Filmprint "Backup" ?

[Jarek Zabczynski]

So I've been meaning to bring this up and get some thoughts on this.

I got to spend some time with director John Landis recently and we spent a lot of time talking about the transition to digital vs film. He's on the Alexa camp for now, but that's not the point. He made a pretty big deal about the importance of having a filmprint made of your film even if it was shot 100% digital. Saying it's the only way to make sure you have a "physical" copy. He made a lot of sense. He made a comment that digital files also "rot" in their own way. I guess he means file corruption and lost data and whatnot. Yes, we make sure we have our files backed up in multiple places, but does our media ever actually "exist" unless we can hold it in our hands? If you don't have a computer or digital playback system you don't have anything do you? You can hold your filmstrip up to a light and "see" your film right in front of you. He said that for this very reason alone, film will not go away. That it is a key aspect of preserving our films in the digital age and into the future and that many people are overlooking this critical part of it.

 

[Andy Jarosz]

A post I made a while ago...

I had this sweet idea once, you could store images on sheets of paper by just writing the hex color code of each pixel. Then you could keep your images for darn near forever, and re-digitize with freely available OCR software.

And then I realized just how much paper it would take to print millions and millions of characters. A second of 1080p would be over 7,000 sheets of 8.5x11. And cost about $100.

 

[Phil Holland]

More of the same out here in LA LA land. Master negatives and prints are being made fairly regularly of digital films and filmy films.

I don't think it's being overlooked at all. It's been going on for a long time. It will however become more expensive and difficult to do.

 

[Uli Plank]

They won't last very long regarding color. The only way they can be much more stable than digital is three separate B/W prints from each primary color.

 

[James Press]

Multiple LTO tapes in multiple locations is surely a lot more practical...

 

[Jarek Zabczynski]

I think he also meant preserving beyond our time. Without modern technology our digital files and media is useless, whereas a physical optical copy is relatively easy to figure out. The same way a record might outlast a CD or mp3 files on a disk.

 

[Björn Benckert]

The flintstones could not guessed that we would have the carbon 14 method today... So I think tomorrow they read our old broken 14tb hard drives like a post it note.... so I'm not so worried.
I also been in business of resorting old film.. like 15 years old... and that is believe it or not that is usually in very poor condition and "rotten" is just the first thing that comes to mind.. The celluloid getting stiff changing length, running hard on the perf, scratches etc... So I do not think film is the most logical way to store it for the future.

If kept on multiple drives I think you are best set for the future. I do not believe in tapes, we been trough quite a few formats there and the decks dies, and the tapes are getting older / are more likely to fail than drives are my experience.

 

[Patrick Tresch]

Recycle it all!

Just kidding.

I wonder why the mordern humanity became so egocentric to belive that everything we do is worth been kept for "eternity"?

Pat

 

[Tim Watson]

Recycle it all!

Just kidding.

I wonder why the mordern humanity became so egocentric to belive that everything we do is worth been kept for "eternity"?

Pat

Art being the only thing that lasts through the ages....(or something people care enough about to preserve)

ironically mankind feels the need to have a total onslaught of destruction towards knowledge and science every now and then...

Just making the comment really.

 

[Patrick Tresch]

Art or stones?

What are the keys for our "modern civilisation" to understand their Art?

Pat

PS: use LTO ;-)

 

[David Mullen ASC]

http://www.wilhelm-research.com/pdf/HW_Book_20_of_20_HiRes_v1a.pdf

Stored at ideal temperature and humidity, it takes well over 100 years now for the least stable color layer (yellow) on modern MP color negative and inter positive film to fade more than 10%. Obviously b&w separations have even more stability. Either way, you are talking about more stability than current digital storage physical media.

No one is talking about only storing movies on film, but it never hurts to store things in multiple formats in multiple locations. Many things have only survived because multiple copies were made. My viewpoint is that rather than debate whether film or digital is better for long-term archiving, use both.

 

[Todd T]

Hasn't anyone created a printer that, for instance, prints high-density barcodes on film stock so that we can preserve the native data files?

