This is topic Scanning film @4K for archiving, reissuing, and ultimate home theater in forum Film Handlers' Forum at Film-Tech Forum ARCHIVE.
To visit this topic, use this URL:
https://ft-forum.com/ft/cgi-bin/ubb/ultimatebb.cgi?ubb=get_topic;f=1;t=006081
Posted by Mark J. Marshall (Member # 1409) on 04-19-2004, 08:07 PM:
Interesting article over at the New York Times website.
Original Story Here Text Below....
600 Macs, 4,000 Lines, One Giant Leap for DVD's
By FRED KAPLAN
Published: April 18, 2004
On the second floor of an unassuming office building on the edge of Burbank, John Lowry is forging what might be the future of the DVD — and, with it, the way that classic films will be stored, preserved, telecast and watched.
Mr. Lowry, who has worked for decades at enhancing video imagery, is responsible for some of the best-looking DVD restorations in recent years, including transfers of "Casablanca," "Singin' in the Rain" and "Once Upon a Time in the West."
Since last November, he has been immersed in a project that promises to advance the state of the art — and that has been kept secret, even among most industry insiders, until now.
What he is doing will make a DVD look nearly as sharp and detailed as a 35-millimeter film print. It will produce images with six times the resolution of today's high-definition television sets. In video quality, it could turn home theater into a true rival of the neighborhood cineplex.
Walk into the suites of Lowry Digital, the company that Mr. Lowry started six years ago, and the first sight that strikes you is the computer bank — rack after rack of Macintosh G5 computers, 600 of them, holding a combined memory of 2,400 gigabytes.
Beyond this room is a super-sanitized, temperature-controlled chamber. Inside, a technician wearing a white smock and cap monitors a pair of machines called the Imager XE-Advanced, made by the Imagica Corp.
The Imagica machines are ultra-sophisticated digital film-scanners. They are loaded with reels from the original negative of the 1967 James Bond movie "You Only Live Twice."
The spools advance slowly, one frame every four seconds, which is how long it takes the Imagica to scan across a frame 4,000 times — a process known as 4K scanning.
During the scan, the machine creates a digital replica of the frame, consisting of 4,000 horizontal lines of data. A cable then transmits this data to a hard-drive server in an adjoining room.
To put the magnitude of 4,000 lines in perspective, a television displays broadcast signals as 480 lines. High-definition televisions have up to 1,080 lines. (The greater number of lines, the more detailed the image — the more closely it resembles a seamless, lifelike picture.) Impressive as HDTV looks, 35-millimeter film has far more color and detail. Engineers calculate that 4,000 lines of data would be needed to reproduce all the visual information in a frame of film — exactly as many lines as the Imagica delivers.
So, if it scans an original camera negative, as it's doing with "You Only Live Twice," it creates a data file that's a virtual duplicate of the negative.
By contrast, most DVD's these days — good as many look — begin with a compromise: they're scanned at just 1,080 lines, at most 2,000 (sometimes as few as 480), and the source is almost always not the original negative but a copy. When you start with a copy, Mr. Lowry said, "you're immediately losing lots of details. Colors are less pure, too."
In other words, a DVD that's scanned at 4K from an original negative should look better than the best DVD's today. That's the theory, anyway.
MGM has hired Lowry Digital to make 4K digital masters of nine James Bond films, including all of those starring Sean Connery. I have watched scenes from a high-definition transfer of these masters on monitors at Lowry Digital. I've also seen a DVD, which Mr. Lowry gave me, on my TV set at home.
The scenes look as brilliant as anything I've seen on a video disc — and better than any video of a color movie that was shot 35 to 40 years ago. Colors are saturated and natural. Gardens have dozens of shades of green. Flesh tones are uncannily lifelike. Shadows look like shadows, not gray blots. Motions are smooth, not jumpy.
MGM executives decline to say when they'll be releasing these Bond DVD's — or anything else about the project, except to confirm that it exists. The new discs won't be out until next year at least — perhaps in part to avoid angering consumers who bought the 20 Bond films in three boxed sets that MGM put out just last year. Those discs (which Mr. Lowry had nothing to do with) tend to look grainy, blotched and flat.
They were made with the materials at hand — faded film stock and high-definition (sometimes standard-definition) scanners. Studios frequently use 4K scanners for computer animation and special effects, but few have even considered 4K-scanning of entire movies for DVD. It's an expensive operation. An Imagica scanner costs about $300,000. The G5 computers cost $3,000 apiece. The software, servers and so forth aren't cheap either. All told, mastering a DVD in 4K costs two to four times as much as doing it the usual ways.
The attraction of going this route is that it produces not just better-looking DVD's for now but a foundation for formats of the future.
In its current format, an image on a DVD consists of about 500 lines — no more than a DVD player can "read" or a regular TV set can display. Even a 4K master has to be reprocessed to fit the DVD format. (It will still look better, just as a photo taken by a Nikon looks better than one taken by a disposable camera, even on newsprint.)
Some time in the next decade, though, DVD's will probably be supplanted by high-definition DVD's. High-definition TV may evolve into ultra-HDTV. (Engineers in Japan have built prototype discs and monitors that display 4,000 lines.)
A hard-drive master that contains a 4K scan could serve the same function as a film negative — a source of copies, for whatever medium. Unlike film, it won't fade; and as video technology improves — as its resolution becomes higher and higher — there will be no need to make new masters; 4K is high enough to accommodate the changes.
"We're making an archive — for DVD, film, digital cinema, HDTV, TV, whatever — that will last the next two or three generations of technology," Mr. Lowry said.
Thirty-five years ago, Mr. Lowry, who is now 71, patented a method of cleaning up NASA's live televised transmissions from the moon. Six years ago, as the DVD took off, he set up Lowry Digital — then a two-man R & D shop — to apply his techniques to digital restoration.
He hired a photographer to make a short 35-millimeter film clip of some children playing soccer on a lakeshore. He paid a local lab to transfer the film to digital video, using a 4K scanner. The picture was clear, sharp, detailed. He then processed the images with his film-restoration software, which he'd programmed onto some Macintosh G4 computers. (The effort took months, as the faster G5's weren't out yet.) The processed picture was clearer, sharper and more detailed still. He could see every divot on the turf. What had once looked like a smudge in the background was now recognizable as a boat on the lake.
In January 2000, some executives from Warner Brothers saw his demo. They were so impressed, they faxed him an order the same day to restore the masters for three DVD's: "Gone With the Wind," "Now Voyager" and "North by Northwest." With the advance money, he bought the computers he needed to do the job.
Since then, he has bought hundreds of computers, hired a staff of 30 and worked on 80 DVD's — including the long-awaited DVD of "Star Wars" — erasing wear, tears, dirt, scratches and other ravages of age. (In the early days, he sometimes erased too much. By his own admission, his restoration of "Citizen Kane" is too clean; the natural grain of film is gone; it looks like a video. He later figured out how to fix flaws while preserving grain.)
Many restoration specialists do this sort of work manually, often frame by frame. Mr. Lowry may be the first to do digital restoration while the images are still in a digital format — a bunch of 0's and 1's inside a computer — before they're transferred to video.
Not long ago, MGM sent a camera crew to interview him about his restoration techniques for a "special feature" to be included on the Bond DVD's. The question now is how long it will take MGM to release these DVD's and how long it will take other studios and digital-mastering houses to catch on — in short, how long it will take for the future to begin.
