Kodak files for bankruptcy

[Mike McEntire]

This is really bad news. I can't imagine how much this is going to affect our whole industry. I just hope that some companies can keep film alive. I have not been much help lately. Sadly I just threw out about 5 thousand feet of 16mm Kodak film that had been sitting in my fridge for the last 7 years. I of course saved a few cans for paperweights but now I wished I saved it all for other decorative purposes. I built my career on Kodak Super 8 and 16mm film and really wish all the people who worked there the best.

 

[Jeff Kilgroe]

Sorry if this has already been posted. Just came across this. Make sure you get at least to a minute in.

If I recall, that came out in '96, not 2006 as the YouTube text claims. As if the references to PDA's and mini-disc aren't confirmation enough... Didn't go double check. All I know is that was the proper vision for them and they completely lost sight of it in the years to follow. Most everything mentioned in the video like facial-recognition, in-camera automatic color and contrast adjustment, etc.. Stuff that he says "try and patent that!". Kodak holds the patents on much of that. One problem is that conceptual patents don't hold up that well, often not at all, and everyone simply went around them. You have to patent the specific algorithms which make it work and there's a million different ways to recognize a face via software algorithm, and probably just as many that haven't been invented yet.

One can watch this video and think back to the early to mid '90s and how Kodak was pioneering digital photography like no one else. Unfortunately someone steered the ship aground. Sad... I truly feel that the way Kodak was run over the past 15 years is nothing short of criminally negligent. They were the best at one time. They still could be the best if only they recognized their industry for what it was (and at one time they did just that). Unfortunately, they decided that they were a "film" company and not the "imaging" company they needed to be... R.I.P., Kodak. Not sure how you're going to crawl out of this one...

 

[Steve Sherrick]

One interesting aspect of my research was an "OPTOARCHIVE" - an archival alternative for digital masters to protect them for 100+ years (KODAK was actually confident about 300+ years figure) - something that is still missing today. It was ironical to us that the true future of film is to preserve digitally shot and finished movies...

I believe that KODAK will get through this difficult time and emerge with new face and strategy. There is very much still a place for KODAK in our industry and will be there for some time to come. Let's hope they can see it before is too late...

KODAK ain't dead...

Peter

This is what I would love to see them throw a lot of effort at. They have one of the best archival technologies with film. Seeing them develop an advanced archiving solution with life spans like what you're talking about would be incredible. I'd love to see them emerge as a leader again. Even creating advanced media storage systems that could be implemented by the digital cinema camera manufacturers and then trickle down into the consumer realm. In other words, Kodak needs to LEAD not follow. Question is, do they have the capital to do it?

 

[Jim Davis]

Nice find Jeff. It is sad. I grew up with Tri-X. They have been SOOO misguided for so long.

We can only hope that someone can guide them back.


They thought they could walk on water. Just like my buddy here shot on Tri-X, over 30 years ago.

 

[Bill Anderson]

A Codec moment. ( :

 

[Gavin Greenwalt]

And they'd employ 400 people.

Yep, my thoughts exactly.

Welcome the closing noose of digital efficiency. Imagine how many employees Kodak used to have for manufacturing, processing and printing photos. Now you sell someone a 7D and a lightroom and that's that. Even if Kodak released 5 sensors (1 FF, 1 Micro 4/3rds, 1 2/3" point and shoot and one cell phone sensor" and we 5x'ed RED's employees we're looking at 2,000 employees. Kodak has 9x that many employees. Put RED in their shoes and they would have to be selling 140k epics or 500k scarlet packages every year to match Kodak's revenue.

Kodak does make point and shoot cameras. So it's not like they haven't embraced digital. It's just that there isn't as much money in it and the barrier to entry is so much lower. RED is a perfect example. Starting up your own film line... horrendously expensive compared to creating a new digital camera. The RED One was largely off the shelf parts except for the sensor.

 

[Peter Majtan]

This is what I would love to see them throw a lot of effort at. They have one of the best archival technologies with film. Seeing them develop an advanced archiving solution with life spans like what you're talking about would be incredible. I'd love to see them emerge as a leader again. Even creating advanced media storage systems that could be implemented by the digital cinema camera manufacturers and then trickle down into the consumer realm. In other words, Kodak needs to LEAD not follow. Question is, do they have the capital to do it?