 

[Jeff Kilgroe]

I find the common fallacy people have when considering digital archive methods is they try to approach it from a "back it up and forget it" point of view. Current digital media, even the best there is, only has a shelf life of about 30 years, possibly as much as 50 years. Of course, none of that is really tested until those timeframes pass. Digital information requires regular maintenance and upkeep. If done properly, it can be sustained indefinitely and will suffer no generational loss or degradation. One can argue (and I often have) that the greatest weakness of digital archives, which is that need for continuous upkeep and reproduction or migration, can in fact enable its greatest strength -- the ability to be infinitely replicated and distributed with no alteration or loss of quality.

I can't argue against making a film backup of your digital production. One can never argue with an added layer of assurance. And to further add to this, the digital and analog archives can reinforce one another through time. As the analog film prints begin to decay (and they do), new ones can be generated from the digital archives.

 

[David Mullen ASC]

Well, digital soundtracks are obviously copied onto a film print, but you're asking about storing data as an image of black silver grains stuck to a strip of plastic (Estar). I'm sure it can be done but the question is whether it would be any more machine-readable than sticking the data etched onto plastic disks, such as on CD's, and any more stable.

Jeff is correct, the only sane and reliable method of long-term data archiving involves periodic cloning to new physical media, etc. The trouble is that a lot of private archives are poorly funded and minimally staffed, so whatever systems are used have to take that into account.

 

[Björn Benckert]

http://www.wilhelm-research.com/pdf/HW_Book_20_of_20_HiRes_v1a.pdf

Stored at ideal temperature and humidity, it takes well over 100 years now for the least stable color layer (yellow) on modern MP color negative and inter positive film to fade more than 10%. Obviously b&w separations have even more stability. Either way, you are talking about more stability than current digital storage physical media.

No one is talking about only storing movies on film, but it never hurts to store things in multiple formats in multiple locations. Many things have only survived because multiple copies were made. My viewpoint is that rather than debate whether film or digital is better for long-term archiving, use both.

Even so Mr Mullen, but the money spent on keeping those copies on ice is probably higher than making one digital dub to a new digital format each year of those 100 years. So after 100 years you would have 100 of digital copies on 100 different hard drives. That would be more safe than having one frozen film strip that has faded about 10%...

... I actually have restored some old Ingmar Bergmans films about 10 years back that where telecined back to uncompressed 8bit 2k. They had all been in the freezer but for one reason or the other they all had quite a bit more than 10% loss...

Also, you have to take into account that going from digital to film is quite a loss. Then going back tot digital is also a loss. just the transfer is probably more than 10% and staying on film introduce problems as picture instability, flicker, color inaccuracy gate weave etc.

I think that if Kodak had this one nailed down they would sell celluloid backups like hotcakes but as it looks the market is not really taking that route.

I do not know for sure what your digital final master copy including sound weights. But I would estimate bellow 30 Tb. All ready today that amount of storage is far less cost then a negative print.
If 5 years the digital storage costs has probably fallen down to a 1/4 of todays prices so to make perfect digital clones in the future will probably be a lot cheaper than today.

I also think when RED introduce R3D as a mastering format then you can have your last 3 films laying around in the desktop folder of your laptop

 

[David Mullen ASC]

Keep in mind I said "modern" color negative stocks -- pre-1980 color negative films had a lot poorer dye stability.

Many studio features are archived on film, but Kodak isn't going to get rich supporting that considering the average feature is only 10,000' long!

Of course film storage is also a flawed process and you cannot physically copy a piece of film with no image loss (though a digital scan at a higher resolution than the original is pretty close to creating a lossless copy.)

The thing with cold storage is that it is fairly low-maintenance compared to having someone go through a digital archive and periodically back things up, not that that couldn't be automated but it generally isn't today. The truth is that most things get vaulted and left alone for decades, so whatever method is used had better factor that in.

 

[Todd T]

Longevity of the media is my point. As was pointed out earlier, film stock has a well known aging characteristics and very long shelf life, recordable CD/DVD media we are finding out, not so much. hard drives even less.

Optical barcodes on film are as easily readable as any frame in "archived" analog film. A modified film scanner can directly read the data from the barcode in each frame, or and unmodified film scanner will save them all to a file that would be trivial to process on even the most modest computers.

 

[Todd T]

Barcodes are currently printed on film; how much data could theoretically be stored if the entire frame had barcodes like this?