Posted by Joe Redifer (Member # 3) on 04-19-2004, 08:15 PM:
600 Macs and only 2,400 Gigabytes? That doesn't make any sense. The crappiest Mac is a laptop that comes with a minimum of 30GB of hard drive space. That'd be about 18,000 GBs if he used 600 low end laptops. But no. He is using G5's. The crappiest G5 comes with a minimum 80 GB hard drive standard. You can't go any lower at all. Multiplied by 600 and you get 48,000 GB. Whoever wrote the article is obviously a dumbass. His family should be kidnapped. I refuse to read the rest of the article as it is surely plagued with tons more idiot mistakes like that, and is therefore not relevant at all.[av=d]
Unless of course he is talking about RAM. Then the article might hold some truth. But it was not made clear. Bad writer!
Posted by Daryl C. W. O'Shea (Member # 1303) on 04-19-2004, 08:19 PM:
Whoever wrote the last post is obviously a dumbass.
The article actually said, 2400GB of memory. Which is most likely correct since early G5 mainboards had two PC2700 (DDR333) slots with a total capacity of 4GB. Multiply that by 600 and you generally get 2400GB.
edit for Joe's edit: They did say memory... which is RAM. Not storage, which'd be a hard disk drive or other media. It was clear. Bad reader!
Posted by Joe Redifer (Member # 3) on 04-19-2004, 08:22 PM:
If he's using early G5's, I feel sorry for him. The G5's that matter can all be expanded to 8GB of RAM. If he was serious about his project, he'd replace all 600 computers immediately![av=g]
Edit for Daryl's edit to my edit: maybe for computer geeks like you, but not to people who have a life like me!
Posted by Daryl C. W. O'Shea (Member # 1303) on 04-19-2004, 08:29 PM:
It'd be so much faster with the new board's (four) PC3200 slots too!
Actually, it likely wouldn't make a difference since memory size and bandwidth probably aren't going to be the limiting factors in this application. Rather how fast you can push the data through the processor 64 bits at a time.
You can't cater to every person's incorrect understanding of well established naming conventions!
Posted by Bruce Hansen (Member # 281) on 04-19-2004, 08:51 PM:
Whoever wrote it is confusing lines of resolution, and scan lines. He also does not understand much about scanning.
Posted by Adam Martin (Member # 641) on 04-19-2004, 08:52 PM:
I stopped reading here:
quote: From some poor journalist who doesn't know his subject:
how long it takes the Imagica to scan across a frame 4,000 times — a process known as 4K scanning.
Posted by Daryl C. W. O'Shea (Member # 1303) on 04-19-2004, 08:57 PM:
Although not very descriptive, it's sort of right. Each line is scanned laterally, across the frame. Of course 4K scanning denotes 4096 horizontal pixels, not lines.
quote: IMAGICA INTRODUCES IMAGER XE – ADVANCED – DIGITAL FILM SCANNER
September 08, 2003
IMAGICA INTRODUCES IMAGER XE – ADVANCED – DIGITAL FILM SCANNER
Newest scanner offers faster speed and quality. Perfect for digital intermediate process
September 8th, 2003...IMAGICA Corp., Tokyo, Japan, the leader in digital film scanners,
introduced today a new digital film scanner, the IMAGER XE - ADVANCED. IMAGER XE -
ADVANCED is the newest generation in IMAGICA’s product line of digital film scanners.
The IMAGER XE -ADVANCED, designed as a true scanner, scans 4096 pixels across a 35mm
Academy frame and provides Cineon format with full RGB color space(14bit A/D), as well
as consistent color and density, frame after frame.
The IMAGER XE -ADVANCED scans 2seconds/frame for 2K and 4seconds/frame for 4K
resolution, significantly faster from the current IMAGER XE®. Differences from the
current generation IMAGER XE® include a new 5K- camera system, new internal processing
system, and the network is Gigabit-Ether. In addition to 35mm, the IMAGER XE -ADVANCED
also supports 16mm and Super 16. In addition, an 8K-camera system is available as an
option. The 4K image is sub-sampled from 8K.
There is an upgrade path for current IMAGER XE® customers who would like to upgrade to
the IMAGER XE -ADVANCED.
IMAGICA is continually advancing its scanner technology to meet the rigorous demands
required by post-production facilities around the globe, and provides an upgrade path
to preserve the investment by these facilities. The new IMAGER XE -ADVANCED is the
latest example of IMAGICA’s commitment to the marketplace Renowned for its customer
service and superior technology, IMAGICA is indeed the market leader in the post-
production community for digital film scanners.
Mr. Makoto Tsukada, Director of the Engineering Solutions Division at IMAGICA, says,
“There are many facilities that require both speed and quality, and don’t want to
compromised with scanners of lesser quality and performance from other manufacturers.
These facilities also want the peace of mind knowing that they’re investing in
technology provided by a well-established company who will continually improve its
scanner products far into the future. IMAGICA has been providing digital film scanners
to this market for many years. We’re experts in this field, not newcomers, and we
provide unparalleled support and service, which is why facilities continually invest
in our scanners to meet their scanning needs”.
IMAGICA scanners have been used on such films as Lord of the Rings: Fellowship of
the Ring and The Two Towers, Spiderman, Stuart Little 2, Harry Potter and the
Sorcerer's Stone, Cats & Dogs, Enemy at the Gates, The Planet of the Apes, Scary
Movie 2, Cast Away, and many others.For further information, contact:
U.S., Canada and South America Sales
IMAGICA Corp. of America
5301 Beethoven Street, Suite 109,
Los Angeles, CA 90066, USA
tel: 310-306-4180 fax: 310-306-2334
e-mail: sales@imagica-la.com
Europe, Asia and Oceania Sales
IMAGICA Corp.
2-14-1, Higashi Gotanda,
Shinagawa-ku, Tokyo, 141-0022, Japan
tel: +81-3-3280-1119 fax: +81-3-3280-7527
e-mail: products@lab.imagica.co.jp
###
All product names are trademarks of their respective companies.
September 8, 2003 11:29 AM
Posted by Adam Martin (Member # 641) on 04-19-2004, 08:59 PM:
I realized that as I was typing. But a casual reading comes across as "the frame is scanned 4000 times".
Posted by Bobby Henderson (Member # 840) on 04-19-2004, 10:39 PM:
Here's the thing I think it stupid out of this deal: that somehow those 4K scans will be a "giant leap for DVDs." Is the guy on drugs or what?
You lose lots of detail when down-converting something mastered to HD video over to 480p. Just how is a 4K source supposed to show any more detail in 480p than HD video down-converted to 480p? Because it is "digital"? Idiots. They're not going to see any apparent benefits of those 4K scan until video storage formats catch up with that resolution.
4K might be good for some archival processes. Although moves should be made in the future for better standards like 6K or 8K scanning when computing power, RAM prices and disc space reach the same level of affordability --which shouldn't be long.
The area where 4K needs to be used is in dealing with current films and digital intermediate processes. I watched "Kill Bill" Volume 2 this weekend, and the black and white stuff (along with the Pei Mai scenes) looked video-ish. Electronic noise, shimmering and even some hints of pixellation were popping up on screen. That stuff wouldn't show up so easy if done at 4K.
Posted by John Pytlak (Member # 331) on 04-19-2004, 10:44 PM:
quote:
Electronic noise, shimmering and even some hints of pixellation were popping up on screen. That stuff wouldn't show up so easy if done at 4K.