Steve - they would not need any massive capital as we already had working prototypes back in 2005/2006. KODAK actually developed a grainless film emulsion for this exact purpose. There are many higher budget movies that had their digital masters recorded back to film separations for archival purposes, but they have stopped shy of what we were trying to achieve - a hybrid digital/analogue solution, whereby the actual digital data were recorder onto film, not just the optical image. I personally think that this project alone on its own as a "separate business unit" would be more then profitable for KODAK...

 

[Gavin Greenwalt]

Steve - they would not need any massive capital as we already had working prototypes back in 2005/2006. KODAK actually developed a grainless film emulsion for this exact purpose. There are many higher budget movies that had their digital masters recorded back to film separations for archival purposes, but they have stopped shy of what we were trying to achieve - a hybrid digital/analogue solution, whereby the actual digital data were recorder onto film, not just the optical image. I personally think that this project alone on its own as a "separate business unit" would be more then profitable for KODAK...

I just can't imagine that being the case. LTO keeps evolving so fast. Sure one film is great but in 4 years when they release it we'll have what 8TB per LTO tape. Redundancy trumps reliability in my opinion.

It would just postpone the inevitable until you have 16TB drives and then 32TB drives... A 4k film is what 3TB? That's one LTO tape today. Soon you'll be able to hold 2 movies in 4k, 4 if you use really mild compression. In 10 years you'll be able fit 50 films onto an LTO tape. How many movies does a studio release in a year? It'll cost $50 to make a copy of their entire year's library. Make 500 of them for the cost of a single optical backup. I would wager that 500 copies are safer than 1-2 optical copies.

And that'll ignore all of the home 4k copies in case everything goes to shit. There'll be millions of those. And pirates will have it on web servers... etc etc... Losing important films isn't a problem in the near future. It would be like losing a CD.

 

[Marc Wielage]

Steve - they would not need any massive capital as we already had working prototypes back in 2005/2006. KODAK actually developed a grainless film emulsion for this exact purpose. There are many higher budget movies that had their digital masters recorded back to film separations for archival purposes, but they have stopped shy of what we were trying to achieve - a hybrid digital/analogue solution, whereby the actual digital data were recorder onto film, not just the optical image. I personally think that this project alone on its own as a "separate business unit" would be more then profitable for KODAK...

Recording digital data on B&W film has been suggested before. The MPAA "Digital Dilemma" white paper report touched on this. There are reasons why even that might be problematic. Issues with MTF and focus, plus bit-depth getting screwed up by noise, are all problems, as is film warpage over time and other age issues. On the other hand, I would trust a piece of film 50 years from now more than I would trust an LTO tape. (To repeat a story I've told before, I once worked on a feature project where a 7-year-old LTO tape had files that were badly corrupted, and I know of another major project backed up to DTF that was disastrous about 9 years later. I think the latter had to be quietly re-done and fixed at a cost of about $500,000.)

Grainless film is not possible in my opinion, because the silver halide particles that capture the image are grain. The key is figuring out how to get the size of the grain down to the point where it's practically invisible. Kodak's most recent 35mm Vision 3 internegative, 5242 (intended mainly for film-out recorders), is probably the film you're thinking of. At some point, I think the noise characteristics of the transfer device and the display device might be worse than the film, depending on exposure. I seem to recall that the effective ISO of films like this is like "10" or something, so we're talking very, very, very slow, fine grain stock.

It would just postpone the inevitable until you have 16TB drives and then 32TB drives... A 4k film is what 3TB? That's one LTO tape today.

The moment the multiple terabyte drives started to come out, I said, "wow! This will be the most inexpensive and easy way ever to lose tons of data at one time!"

The big issue for tons of data is still transfer and copy speeds. Sure, you can theoretically have a 100TB drive, but how long is it going to take to get the material from here to there? How many of you have waited hours to download and/or convert camera files (from any camera, not just Red) to editorial files? This is a huge bottleneck. Even with new developments like USB 3 and Thunderbolt, it takes hours and hours and hours to deal with, say, 4 or 5TB of material per day. And you need multiple copies in different places on different formats for a true, reliable backup.

Film is no different. I know of many studios that struck multiple INs and IPs for big-budget features, then stored them in different vaults around the world. One big one is the underground salt mines in Hutchinson, Kansas -- lots of important audio, video, film, and other materials are down there. The theory is, even in the event of nuclear attack, stuff 1500 feet under solid rock will probably be safe, assuming it doesn't get a direct hit.

