluxe Duke Spiderman 3 Power Gold Director's Cut Nukem Forever".

I think you misspelled "Blade Runner, The Final Cut"

I must be missing something here...

[Pojut (1027544)]

How does it cost more to store a bunch of files on a few duplicate hard drives than it does to maintain the facility AND personnel required to keep film negatives in excellent condition? I mean, isn't that one of the advantages to an all-digital film? Everything gets stored as a 0 and a 1, and can easily be duplicated however many times you want with no loss or degradation to the original source?

Someone care to explain why it costs so much to buy a few hard drives?

Re:

[Tsunayoshi (789351)]

Just a complete WAG here by someone who has no knowledge of the exact process:

It is not just buying another drive. Other costs include:
- power for the drive(s)
- power for the server(s) using the drive(s)
- costs of the backup architecture for DR
- costs of cooling the datacenter housing all of the above
- maintenance agreement costs for all of the above
- costs related to the admins who manage all of the above (salary, benefits, etc.)

I am missing quite a few things in there as well, such as off-siting DR copie

Re:I must be missing something here...

[pla (258480)]

It is not just buying another drive
$300/TB, currently.

power for the drive(s)
Approaching zero (minus a few hours per year for making a copy) if you store them offline.

power for the server(s) using the drive(s)
Ditto.

costs of the backup architecture for DR
A minimum-wage drive-jockey and a handful of PCs with EZ-Swap drive cages.

costs of cooling the datacenter housing all of the above
AKA "the dry and somewhat temperature controlled (40-110F) basement of any office building in the world"

maintenance agreement costs for all of the above
See "minimum-wage drive jockey" and add a broom.

costs related to the admins who manage all of the above (salary, benefits, etc.)
See "minimum-wage drive jockey".


And that presumes they use HDDs and make a new copy once a year (keeping a few years as redundant backups and "working" masters)... Although I normally consider tape drives a waste of time and money, in this situation, they seem even more ideal than HDDs. The "handful of PCs" cost goes up, but the cost-per-copy drops drastically.

Even if you replace "minumum-wage drive jockey" with "qualified IT professional or three", I can't see how you'd get anywhere near $12k per year.

Re:

[Tsunayoshi (789351)]

As I responded to someone else who mitigated costs in the same manner:

If your business machine depends guaranteed access to millions of $$ of digital IP, are you going to rely on "minimum-wage drive jockeys" swapping out cheap disks to archive your data?

Re:

[Falladir (1026636)]

The article talks about the reasons that storing a digital movie is expensive, but it doesn't break things down or give us any hard facts. The one truly important question that any reader should be asking (if he's a computer user) is "how many Gigabytes or Terabytes are we talking, here." I think the answer is in this case a few dozen terabytes, and while the cost of storing that much information is kind of meaningful now, it's decreasing continuously. That the article ignores this trend is a serious lap

Re:

[dtolman (688781)]

Your numbers are way off. Actual filming tends to last around 6-10 weeks. But during that time they are filming 10-12 hours per day, with multiple film crews, each with multiple cameras.

Television shows aren't much better. I have a friend who does TV editing - a major complaint he has is that there is dozens of hours of footage for hour long TV shows now - movies are worse. Major motion pictures can have over 200+ hours of footage for a 2 hour movie. Here a few cites I could find with a quick google on "

Re:I must be missing something here...

[Cromac (610264)]

It's to bad there isn't any warning when a drive interface is going to be retired, also to bad it happens so often people can hardly keep up...oh wait...

Even if it did cost a quarter million a year that's still a fraction of the salary the so called "talent" makes for the big movies, there is plenty of money in the movie industry to pay for a datacenter for long term storage of the film.

Maybe the movie industry should hire some people from Google to help them design a large scale redundant storage facility, Google seems to have the entire web cached, adding movies - even at a few TB each - shouldn't be a problem for them.

I have dozens of 20+ year old CD's

[gatkinso (15975)]

that still play just fine.

Just an observation.

unedumicated

[peas_n_carrots (1025360)]

This story must have been written by a journalist clueless in the ways of technology. How does storing a hard drive in a salt mine any more costly than storing a film version? Where does the extra electricity come in? Have one primary version, make a backup (or 2 or 3) and put them in storage. If you're paranoid, verify and/or re-duplicate every few years. The cost of verifying regularly vs reconstructing degraded film should be a wash at worst. It should easily favor the digital versions.

more people who don't know how to properly

[geekoid (135745)]

calculate computer costs.

Not a dupe!