Pretty much the findings of image scientist Dr. Roger Morton and his team, as reported in several recent SMPTE papers:
Roger Morton on Film Scanning
quote:
Digital film mastering is now typically done at 2K resolution, however the technology for working at 4K resolution is available, and Morton says it is expected to become a more practical and affordable tool in the foreseeable future.
SMPTE Paper May 2003
quote:
An Introduction to Aliasing and Sharpening in Digital Motion Picture Systems
By Roger R. A. Morton, Michelle A. Maurer, and Christopher L. DuMont
Abstract
Many interacting factors affect image quality. This paper discusses the measurement of some crucial factors and reviews their interactions. Specifically, how pixel count can interact with the image-quality factors of limiting resolution, aliasing ratio, and shape of frequency response. A new tool designed for onscreen image-quality measurements for both film and electronic projectors is described, and actual examples of aliasing artifacts are shown in still images from digital motion picture systems. It is then illustrated how aliasing artifacts can arise, although Nyquist sampling requirements are satisfied. Finally, some of the interactions between limiting resolution and frequency response shape for still and moving images are explained.
SMPTE Paper July 2003
quote:
Relationships between Pixel Count, Aliasing, and Limiting Resolution in Digital Motion Picture Systems
By Roger R. A. Morton, Christopher L. DuMont, and Michelle A. Maurer
Abstract
This paper analyzes how pixel count affects one type of aliasing artifact and image rendition near limiting resolution. A previous paper1 showed that aliasing artifacts take many forms. This paper focuses on the aliasing artifact identified in the ISO 12233 standard and identified as Type A aliasing in the prior paper. A relationship termed “Type A aliasing equation” is presented, which predicts aliasing as measured by the ISO 12233 standard. It is then demonstrated that this equation predicts the best unreconstructed aliasing performance for digital motion picture systems and subsystems, thereby defining one characteristic of an ideal system. The predicted result is then compared by the Type A aliasing equation and the measured aliasing performance of 20 different digital motion picture systems and subsystems. It is also shown that the Type A aliasing equation is superior to the classical Nyquist theory as a predictor of aliasing performance of digital motion picture systems. Finally, the equation is used to compute the minimum number of pixels required for a given aliasing level and limiting resolution, and data is presented to determine the pixel count required to render a given limiting resolution.
[ 04-20-2004, 07:43 AM: Message edited by: John Pytlak ]
Posted by Zach McQuaid (Member # 613) on 04-20-2004, 12:20 AM:
huh?
Posted by David M. Dorn (Member # 2298) on 04-20-2004, 12:37 AM:
The use of clustered G%'s is not new. The Berkley/Manneheim/Tennessee list of super computers as of Nov. 2003 says:
Quote:
The third system ever to exceed the 10 TFflop/s mark is Virgina Tech's X Cluster Institute measured at 10.28 TFlop/s. This cluster is built with the Apple G5 as building blocks. It uses a Mellanox network based on the new Infinband technology as interconnect.
The VA computer is using 1,100 G5's each with 4gb memory and a 160gb harddrive. The two computers ranked faster each cost $250-400 million. The VA comuter cost $5 million!
The real hold up for a 4K standard is data storage. The Dalsa camera can originate full frame 35mm 4K images in RGB with 14bit depth....but the data rate is about 1 terabyte for 15 minutes of filming.
Posted by Daryl C. W. O'Shea (Member # 1303) on 04-20-2004, 12:44 AM:
Heck, if I can manage over a terabyte in my desktop computer, the industry can afford 10 TB a feature.
Posted by Leo Enticknap (Member # 534) on 04-20-2004, 01:55 AM:
quote:
Many restoration specialists do this sort of work manually, often frame by frame.
Because in many scenarios, automating the process of examining and correcting defects using software produces flawed results; and I certainly don't know of any computer-based system which can carry out an automated technical comparison of multiple source elements. If you're starting with a camera negative in relatively good condition (as with the You Only Live Twice example), then technical selection is a no-brainer. And if there is only one form of consistent damage (e.g. dye fading), then the restoration process can be automated efficiently with very good results. But what happens when you're faced with 20 source elements of the same material, which may be a mix of intermediates, release prints, separate titles and opticals and all sorts? And furthermore, what if the element you select for duping has a combination of many different kinds of damage in various sections and to differing levels of severity? In that scenario there's no substitute for an archivist making decisions using his or her eyes. As Martin Koerber pointed out in a 'restoration of' extra on the Metropolis DVD, high-resolution digital image manipulation has a huge amount to offer in terms of implementing restoration decisions, but in most cases, only when it's under the control of an experienced operator. Just scanning the film and telling the computer to get on with it will often yield inferior results.
quote:
Mr. Lowry may be the first to do digital restoration while the images are still in a digital format — a bunch of 0's and 1's inside a computer — before they're transferred to video.
Whether he was the first or not, it is increasingly becoming a routine tool, even for smaller archives. For example, one of the regional archives in Britain has recently bought the Diamant software, which runs on a normal high-end PC. They send films to a facilities house for scanning. The data comes back on some sort of tape streamer and is then worked on in 10-minute segments before being outputted back to another tape and sent back to the lab for output on the required format. This can be anything from 35mm to VHS. Both the original, unrestored 'as scanned' data tapes and the restored data are archived as master elements, along with the film originals.
So it seems to me that the only significantly new thing Mr. Lowry is doing is to scan the 35mm originals at far higher resolution than he'll ever need for any sort of retail video format, unless we get to the stage of having 20ft screens in our living rooms. Personally I'd say he was throwing money down the pan, because the cost per foot of scanning this stuff is likely to drop massively in the next few years. Surely it makes far more sense to preserve the originals in the ideal atmospheric conditions (where they would survive without any significant further decomposition for centuries), and scan them to the required and most cost-effective resolution as and when you need them in digital form.
Posted by Mike Olpin (Member # 1115) on 04-20-2004, 03:57 AM:
Hey, at least he got this right!
quote:
Impressive as HDTV looks, 35-millimeter film has far more color and detail.
You can watch a really awsome video about Virginia Techs G5 supercomputer/cluster here .
Posted by Mattias Ohlson (Member # 1662) on 04-20-2004, 07:38 AM:
I have read about scanning the full negative at 8kx6k to avoid artifacts and then processing that to store at 4k.
Dvds should improve if you scan the orignal camera negative at 4k compared with scanning a nth generation copy at SDTV resolution. That is not difficult to understand.
A new dvd format at 2k should look pretty good from a 4k master like that one talked about in the article.
Posted by John Pytlak (Member # 331) on 04-20-2004, 10:54 AM:
Generally, higher quality origination (e.g., 65mm) or higher resolution scanning (4K vs. 2K) will result in higher image quality, even if the final display system is lower in resolution.
As good as Super-16 origination is for both SD and HD television, 35mm is usually discernably better. And 65mm even better.
Posted by Mattias Ohlson (Member # 1662) on 04-20-2004, 04:59 PM:
It was Imagica that was mentioned in this thread. They are upgrading to a 10k ccd sensor. That way you can scan at twice the archice/master resolution to improve quality.
Posted by Leo Enticknap (Member # 534) on 04-21-2004, 01:27 AM:
quote: John Pytlak
Generally, higher quality origination (e.g., 65mm) or higher resolution scanning (4K vs. 2K) will result in higher image quality, even if the final display system is lower in resolution.