 

[Peter Majtan]

Gavin, the main issue with tapes are both the true longevity and reliability of the format and also the actual readers. Try to get a DDS reader these days. Try to get a current gen LTO reader in 25 years - good luck with that...

Marc - sure film has all those drawbacks you've mentioned, but that is precisely the research I have done - to deal with and overcome these issues. And when I say recording data to film - I don't mean just recording the digital images onto film - I'm talking recording actual digital non-image data onto film. The best part of our solution was the fact that you don't have to worry about trying to find a proprietary reader 100 years from now - any film scanner will do - if you can see the film - you can read the data...

PS: the film stock I'm talking about is actually grainless - no kidding. I don't think it has ever been released to general public - and it is not suitable for traditional image acquisition - its sole purpose was archival...

 

[Harald Schossmann]

I agree with Peter. While film is gradually vanishing form the production pipeline there will always be a place for it when it comes to archival. Even if the problem of reliable long-term data storage will be surmounted the question of data compatibility still lingers. Still have that tape deck you used on your Commodore 64 to load some games? Yup same story when you try to read some LTO tapes 50 years from now.

Why has this grainless film stock not been released? Did it never make it past the testing phase?

 

[Gavin Greenwalt]

. Try to get a DDS reader these days. Try to get a current gen LTO reader in 25 years - good luck with that...

That's the weakest argument in favor of optical in my opinion:


http://www.ebay.com/itm/HP-35470-00...e_Internal&hash=item4154d08d45#ht_2817wt_1139
DONE.

Hell let's go back 25 actual years to the IBM 3480....
http://www.ebay.com/itm/IBM-Fujitsu-F6479C-18-36Tracks-200-800Mb-IBM-3480-3490E-100-compatable-/130627086497?_trksid=p3286.m7&_trkparms=algo%3DLVI%26itu%3DUCI%26otn%3D3%26po%3DLVI%26ps%3D63%26clkid%3D5773022462023357946#ht_500wt_1180
Done!

Now find me a working/maintained Arrilaser in 25 years. Good luck with *that* when almost nobody shoots film anymore. I bet you could still do it, but it would take more than 10 seconds and $400.

 

[Tom Lowe]

You snooze, you lose.

 

[Peter Majtan]

That's the weakest argument in favor of optical in my opinion:


http://www.ebay.com/itm/HP-35470-00...e_Internal&hash=item4154d08d45#ht_2817wt_1139
DONE.

Hell let's go back 25 actual years to the IBM 3480....
http://www.ebay.com/itm/IBM-Fujitsu-F6479C-18-36Tracks-200-800Mb-IBM-3480-3490E-100-compatable-/130627086497?_trksid=p3286.m7&_trkparms=algo%3DLVI%26itu%3DUCI%26otn%3D3%26po%3DLVI%26ps%3D63%26clkid%3D5773022462023357946#ht_500wt_1180
Done!

Now find me a working/maintained Arrilaser in 25 years. Good luck with *that* when almost nobody shoots film anymore. I bet you could still do it, but it would take more than 10 seconds and $400.

Right, SCSI. Let's see - where can I get PCIe SCSI card? I am sure if You look hard enough You might find a solution to get a working DDS reader where You can stick your tapes. Now do You honestly believe You could rely on it to actually read the data? What about 15 years from now?
Gavin - don't take me wrong - tapes have their place in medium-term archival, just as HDD's are great for short-term and near-line archival. But for long-term (50+ years) the only option is film. Now it is nothing new for high-production masters to be recorded on separation stocks for archival purposes, but what we were working on was to actually record digital data (not just the images) in a way they could read optically 100 years from now without proprietary system. Who know if ARRI is still going to be around in 100 years (or RED for that matter) - but you can be sure there will be some kind of optical scanner device that could read the data...

Harald - That stock is meant for archival only and I believe it is actually being used for such a purpose. I am sure any film archive will have a way of getting that stock from KODAK. This is just not something You sell on your website, as You can't stick it into a camera and shoot images...

 

[Gavin Greenwalt]

?

http://www.adaptec.com/en-us/products/controllers/hardware/scsi/entry/asc-29320lpe/

And yes I do expect if you took $XX,XXX that it would take to make an optical digital master and made ## Duplicates at $100 each you would amongst them have a working copy in 50+ years. But you're looking at it wrong. You don't let a tape sit on a shelf for 100 years.