 

[Jeff Kilgroe]

Longevity of the media is my point. As was pointed out earlier, film stock has a well known aging characteristics and very long shelf life, recordable CD/DVD media we are finding out, not so much. hard drives even less.

Keep in mind that those CD/DVD, etc. optical discs -- the writable ones -- are based on the very same emulsive film processes that are used in film stocks. Kodak was one of the leading pioneers of optical media back in the day. Practically invented the writable disc. Oh, how the mighty have fallen...

Optical barcodes on film are as easily readable as any frame in "archived" analog film. A modified film scanner can directly read the data from the barcode in each frame, or and unmodified film scanner will save them all to a file that would be trivial to process on even the most modest computers.

Optical barcodes or other data patterns on film don't really help us here. They are prone to loss and corruption the same as any other area on a film print is. No better/worse than an optical disc. As soon as data density increases to a certain point, there is no longer a means by which to escape the organic transitional shift and decay of the film itself. If we were to encode a high-quality and highly compressed (via wavelets or fractals) 4K frame into a film print via a computer-readable data pattern, you're looking at 3 to 7 MB of data per frame. The best film stock has a median grain size of around 5 microns... If we encode to 35mm film using 36x24 frame dimensions, that gives us 7200 x 4800 grains. We could assign more than a binary value to them based on their emulsified "exposure" value, perhaps even go as far as squeezing a 16bit value in each. Or enough shades of grey to represent values from 0 to 65535. That sounds fine and great... That's 34.56 Million grains at 16bits each. Then again, it's not so great, because that's an equivalent of 69 Megabytes. We can already encode 1000 times that amount of information directly onto silicon or store it electrostatically on NAND flash or magnetically on hard drive platters in an area that is 1/10th the size.

Reality check... the above film example isn't going to work. There is nothing built into those numbers to account for the decay and mutation that occurs over time in the emulsive process. The grain sizes are far too small and individualized to hold up to these forces such as dye shift, bleed and fade. What may account for a visual loss of 10% in maybe 25 years would be catastrophic and account for a near total loss on a binary data level.

This is where I think that with the current state of technology, both analog and digital, as it pertains to archival process, there is no one right answer. All the available technologies have flaws, so it's best to be diverse in how, when and where we archive our information. Digital archives are probably the most robust at this point due to the replication and dispersion ability. The ability to do so without incurring any generational loss. But they suffer in that they need to be maintained. Film requires maintenance, but less so. Unfortunately it decays and all we can do is slow that process somewhat. And any further replication from an analog source will always show generational loss and loss of detail compared to the original. I think for now, these analog and digital technologies can offer the most if used together. Ultimately, I see digital as the winning method. It's just that the technology to make it so has yet to be developed.

I don't buy into all the apocalyptic scenario reasons often spouted against digital archives. Sure, they are not human-eye-readable as a film strip is, but film is not readable in its entirety. The audio is still encoded and it still takes mechanical means to present the film in proper potion. If our civilization is to undergo an apocalyptic tragedy, then the odds of there being anyone around to properly collect and maintain these film archives is pretty unlikely. If there are, then chances are there will be means by which to retrieve the digital archive contents as well. In the grand scheme of things, none of us can say whether our civilization will survive the next century or millennium. What if we don't? Who's going to view the film then? What if it's another 10,000 years before another capable intelligence comes along and by then, the film is nothing more than dust? 100 year lifespan on film stock is pretty meaningless, IMO. It's not long enough to transcend multiple human generations without significant decay. It's nothing more than a pico-second blip on the cosmic radar. If human civilization is to endure, we will eventually have to spread from Earth and take our knowledge and history with us. Digital mediums will make this possible, analog will not.

 

[Todd T]

Sorry Jeff, I moved my example.

I agree, nothing will last for millennium. I am just postulating what the most effective use of current technology/ease of recovery/production and storage cost and workflow.
Barcodes are designed to account for degradation in media and they don't have to printed at the highest possible resolution at the expense of using more media.
Optical media, while not a dense as silicon or magnetic media, does allow for easier recovery in the future. I still have an 8-inch Floppy disk and several jaz cartridges that I have no way to read.

Leapfrogging from one media format to the the next is easier for individuals than dedicated archival storage units or buildings, but reverse is true for businesses with large archives.