No arguments with that, but I can't see how it justifies a business case for archival scanning at 4k for the purpose of DVD mastering here and now.
Firstly 9.8mbit MPEG-2 is currently the highest definition that any domestic video playback system will support. This is likely to remain the case for at least the next few years, simply because the market saturation of this format is already so high (after all VHS is still in widespread use, almost two decades after it was launched). There is a lot of hardware and software out there, and its owners are not going to buy new systems which are incompatible with it in a hurry.
Secondly, decades of research by scientists and archvists has established that film - even nitrate - will remain stable, literally for centuries, if it is stored in optimum environmental conditions.
Thirdly, the cost of scanning film is coming down and the resolution of the output is going up. It will be a lot cheaper to scan a reel at 4k this time next year than it is now.
And finally, there are big question marks over the suitability for long-term storage of just about any offline digital data carrier you care to mention. When Mr. Lowry says that his digitised movies 'will last the next two or three generations of technology', does he realise just how short a time frame that represents? Try reading a floppy disc or playing a DAT audiotape written a decade ago and it's likely to be corrupted beyond redemption.
Assuming that the film elements have been properly archived, why sink all this money into speculative scanning now, the benefits of which probably cannot be seen on any existing domestic format? You will almost certainly produce a cheaper and better quality result waiting until you actually need the data in order to produce it.
Posted by Mattias Ohlson (Member # 1662) on 04-21-2004, 02:10 AM:
I believe consumers deserves as good a dvd quality as can be produced so I welcome this. Also 20Mbit Mpeg-4 kind hd-dvd at 1920x1080 in the segmented frame format are not far away. That justifies the scanning even more not to talk about digital cinema at 4k with bitrates perhaps well over 100Mbit.
You can always buy towmorrow but that is a lame argument.
Posted by Patrick de Groot (Member # 858) on 04-21-2004, 08:57 AM:
quote: Bobby Henderson
The area where 4K needs to be used is in dealing with current films and digital intermediate processes. I watched "Kill Bill" Volume 2 this weekend, and the black and white stuff (along with the Pei Mai scenes) looked video-ish. Electronic noise, shimmering and even some hints of pixellation were popping up on screen. That stuff wouldn't show up so easy if done at 4K.
quote:
from other thread:
“Casino was the only film I’ve shot in Super 35 format,” he says, “and I was terribly disappointed with the release print quality. It was devastating to me. Quentin had a similar experience with Reservoir Dogs. Films stocks and lenses have improved, but there is still a chasm (separating the Super 35 and anamorphic formats).”
Why does 2K scanning of Kill Bill 2 makes sense if they are worried about image quality? They must surely have the budget to use 4K scanning if the digital intermediate process is used because it is after all a blockbuster movie...
Btw was the digital intermediate process also used on Kill Bill 1?
Posted by John Pytlak (Member # 331) on 04-21-2004, 10:28 AM:
Repeating the link to Kodak interview with cinematographer Robert Richardson on "Kill Bill" production:
http://www.kodak.com/country/US/en/motion/students/filmtech/killBill.shtml
quote:
The problem was that Tarantino had written specific "zoom" lens shots into the script. Richardson felt that the anamorphic zoom lenses available from Panavision didn't provide a fast enough T-stop to properly execute those shots. That drove the decision to use a digital intermediate process. That approach enabled Richardson to shoot with spherical lenses in Super 35 format and "squeeze" the images into a 2.4:1 aspect ratio during digital mastering. Basically, the edited cut of the negative is scanned and converted to digital files. The images are timed for shot-to-shot and scene-to-scene continuity in a digital suite, and then recorded out to an intermediate film used as a master for release printing. It's a familiar tactic for Richardson, who has shot many commercials that have been converted to digital format and manipulated in telecine suites. However, this is his first experience with a "film-out" from a digital intermediate.
He suggested shooting Kill Bill in three-perforation 35 mm format. That trimmed film and lab costs by 25 percent. Richardson explains that essentially covered costs for the digital intermediate process. There is no downside, since the digital files can be recorded out to a four-perf 35 mm color intermediate film with no loss of picture quality.
Front-end processing and release printing were slated for Technicolor Labs, in Los Angeles. Tarantino agreed with Richardson's suggestion about creating and timing a digital file at Technique, the lab's subsidiary in the Los Angeles area.
Posted by Jeffry L. Johnson (Member # 453) on 04-21-2004, 12:47 PM:
Prior to the New York Times article, Apple covered Mr. Lowry in one of its email newsletters.
John D. Lowry quote:
By Joe Cellini
If you’re a real fan, you already own the recently-released “Adventures of Indiana Jones” (”Raiders of the Lost Ark,” “The Temple of Doom,” “The Last Crusade”) DVD set, and you like what you see: ridiculously clear images, bonus documentary, nice extras.
Chances are you’d be equally impressed by what you don’t see, conspicuous flaws removed from the films by master digital restorer Lowry Digital Images. “One of the biggest problems we had with ‘Raiders,’” says CEO John Lowry, “was a blue line cutting across the actors’ faces right down the center of the first part of the movie, about 35,000 frames of scratches.”
And if you’re loading the movie now to look for traces of the scratch, don’t bother. Lowry’s digital botox, powerful proprietary software running on a formidable stack of Power Macs, removed all traces of the unsightly line that distracted so persistently from the more compelling vertical furrows on the brow of Harrison Ford.
Jim Ward of LucasFilm, who selected Lowry Digital Images, Inc., to restore the films, says his company first considered doing the clean up themselves. “We could have done a lot of the dirt removal, but we knew it wouldn’t be as good. We owe it to filmmakers like Steven Spielberg and George Lucas to make sure that these films are pristine.”
First Mover
By all accounts, few people do pristine as well as Lowry, who holds original patents on noise reduction in film. “The first clean up of this kind that was ever done was in 1971 when I processed the Apollo 16 and Apollo 17 pictures from the moon,” he says. “So I’ve been at this a little while.” Recently he was awarded the James A. Leitch Gold Medal from the Society of Motion Picture and Television Engineers for his career contributions to digital technology and motion imaging.
Lowry set up Lowry Digital Images in 1988, intending to develop digital tools for sale or license, but first taking time to have his crew clean 20 movies to learn precisely where the problems were, so he could build correct solutions. “Now that we’ve done 80 movies, we’re still discovering new things, but far fewer,” he says.
His timing and tactics proved as good as his technique. Explosive growth in digital distribution channels — DVDs, cable, HD TV, digital projection theaters, even the Internet — has every major studio looking at digital restoration, and looking at Lowry.
Film To Go
It’s easy to believe that movies are forever. But a different message plays across the banks of busy monitors at Lowry’s offices in a converted aerospace plant in Burbank, CA: it says, without significant intervention, the only inevitable thing about film is that it decays over time.
You may have already concluded from watching movies through what looked like tape come unstuck from dirty floors — the swimming motes, the glowing hairs — that film is a fragile medium. You may not know that films before 1950 were shot on nitrate-based cellulose film, an unstable medium prone to disintegration, even conflagration. Fewer than half of those films exist today.
Dirty Movies
But when I suggest to Lowry that fans of recent films like the “Indiana Jones” movies might be surprised that the works of high-profile filmmakers with high-tech production cred would be in anything but pristine condition, he sits me down to explain the facts of film.