In 10 years you'll be able to fit 50-100 films on one tape. A studio like MGM has what 5,000 films? That's 50-100 tapes. Every decade you pay someone for a month's worth of copying from old tapes to new ones. Over 100 years that's 1 year's wages for one person. A very very tiny expense, smaller than the shear volume of archive space you would have to rent for optical masters. So in 2022 you hire an IT guy to make sure the robotic LTO machine successfully creates your 100 tape masters and does that 10 times. 1,000 tapes. In 2032 you do it again only this time you need 5 tapes x10 copies 50 tapes. In 2042 you set aside a little time to copy them all to one tape 50 times just to be sure. Every decade you hire a guy to check in on your 50 tapes and replace them. In 50 years you can also have a mountain of USB drives which can hold every single film.

And that's *worst* case scenario. In all likelihood the majority of every single generation of tape generation will be fine even in 20+ years. And you'll have dozens of copies.

 

[Marc Wielage]

And when I say recording data to film - I don't mean just recording the digital images onto film - I'm talking recording actual digital non-image data onto film.

I think that's what I said. Again, resolving little teeny-tiny bits, especially when there's 52MB per frame in a 4k image, is gonna be tough in a 35mm frame only 1.4" x 1" (roughly a full-ap 35mm frame). The good news is, creating a 35mm transport to play back the data would be pretty easy: sprockets, pull-down claws, gates, and all those mechanical parts are relatively easy to make. I think getting a hard drive or an LTO tape to play back in 50 years will be a lot harder.

I don't know why Kodak didn't pursue the idea of using their films to optically store data, because it was certainly possible. I have a friend who actually had the same idea more than 20 years ago, using it for digital audio storage, but he was never able to get it going. That would've been very easy, given that there's only a fraction as much data every 24th of a second in sound compared to picture.

As far as I know, the studios are still using regular 4K images going to negative stock for archival purposes. There are a few studios that strike YCM protection negatives, sometimes doing a frame for Red, the next frame for Green, and the next frame for Blue, then marry them all together later on. This would have three times the pixel density of a regular piece of film, assuming you can line up the images and not have any issues with grain. This also works better than having 3 completely separate pieces of film, the way it was done in the old days.

In 10 years you'll be able to fit 50-100 films on one tape. A studio like MGM has what 5,000 films? That's 50-100 tapes. Every decade you pay someone for a month's worth of copying from old tapes to new ones. Over 100 years that's 1 year's wages for one person.

Have you ever used an LTO-5? Do you know how long it takes to put 1.5TB on a tape? It's 140Mb per second. Do the math -- it takes a long, long, long time... roughly 4 hours, if you don't verify the data. And that's not even a full reel of uncompressed 4K data.

Don't forget it's not just the number of movies; it's the number of versions of movies. I once visited the MGM vault in LA and saw all the film and sound archives for Gone with the Wind. I think there were at least 20 shelves, floor to ceiling, stretching about 30 feet, with a couple of thousand rolls of films -- Japanese versions, subtitles, TV versions, negatives, prints, outtakes, screen tests, documentaries, optical tracks, mag tracks, mono mixes, stereo mixes, titles, optical effects... it was a stunning amount of film, just for one 3-hour movie. Multiply that times thousands of movies, and then consider that most of these images have not even been scanned to data yet. Modern movies have 3D versions, flat versions, versions for different color space, long versions, short versions, censored versions, unrated versions, TV versions, textless versions... it goes on and on.

This is a huge problem. And all these movies are falling apart. I've said it before, but it's fairly well known that at least half of the movies made prior to 1950 are lost or deteriorated to the point where they are no longer complete. Two decades after that, I bet half of the TV shows ever broadcast are gone. Archiving is not an easy or simple business.

 

[Peter Majtan]

...To repeat a story I've told before, I once worked on a feature project where a 7-year-old LTO tape had files that were badly corrupted, and I know of another major project backed up to DTF that was disastrous about 9 years later. I think the latter had to be quietly re-done and fixed at a cost of about $500,000...

I rest my case...

Gavin - once again - tapes DO have a place in MEDIUM-TERM (10~50 years) archival applications. But You are completely missing the logic behind LONG-TERM (50+ years) archival. The same logic You use for defending tapes I can use against you with HDD's. Why bother with expensive LTO writers/readers where SATA docks are $20,- USD, not to mention the cost and (ever-increasing) capacity of HDD's. The same way You can have an IT dude copying data over to larger HDD's every 2~3 years or so - why bother with tapes...