[MiniMike (234881)]

This is just Slashdot's method for reducing storage costs!

Space Reduction?

[davidsyes (765062)]

Sounds like they need Slimfast or Sego....

CelluLOSE in humans is fat?

CelluLOID in film is SLIM

ANY ideas for product names (other than CompressFAST)?

Re:

[avandesande (143899)]

Cellulose is also a name for the fibers found in wood. Film stocks were originally made from cellulose nitrate and then cellulose acetate (much safer).

Another Idea

[avandesande (143899)]

How about they just shitcan everything and spare us another needless re-release?

My God! We've done it all wrong!

[PHAEDRU5 (213667)]

Back to analog, everyone.

Someone needs to come up with a new method

[Mike Buddha (10734)]

Someone who cares ought to come up with a method of transferring digital information to celluloid so that it can be stored with the cheaper storage costs. I'm not talking about a print, but storing binary files on film. A 70mm reel ought to hold a ton of properly formatted digital data and error correction.

Thank God

[edwardpickman (965122)]

That's a relief it's a dupe. For a second there I thought I'd zoned out and it was still two days until Christmas.

WTF??

[IchBinEinPenguin (589252)]

$200K to keep a few bits from rotting?

...just as Hollywood's writers began their walkout.

Oh... that explains it.
It's a conveniently timed report to bolster a negotiating position: "you can't possibly ask for more money, look how much it costs us to store this stuff!!"

Re:

[GregPK (991973)]

I could see the cost exploding if they keep the data in a data wharehouse so that they can actively access it at any given time. However, if they were to put it on laser disc, blue ray, dvd, HDDVD and a hard drive. Then, leave it sitting in a vault they wouldn't have to worry about it. The hard part is still having ready easy access to the original file.

Read the previous /. story

[davidwr (791652)]

Most of the points worth discussing were brought up there.

Re:

[gurps_npc (621217)]

I work with DVDs and storage.

DVD's on average begin to experience decay within TWO YEARS of creation.

While the minimal decay is not noticed by you or me, it is noticeable by machines that copy things.

To obtain the same high end storage with no detectable loss offered by raid storage, you pretty much would need to copy the DVD's every 18 months or so. Expenses for doing this mount up pretty quickly.

Re:

[orclevegam (940336)]

Are you talking DVD-Rs, or pressed DVDs, and what sorts of storage conditions? I would think a pressed DVD inside a nitrogen filled hermetically-sealed safe should be good for a very long time indeed.

Re:... what?

[Chrisje (471362)]

You are obviously not an electronic mass storage professional, because your answer could not possibly be further from the truth. The fact that it got modded 2, Interesting) is interesting in that it proves that people on /. haven't got a clue either.

A hard drive is a mechanical part that will cease to function if the lubrication (I kid you not) goes dry. So sticking a whole bunch of hard drives in a safe for ten years most likely results in you scrapping 8 out of 10 disks for mechanical reasons. Then the magnetic information that is stored on those disks will degrade with time even under perfect conditions. This is why the shelf life of data on an inactive hard drive doesn't surpass 2 years.

DVD's and CD's supposedly should last for 20-100 years depending on whose marketing bullshit you are reading, but in practice up to 15 years is the maximum before the thing starts degrading. Tape suffers, albeit less, from the same ailment hard disks suffer from, even the current batch of LTO-3 and 4 WORM media.

The current generation of MO or UDO drives however use a laser to heat up particular clusters of particles after which it uses a magnet to create the 1 respectively the 0. This means that they are (nigh) impervious to magnetism or heat as long as those two are not combined. MO/UDO is therefore the only medium that will survive for long times on a shelf.

The obvious solution therefore, since HDD's are getting cheaper and bigger, is to stick all that data on active hard-disks, and keeping it alive. Keeping it alive means also having to do backups. All of this requires system administrators. And rules, management, business processes and whatnot, and at the end of the day you will have managed to build an expensive data center. It works, but not as cheaply as putting boxes of film in a basement for 50 years, sorted by title/alphabet.

Obviously, the physical survival of the media is not the only worry, we're also aware of the fact that the .mod file I could play out of my LPT-port-sound-contraption 18 years ago is now useless because mod players and those devices are far from ubiquitous (I found the .mod format converter, but can't find any schematics for that capacitor-LPT-sound-thingy I put together back in the day).

But all that aside, this article is a dupe. And so are the comments claiming it's a dupe. I'm getting a strange sense of Deja-Vu, because it's not the first time I see ignorance on the subject of electronic data management either.