“In ‘Indiana Jones,’ we removed a piece of dirt from only every frame or two,” he says. “For two versions each of three movies, that’s half a million pieces of dirt. But for ‘Roman Holiday’ (1953), we removed hundreds of pieces of dirt from each of the 170,000 frames. In a typical old movie, you’re talking millions of pieces of dirt.”
Degrees of Degradation
As we walk the lab, Lowry outlines the three-part problem faced by film restorers, making frequent visual references to convenient examples from failing classics. “The first category is wear and tear. A popular movie tends to be more damaged simply because it’s been printed so often. Real bombs that never got printed have no broken sprocket holes, not a lot of dirt, no scratches. Even the most careful physical handling still adds dirt, scratches, blemishes and even tears. Film is a delicate physical medium.”
Lowry’s second category is age. “Just plain time results in fading and flicker. A sky might pulse from yellow to blue in a beat. On films from the 30s and 40s, you get this vinegar syndrome, and the film turns to dust.”
Third and most lethal, Lowry explains, is multiple generations of optical copies. “They’re the worst degradation that can happen to film because fine detail disappears with each optical duplicate made. We can fix dirt, scratches, tears, and flicker, but if the information is gone, it’s very hard to put back.”
Pressing his case, Lowry brings the evidence. Literal wear and tear in a frame from “North by Northwest” shows Cary Grant in a cornfield, torn in half at the waist. Aggregations of grain and dirt occlude third-generation dupes of “Roman Holiday” and “Sunset Boulevard,” the best extant versions of either film, because originals were lost.
The Digital Fix
Then from deep in the third circle of celluloid hell, Lowry shows the way out. Summoning “after” images on a screen to compare alongside the “befores,” he pulls up a perfectly intact Cary Grant. Also Audrey Hepburn in a stunningly photographed black-and-white Rome, and Gloria Swanson as Norma Desmond, both looking ready again for their close ups.
It’s clear that Lowry’s software has repaired the frames, less clear how it found enough visual information to do so. “The motion picture camera presents lots of information to the film frame, to the emulsion,” he says. “The emulsion has a grain structure that allows you to collect certain information, but prevents you from collecting other information. We’re in the business of extracting that information. By borrowing and combining information from a number of frames, you can do some interesting things.”
Suspecting that he lost me at emulsion, Lowry tries again. “It’s like cleaning your glasses so you can see more. We want to remove the veil so that you can see everything that’s in the original.”
Automatic Software
Precisely how Lowry cleans the glasses is the stuff of proprietary algorithms. But the process is fairly transparent. First Lowry’s project managers scan film or tape into high-resolution digital format, note standard problems and tweak the parameters in the proprietary software (for example, to stabilize jitter and weave). Then they process the digital frames on Lowry’s battery of Macs.
Once the file has been through the process, the imagery is eye-checked frame-by-frame, and then checked with clients. “We go through it to make sure the algorithms did their job properly,” he says, “but the algorithms now are very sophisticated and so there are very few do-overs.”
Digital Ins and Outs
Impressively, Lowry’s software works with images in any size and format, including NTSC, PAL, HD, 2K, 4K. And once Lowry creates a pristine digital master, his clients can output to any of those formats to cover everything from a new film to digital projection to DVD.
Besides cleaning better than other technologies can manage, Lowry’s software enhances his clients’ translations across digital formats by avoiding unnecessary early compression. “Because we do our processing at 32-bit floating point, we expand everything we get. We avoid compression in production stages because it usually does terrible things to pictures,” he says.
The eventual compression required to get a movie onto a DVD can leave artifacts if there is a lot of noise in the images, but Lowry’s robust processing yields even compressed images that are artifact free.
Does Compute
Lowry believes that personal computers, even more than the generous studio budgets for DVD production and distribution, have made digital restoration viable.
“To do the work we are doing today 10 years ago, would have cost tens of millions of dollars for a computer facility,” says Lowry. “The economics of it, without question, would have been impossible. But along came the Power Mac G4 and the Power Mac G5, and suddenly we had an answer.”
Jim Ward makes the point that Lowry’s economic model, as much as his algorithms, brought him to Lowry. “His software does 95% of the work,” says Ward. “There’s a huge economic difference between letting 100 Power Mac G5s render all night long versus having 8 or 10 guys sitting at workstations for a much longer period of time eyeballing pieces of dirt.”
Powering Through
Lowry plans to acquire another 100 Power Mac G5s, and he’s clear about why. “Obviously speed. Both the Power Mac G4 and Power Mac G5 have outstanding 32-bit floating point processors, and we make very good use of them. Another major factor is RAM capacity, important in image processing. The new Power Mac G5s are running with 3.5 gigabytes of RAM in each computer.”
Mac OS X figures in Lowry’s plans for working in higher-res formats. “The studios say they need 4K to get out all the information hiding in their films. So we need access to memory, because these frames are all computed while they’re in RAM. Mac OS X gives us the ability to use virtual memory if necessary to swap around.”
Too Clean?
Lowry’s process, while undeniably effective, is also somewhat controversial. Some industry critics have claimed that digital processing cleans a film too well, stripping it of distinctive grain.
“Generally speaking in the industry, grain removal equals detail removal,” says Lowry. “But we’ve been able to reduce the granularity while increasing the detail. I don’t know that anybody else is doing that in an effective way. As a matter of fact, most of the grain we reduce is grain that crept into the film through many generations of optical duplications. The philosophy of my critics, that every grain in a film is sacred, and to leave grain untouched, is a bit ridiculous in my mind.”
Jim Ward, the happy customer, concurs. “The issue of how much grain to keep in the trilogy was Steven Spielberg’s call. It was really about exact balance, which John has the ability to modulate. Steven wanted a little grain — that’s what he likes and prefers — and you see that in the end product, but with beautiful clarity and color saturation.”
Back to the Future
Industry interests are broadening Lowry Digital beyond standard restoration, including amazing experiments like taking smaller formats to IMAX and re-mastering any format for digital projection. “We’re finding new things to apply the software to every day,” says Lowry.
But restoring films remains Lowry’s main business. He’s working again with Jim Ward and Warner Bros., just finishing a restoration of George Lucas’ first film, “THX 1138.” “After we saw what he could do with ‘Indiana Jones,’” says Ward, “we definitely wanted back in the queue.”
Asked what would make an ideal next project, the man who for more than 30 years has pulled the long levers behind a closed industry curtain, makes just the right call. “‘The Wizard of Oz,’ from separations.”
John D. Lowry
1. Restoration Software
2. Removing the Veil
Great Saves
“The Ghoul” (1933)
Worst-case scenario restoration from jittery, flickery, dirty print. Looks like HD on DVD.
“Sunset Boulevard” (1950)
Now famously ready again for close up with DeMille after Lowry’s reclamation from a third-generation dupe neg. Astonishing details rescued from grain and dirt (300-400 pieces per frame).
“The Importance of Being Ernest” (1952)
Red flicker, processing streaks in dark areas, overall flarey. Flare reduced, flicker gone.
“Roman Holiday” (1953)
Lowry’s most challenging restoration — grain-greyed whites, dirt, flicker, jitter and weave, tears, scratches. Now on DVD as it looked 50 years ago, with restored screenwriter credit for blacklisted Dalton Trumbo.
“North by Northwest” (1959)
Lowry’s first restoration, this very dirty job proved the concept of his approach. All dirt removed, “brutal” tear resolved, visible effects wires edited out.