Gavin - I don't want to pick a fight - all I am saying is that there is a purpose for film, but ironically it is not at the acquisition stage (beginning of a film project), but rather at the archival end. This thread is about KODAK and film, let's keep it on topic...

Peter

 

[Jeff Kilgroe]

Right, SCSI. Let's see - where can I get PCIe SCSI card? I am sure if You look hard enough You might find a solution to get a working DDS reader where You can stick your tapes. Now do You honestly believe You could rely on it to actually read the data? What about 15 years from now?

Er... PCIe SCSI cards are plentiful. Adaptec, ATTO, SIIG, Symbios, 3Ware, LSI and more all still make them. The whole "how do you read this in 25 or 50 years from now" argument is very weak as Gavin points out. LTO is so widespread, the odds of finding a way to read them after a global apocalypse is pretty darn high.

Gavin - don't take me wrong - tapes have their place in medium-term archival, just as HDD's are great for short-term and near-line archival. But for long-term (50+ years) the only option is film. Now it is nothing new for high-production masters to be recorded on separation stocks for archival purposes, but what we were working on was to actually record digital data (not just the images) in a way they could read optically 100 years from now without proprietary system. Who know if ARRI is still going to be around in 100 years (or RED for that matter) - but you can be sure there will be some kind of optical scanner device that could read the data...

Now this is where I have to point out that digital archives are not a solution, they are a process. Backing up is not a choice of format, but a strategy. Your data archives must evolve over time, there isn't any medium out there that can sit for 50+ years without risk of degradation or corruption. AND THAT INCLUDES FILM! Just look at the restoration work that goes into most digital masters of films shot in the 70's and 80's. Look at what was done to the original Star Wars films just to make the 2K masters for the Special Edition and subsequent DVD releases.

With digital, you have the opportunity to preserve the original data forever and it can transcend multiple formats and archival generations. That is the strength of digital. The fact that one form of media may not have a shelf life of over 35 or 50 years is irrelevant as digital allows us to move beyond that and find reliability in replication and redundancy across multiple generations of media. To be kinda blunt about it, anyone who envisions digital archives as a write the media and shelve it scenario is missing the whole point. Anyone who in 20 years from now is trying to hunt down a tape mechanism to read LTO-1 tapes is doing nothing less than trying to fix a major fuck-up.

If you print something to film, it is an analog representation of your original. Each subsequent copy is one generation removed from that which it was copied from. Even in its most ideal scenario, film slowly decays over time and will eventually fade away. Sure, that won't happen to any film prints made in our lifetime, but even in the most controlled environments and on the latest formulations of archival stock, the decay will be readily apparent in 100 years or so. I'm not going to argue that film is not a good archival medium, because it is good for archival use. But let's be realistic here and admit its faults. Even Kodak and Fuji admit these faults. And moving forward, since most delivered studio product is now digitally processed and has a digital master from which to make the film prints, as well as digital deliverables and archives, I see no reason why anyone would pass on making that digital archive. Furthermore, the digital archive is the only means by which to preserve the final mastered production at 100% quality. Film will get close, in an analog transcription, but it's a slow descent from that point for film, no matter what is done.

 

[Tim Whitcomb]

Good points but film is STILL the ONLY medium that all you have to do is hold it up to some light to see the frames. thats why it will rule supreme as the BEST long term archive medium - just my 2 cents. nothing else can just be held up and viewed decades from now. save film

 

[Craig Parkes]

History has shown the best form of archive is often the one that accidentally survives, not the one that is planned and managed. Libraries, museums and storage facilities can become neglected or abandoned, burnt down, get levelled by natural disasters or bombed/ransacked in wars. Securely kept digital copies over a long enough life span are less likely to survive for this reason then unsecurely kept mass distributed digital copies.

Which brings up an interesting point about archival and copyright duration. No digital archival in terms of corporately manufactured products in the digital world should require archival of more than 125 years under current us copyright law, as that is the expiry of corporate copyright.

Basrd in the way storage space is going, if that laws was to stay in place as is (it won't, it will be extended due to corporate lobbying I am sure...) than if something was wanted to be kpet for posterity for humanity, rather than just commercial interests, then simply distributing all the original digital materials on mass and freely under a completely open public domain license would ensure it's survival more than anything else.

It's interesting because in a digital world copyright works under the reverse model to art in an analogue/physical sense.

Physical art can gain value over time due to scarcity, copyrighted digital works will inheritly lose value over time as they approach infinite abundance.