“Dr. Zhivago” (1965)
Long movie, lots of dirt and flicker, none visible on DVD. Image-by-image, a Lowry favorite.
“Once Upon a Time in the West” (1968)
Technoscope, a half-height film format, yielded wide aspect ratio but also twice the grain and jitter. Cleaned, stabilized images produced a stunning DVD.
Show Stoppers (Know Your Notorious Film Flaws)
Dirt
Pervasive, miscellaneous grit that grows through multiple printing of popular films.
Flicker
A symptom of film aging that manifests itself as color breathing, as when the picture flickers from blue to slightly yellow and back.
Grain
The particulate texture responsible for film’s brilliant images can coarsen over generations of prints and actually hide information.
Jitter and Weave
Sprocket holes in film move the film through the camera gate introduce jitter and weave, which becomes more pronounced with successive printings.
Posted by Bobby Henderson (Member # 840) on 04-21-2004, 05:54 PM:
quote: Patrick de Groot
Why does 2K scanning of Kill Bill 2 makes sense if they are worried about image quality? They must surely have the budget to use 4K scanning if the digital intermediate process is used because it is after all a blockbuster movie...
It is the good old "it's good enough" practice, backed up with extending profit margins. Why spend more money doing a 4K-based digital intermediate when 2K is "good enough?"
The computing power and data storage capability is certainly there to handle major productions in 4K. The studios just don't feel like spending the money.
It's the same problem as 65mm photography being very feasible for major productions, but no one shoots a major feature in that since 35mm 'scope or super 35 is "good enough."
quote: Patrick de Groot
Btw was the digital intermediate process also used on Kill Bill 1?
Yes. I don't remember who handled the digital intermediate stuff, but there is a listing in the credits for it.
Posted by John Pytlak (Member # 331) on 04-21-2004, 06:15 PM:
From my posting above:
quote: John Pytlak
Tarantino agreed with Richardson's suggestion about creating and timing a digital file at Technique, the lab's subsidiary in the Los Angeles area.
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 01:34 AM:
That piece about Lowry on Apple's website is some of the most biased BS I've ever come across. The message seems to be 'just buy a Mac and you too can do this.' Furthermore, he appears to know worryingly little about the physical properties of film for someone who is routinely let loose on the o-negs of big features.
quote: Apple website article posted by Jeffry L. Johnson
You may not know that films before 1950 were shot on nitrate-based cellulose film, an unstable medium prone to disintegration, even conflagration. Fewer than half of those films exist today.
...and Apple doesn't want you to know that in appropriate storage conditions, your movie will last a lot longer on nitrate than it will as digital data stored on magnetic media.
quote:
“Just plain time results in fading and flicker. A sky might pulse from yellow to blue in a beat."
While it is certainly true that film scanners and modern flying spot telecines can run the film at any speed you like and then do the sums to make the output scanning rate compatible with either the relevant TV system or a DLP projector without any visible flicker, this is not the result of physical changes to the original over time. If Mr. Lowry's digitised film is outputted to 35mm and then shown at 16fps with a two-blade shutter, it will still flicker.
quote:
On films from the 30s and 40s, you get this vinegar syndrome, and the film turns to dust.”
Odd, given that the film base which is most susceptible to VS (i.e. cellulose triacetate) was not sold on any significant scale until 1948. And although I have come across many warped and buckled acetate elements, I have never, ever seen one which has 'turned to dust'.
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 03:20 AM:
If you read past the pro Mac stuff Is not the point that old movies with copies in a bad shape can be freshened up and put back on a modern film for storage.
Is that a bad thing, saving good old movies? I understand that the copies are not total crap but is a good practice to care for your house.
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 04:34 AM:
It's a very good thing - the issue is how you do the saving.
A basic guide to archival practice can be found here (the International Federation of Film Archives publish a similar manual, available from here, though it's not free to download). Without writing another 150-page manual myself, the generally accepted principle is that you preserve the original element, or the nearest generation you have to it, for as long as possible and in the exact same state as when the archive acquired it.
The rationale for this is that as new copying techniques become available (be they analogue or digital), you will always get the highest quality dupe from the original. This wasn't always the received wisdom. Until the 1980s many archives had a policy of destroying nitrate immediately after duping to acetate, for example. But then the full extent of the VS and dye fading problems became known, and printing technology improved. If you look at prints made from new 35mm preservation masters nowadays compared with those dating from the 1960s or '70s, the difference can be astonishing.
Furthermore, research by the Image Permanence Institute and others (e.g. the team at Manchester Metropolitan University in the UK, lead by Dr. Michelle Edge) revealed that film base is actually very stable when stored in cool and dry conditions. The mantra of 'nitrate won't wait' does not hold true as much as was once believed. Actually it will: accelerated aging tests suggest that when stored at a little above freezing point and at very low humidity, a nitrate element placed into storage whilst still in the first stage of decomposition will be copyable for around 500 years. Furthermore in these conditions, the decomposition process is so slow that you'll have ample warning before needing to dupe it for the last time.
By contrast, the approach Mr. Lowry seems to be taking is to scan the original film element at 4k, carry out extensive digital restoration and then treat the resulting data stream as the new master element for that material. Any properly trained archivist would reject that approach, and for two sets of reasons.
Firstly, there is no digital data carrier known to man which has the proven storage longevity of film. This divides into two sub-issues: the physical integrity of the carrier itself and the existence of the hardware and software needed to recover the data from it in the long term. Experience has shown that you're doing well to achieve a lifespan of 10 years on either score. So, if you're going to commit to purely digital preservation, you will need to undertake a process of what archivists term 'continuous format migration'. This is exactly what it sounds - copying the data onto format B as format A approaches obsolescence. Not only does this introduce data integrity issues, but it's also hellish expensive. It therefore costs a typical archive far more to preserve videotape than film, for example. If the material was originated on video or as digital data in the first place, you have no choice but to preserve that way. But if it was originated on film, migrating to digital for preservation is simply crazy - it makes no sense at all because it is a lot more expensive and you are not even preserving the original element.
Secondly, Mr. Lowry is not even proposing to archive a 'straight' or unaltered digital representation of the film original. He is going to do all sorts of cleaning up and restoration work first. For presentation to an audience, such restoration is perfectly justified. Most archivists would argue that anything goes if, on a balance of probabilities, it is showing a representation of the material which looks and sounds closer to the original in its (subjectively) undamaged state than any extant film element does. But the process of restoration is introducing changes, made by a human being (or a computer programmed by a human being), which alters the nature of the original. Keep the original, and when newer or more appropriate restoration techniques become available, they can be used. But if you've only kept a 'restored' version, you've burnt your boats.
Scanning films, cleaning them up and releasing them on DVD or whatever is fine. I have no problem with this. But no sane archivist would regard the scanned datastream as a long-term preservation master, as Mr. Lowry appears to.
One very positive aspect to digital restoration, though, is that it almost eliminates the need to carry out irreversible treatments on original film elements (e.g. suspect cleaning chemicals, or vacuum-type rehydration processes). This image enhancement can be done in the digital domain, thereby leaving the film untouched.
Posted by William Hooper (Member # 71) on 04-22-2004, 07:09 AM:
quote:
Too Clean?
Lowry’s process, while undeniably effective, is also somewhat controversial. Some industry critics have claimed that digital processing cleans a film too well, stripping it of distinctive grain.
“Generally speaking in the industry, grain removal equals detail removal,” says Lowry. “But we’ve been able to reduce the granularity while increasing the detail. I don’t know that anybody else is doing that in an effective way. As a matter of fact, most of the grain we reduce is grain that crept into the film through many generations of optical duplications. The philosophy of my critics, that every grain in a film is sacred, and to leave grain untouched, is a bit ridiculous in my mind.”
quote:
“Sunset Boulevard” (1950)
Now famously ready again for close up with DeMille after Lowry’s reclamation from a third-generation dupe neg. Astonishing details rescued from grain and dirt (300-400 pieces per frame).
What an egomaniacal, un-self-assessing statement re: 'my critics'.
I really don't like the Sunset Boulevard DVD. It's been scrubbed flat, depleted of depth, & looks like 1970s b&w video. The picture is not very photographic at all. The processes he's using only have grain removal & detail as an effect.
He'd probably love to get hold of a bunch of old masters & clean up that "atmospheric perspective" that they so misguidedly slaved over. It would yield a much crisper, detailed, flat, comic book-like image to his liking.
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 07:09 AM:
Leo
The point I see with a good quality 4k digital master is that it is accessible. This digital master has a high quality. The film can be archived and everyone is happy.
Someone else can scan the movies without doing to much cleaning up and then print to film for storage.
The concept is not owned by one man.
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 07:17 AM:
quote: Mattias Ohlson
Someone else can scan the movies without doing to much cleaning up and then print to film for storage.
What is wrong with creating the preservation master photographically by continuous contact printing (assuming that the condition of the original doesn't rule that out)? It's much cheaper than scanning it, and because the emulsions of the original and duplicate are placed in physical contact with each other, this method should theoretically produce the highest definition duplicate possible.
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 08:38 AM:
Nothing is wrong with that but that does not give you a source to produce
1 digital cinema file
2 hdtv
3 dvds and hd-dvd
I am not debating film for archive.
1 scan the OCN once and then save the negative
2 use the 4k DI as a base for all digital formats
3 print DI to a high quality duplication negative at any time.
Posted by Mark J. Marshall (Member # 1409) on 04-22-2004, 10:11 AM:
How would this process work on a negative that's considered "unprintable?" If the negative is not in good enough shape to run through a normal printer, it may be in good enough shape to run through this machine at 1 frame every four seconds. Or would the registration be so bad that it wouldn't be fixable?
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 10:29 AM:
Exactly, hence the rider 'assuming the condition of the original doesn't rule that out'. If it does, then there are a range of options, both analogue and digital. What is the best bet really depends on what sort of damage you're dealing with and how severe it is. These are described in a lot of detail on pp. 41-58 of the NFPF preservation guide (see link in my last but one post): in some cases, creating a new master digitally can be justified.
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 10:55 AM:
Leo
You have a point but I guess a negative could have some kind of problem that is easier to work with when the negative has been converted to digital.
I am more thinking along the line of having the 4k digital master and then let the film orignal rest for a long time in the archive. This for newer films.
Posted by John Pytlak (Member # 331) on 04-22-2004, 11:22 AM:
At this week's NAB Convention in Las Vegas, BHP introduced a new wet gate printer head specifically designed to print shrunken originals:
http://www.rtico.com/bhp/shrink.html
quote:
Special " -2% " Wet Printing Head for
Copying Shrunken Archival Films
Chicago, IL - Increasingly, the world at large has become more concerned about the potential loss of some of the classic movies due to lack of proper storage and maintenance in past decades. With the growing global market for filmed entertainment brought on by a wider range of distribution systems, new and old products have been and will continue to be very much in demand to fill the pipeline.
One of the problems inherent in old movies is the fact that motion picture film will shrink over time. Therefore making prints from aged and shrunken negatives is a serious problem when raw print stock and shrunken sprocket holes do not line up right.
Recently BHP Inc., the leading maker of professional motion picture film printing systems, has undertaken the development of a 35mm wet printing head for restoration applications. This head is designed for use on films that have shrunk as much as 2%. With this head, the difference between the original film’s sprocket hole spacing and that of the raw film stock is cancelled out.
For details contact BHP headquarters.
http://www.rtico.com/bhp/
quote:
Motion Picture Printers
For years, BHP Inc. has been and continues to be the world's leading manufacturer
of Motion Picture Film printing equipment. Our products are used in film
laboratories the world over. Call us for information and a quotation.
BHP INC.
4700 Chase
Lincolnwood, Illinois, USA
Phone: 647-677-3000 Fax: 847-677-1311
E-Mail: sales@bhpinc.com
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 12:59 PM:
Mr Pytlak
Is this shrinkage of old films a problem when scanning as well as it "was" with printing before this announcement?
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 01:07 PM:
That looks like a contiuous contact head as well (presumably the space above the sprocket drum is where the user adds a light source) - if it is then I'd be interested to know how it's able to keep the source and destination stock in even, flat contact while in continuous motion, without frame line slippage. After all, if a source element was originally 1,000 feet and has shrunk 2%, then that makes a 20 foot difference in length between the source and destination stock. To achieve a good copy by continuous contact printing seems to me to be defying the laws of physics.
quote: Mattias Ohlson
You have a point but I guess a negative could have some kind of problem that is easier to work with when the negative has been converted to digital.
Agreed, there are some scenarios where the damage is so severe that digital restoration is either the best or the only way to create a viable preservation master. I'm simply suggesting that this should be the last resort, not the first. All of which is a totally different purpose from creating viewing copies, for which a high resolution scan offers all the benefits you outline.
Posted by Bobby Henderson (Member # 840) on 04-22-2004, 01:11 PM:
quote: Leo Enticknap
What is wrong with creating the preservation master photographically by continuous contact printing?
quote:
Nothing is wrong with that but that does not give you a source to produce
1 digital cinema file, 2 hdtv, 3 dvds and hd-dvd
I don't agree with this for a number of reasons.
First, others have mentioned how archival safety film (when properly stored) will last longer than any digital file storage media. That is a fact. The digital fans can bring up the color shifting and fading arguments against film all they want. At least the film still works. If you have something go wrong with your digital media, the whole damned file is toast!
Second, it really makes little sense to archive a film digitally when the analog original negative and archival safety interpositives and internegatives will do a better job, and do so more cheaply.
Third, it only makes sense to digitize a film if the resulting file will immediately be used for some kind of video presentation medium, such as a DVD release.
It is useless to digitize a movie in 2K or 4K format only to put the file away in storage. Why? You're not going to use that data for whatever video format is popular 20 or 30 years from now. You're going to take those analog film sources and digitize them using the scanning and video mastering equipment that is state-of-the-art in that age. You'll get much better results.
To use an analogy, just look video mastering 10 years ago. The first THX certified Laserdisc of "The Abyss" was given raves for its video quality. But 10 years later, the same techniques would get a grade of "F" for DVD release. The standards are so much higher now, only after a decade.
Those 4K scans Lowry is doing might be nice for current day film digital intermediate work. But they will be largely useless only a decade or two from now.
Posted by John Pytlak (Member # 331) on 04-22-2004, 03:03 PM:
quote: Mattias Ohlson
Is this shrinkage of old films a problem when scanning as well as it "was" with printing before this announcement?
There are pin-registered scanners that position the film with a full-fitting registration pin, much like an optical printer.
Even continuous scan telecines and datacines can be adjusted to properly handle and position shunken film.
In a continous contact printer, the original and raw stock are wrapped around a large sprocket, so the sprocket pitch and diameter need to be optimized for the pitch differential between the two films to avoid any slippage.
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 03:30 PM:
Sorry, John, I was being a tad dense! Of course, the contact is in an arc, not flat, thus enabling a slight differential between the source and destination stock to offset the shrinkage. That's really clever, and the gain in definition by not needing to resort to optical printing must be very significant.
Posted by Mattias Ohlson (Member # 1662) on 04-22-2004, 03:58 PM:
Bobby H.
I am not saying a 4k will last forever but scanning a 35mm frame at higher and higher resolutions give diminishing returns. You always make the argument how cheap the studios are so now I make use of that point.
Would you say that scanning a 35mm frame at 8kx6k captures 99 percent of what is there? Downsizing this to 4k deals with sampling problems and gives a good working material. I would guess without knowing that this might be good enough for many many years. Later you can improve the release by better compression and higher bit rates if we talk digital. If we talk about creating new film copies we might talk about improved laser recorders used directly instead of going through a duplication negative to reach the release print.
Posted by Bobby Henderson (Member # 840) on 04-22-2004, 04:26 PM:
quote: Mattias Ohlson
I am not saying a 4k will last forever but scanning a 35mm frame at higher and higher resolutions give diminishing returns.
Yes, at some upper limit you will run out of detail in a film frame to scan. I don't think 4K truly reaches that limit. It will be some time before a film scanner can truly capture all the detail, color depth and contrast of a 35mm film frame.
That brings up the issue of limitations of current state-of-the-art film scanners. They don't represent any final state of film scanner development. Improvements will continue to happen in photomultiplier tube, CCD and CMOS technology for film scanners. That will continue to extend capabilities in resolution, contrast and color depth.
Give it a decade and the scanners Lowry is using will seem primitive by comparison. High end scanners 10 or 20 years from now will have much better image capture quality than systems currently in use. They'll also be faster, more economical and more adept at addressing presentation formats in use at that time.
quote: Mattias Ohlson
Later you can improve the release by better compression and higher bit rates if we talk digital.
That's only if you start with a superior, analog film source. It is totally useless to try to interpolate and blow up a low rez videotape production to a higher resolution format. This is exactly why studios need to keep shooting major releases on film, not video. Improved scanning technology is also why film studios really need to consider shooting major features on 65mm instead of 35mm. Future-proof the investment in that film project.
[ 04-22-2004, 08:58 PM: Message edited by: Bobby Henderson ]
Posted by Leo Enticknap (Member # 534) on 04-22-2004, 05:12 PM:
quote: Mattias Ohlson
If we talk about creating new film copies we might talk about improved laser recorders used directly instead of going through a duplication negative to reach the release print.
Agreed. I presume that the current practice of outputting an interneg from the final digital version and thereafter striking release prints using a conventional high-speed continuous contact printer is for economic reasons. But if a viable system could be devised whereby you could 'burn' all the 35mm release prints directly from a digital intermediate, it could potentially take a lot of the consistency issues out of big print run releases which currently exist (unless more of these are caused in developing than printing).
Posted by Tao Yue (Member # 836) on 04-22-2004, 10:06 PM:
In the wee hours of this morning Leo Enticknap wrote:
quote: Leo Enticknap
... without any visible flicker, this is not the result of physical changes to the original over time. If Mr. Lowry's digitised film is outputted to 35mm and then shown at 16fps with a two-blade shutter, it will still flicker.
in response to the snippet in the article about:
quote: Leo Enticknap was the last to post
Just plain time results in fading and flicker. A sky might pulse from yellow to blue in a beat.
While Mr. Lowry did use the word flicker, he also referred to abrupt color shifts, in fact he even referred to fading. The phenomenon being corrected is variable fading over frames, caused by improper storage. It has nothing to do with frame rates or two-bladed shutters and indeed comes from physical degradation of the film.
To use an extreme example, if you stored film vertically in a box that kept the top half in bright light and the bottom half in total darkness, for example, then after a while half of the film will have faded and the other half won't. When projecting or telecining, every half revolution will show an abrupt shift in contrast and brightness. If this happens fast enough (like near the center of the reel), it looks like flicker.
Clearly the writer is not very technically-inclined, so it makes sense for Mr. Lowry to refer to a visual rather than technical description of the problem. Think of it as correcting flash frames introduced during printing, except that it happens more frequently than once per scene and involves a less predictable shift in contrast and brightness.
Posted by John Pytlak (Member # 331) on 04-22-2004, 10:18 PM:
That type of "convolutional repeat" fading is usually seen when a roll of film is stored for a prolonged time near a heat source, such that one side of the roll is constantly warmer than the other.
SMPTE Recommended Practice RP131 specifies storage conditions for long term storage of motion picture film.
Posted by Mattias Ohlson (Member # 1662) on 04-23-2004, 03:00 AM:
Bobby H.
I am always assuming film origination here. For some lower budget "movies" some form of refined video production might be used but that is not part of this discussion.
With your reasoning you can never do anything because of the capabilities available in some fuzzy future. I do not get this part of your reasoning.
Yes, I am sure scanners will improve especially in sensitivity. We might decide that there are other parameters besides "resolution".
Posted by John Pytlak (Member # 331) on 04-23-2004, 05:36 AM:
quote: Mattias Ohlson
We might decide that there are other parameters besides "resolution".
Of that there is no doubt, especially when it comes to sensors used in digital cameras.
Posted by Leo Enticknap (Member # 534) on 04-23-2004, 12:29 PM:
quote:
In the wee hours of this morning Leo Enticknap wrote:
Uhh? The post you quote was at 2312 my time (according to Daryl's nifty local time feature, which squares with my memory), which should mean early evening in MA if I've got my sums right.
Thanks for pointing out his definition of 'flicker', though - I didn't see that he was talking about dye fading caused by inconsistent heat or light exposure.
Posted by Bobby Henderson (Member # 840) on 04-23-2004, 08:57 PM:
quote: Mattias Ohlson
With your reasoning you can never do anything because of the capabilities available in some fuzzy future. I do not get this part of your reasoning.
I didn't say you can't do anything with 4K. 4K would be a much better alternative to the current 2K de-facto standards for CGI and digital intermediate work on contemporary movies. 4K production certainly would have helped the digital intermediate scenes in "Kill Bill." Those scenes would not have looked so video-ish. I just don't see the point of using 4K as a means of long term archival. That's what can make 4K into a solution in search of a problem. It doesn't beat the original analog film elements and will not last in storage as long as those film elements.
If a movie was shot on film and then revived on some home video format 20 years from now, those original film elements are just going to get re-scanned again at that time using up to date equipment of that age. The post production house isn't going to use a 20 year old digital file. The quality won't be acceptable.
Posted by John Pytlak (Member # 331) on 04-25-2004, 04:46 PM:
Mr. Lowry is right in going for the highest quality scanning and digital restoration, even for release on today's lesser video formats. Of course, the original film elements should continue to be preserved (e.g., specifications in SMPTE RP131), as improved future digital restoration (for film and future video formats) can best be done by rescanning the original film elements.
If a 4K digital restoration is done today, it's likely there will be a "film out" for a limited film release to promote the digital release. If so, you have a new film element representing today's "state of the art" restoration, that can be stored for future use, in ADDITION to the original film elements.
Powered by Infopop Corporation
UBB.classicTM
6.3.1.2