• Blogs
• Forums
• Tradeshows
• Partners
• White Papers
• Events

 

Data Driven, Part II

Jan 1, 2006 12:00 PM, By Michael Goldman

Overcoming Storage and Compatibility Roadblocks

The postproduction industry continues to expand existing digital intermediate pipelines toward larger, seamless data pipelines for image acquisition, viewing, manipulation, transfer, editing, and finishing. In the process, new workflows — and differing opinions about them — are cropping up.

As discussed in last month's issue, some major facilities (Efilm, Technicolor Digital Intermediates, and Laser Pacific) have recently instituted competitive new digital dailies services, advanced technology for viewing images and improving communication between DPs and colorists, high-resolution digital viewing theaters, and other infrastructure improvements. Similar advancements are moving forward inside other major DI facilities across Hollywood. Officials representing such facilities almost universally agree that the ultimate goal for such pipelines is “the metadata concept — passing it along seamlessly from one device to another,” in the words of Al Hart, executive VP of engineering at Modern VideoFilm, Glendale, Calif. But how to best get there from here is a topic of much debate. Less debatable is a simple question of logistics: Many in the industry emphasize that the cost of storage for that data-centric future remains, for now, prohibitive for most industry players.

“Storage is the big challenge for that [tapeless] approach,” says Steven Cohen, VP of feature services for Ascent Media in Santa Monica, Calif. “For the DI side of things only, it works fine — you are only a month or two away from the movie's release. But for dailies, you might be a month or a year away from release, and to keep 20TB online or nearline to move the work along could prove difficult and expensive.”

For most companies, a far more feasible approach is to concentrate on expanding their slate of services to create solutions for every individual step in the pipeline and build an infrastructure so that those steps can be seamlessly connected as obstacles fade away.

DI to Screen

he strategy of taking pipeline solutions step by step has permitted large companies like Ascent to experiment with building all movie deliverables — from early dailies to a digital cinema release and everything in between — out of data made from initial scans and stored on servers. Ascent facility Company 3 used this model in 2005 on The Island (see the July 2005 issue of Millimeter), and company officials say that success offers a template for future projects.

“Michael Bay used our new 2K digital cinema projectors to color correct the previews right from dailies,” says Cohen. Although Bay had previously done all his premieres on film, he agreed to a digital premiere for The Island off a QuVis Acuity server. “That version was conformed and color corrected directly from the original 35mm negative [which was] scanned at 2K, conformed at 2K to final picture versions with dropped-in 2K visual effects, and then married to 24-bit audio to take it around the country,” says Cohen. “That's a direct data path for most of it, from DI to the digital cinema master and the release negative.”

Reiner Knebel, VP of digital intermediate services for Ascent Media, emphasizes that Ascent is also developing other tools for the dailies and DI processes to permit closer collaboration between directors, DPs, and colorists throughout the filmmaking process.

Among these developments, Knebel reports, is what he is calling “an Alpha version of a digital printer light system.” Designed to permit DPs to communicate in the language of printer lights as they do in the photochemical world during a digital color-timing session, such a service could be invaluable for digital dailies work, Knebel says. He expects Ascent to make the process available to filmmakers in the next few months.

“In our case, we are working on a system for the scanning process — a new software option for the Spirit 4K that can be used during the transfer of dailies,” Knebel explains. “Our goal is to send metadata with digital dailies that can replicate printer numbers for the DI process later on. This permits increased feedback to the DP during the dailies transfer process — letting him know how his lighting affected film exposure. DPs have never had access to this feedback mechanism in the digital world. It will enable the DP to easily migrate from traditional photomechanical film dailies to digital dailies. Secondly, it will make it easier to maintain and apply color correction settings from the dailies process during the eventual DI.”

Of course, no dailies or DI system will yield maximum results if those creating and approving the imagery don't have a consistent way to view the images at the best possible resolution and color. This is a growing issue of concern in the industry these days. Viewing imagery from remote locations has been going on for a long time, but mainly to judge everything but color.

In November 2005, Ascent announced its Monitor Watch initiative — a software-based technology that permits remote collaboration during dailies or DI work through closely calibrated monitors. Company officials say Monitor Watch will initially be part of Ascent's UP Satellite remote collaboration service, available at its facilities in Hollywood, Santa Monica, and New York, and outposts at various affiliated facilities in cities including Miami, Minneapolis, Dallas, Chicago, Boston, Cleveland, and Richmond, Va. Company officials say the company has equipment and technicians available domestically and internationally for the service, and they claim to be able to expand the service to any location in the world reachable by the footprint of Ascent's satellite technology within a few days.

“Basically, we'll have engineers at different facilities who can send signals directly to monitors in remote locations — monitors that have a special probe attached to a remote PC that measures and analyzes the test signal sent by satellite from our facility,” says Knebel. “The PC will then send a signal back to our engineer via an Internet connection, giving the engineer remote control of that remote monitor. He can analyze and calibrate it as needed to match the one in our suite. This will let a DP supervise telecine color correction or a DI session from a remote location, and everyone can be 100 percent sure they are seeing the same, exact image.”

Incremental Infrastructure

Besides the inhouse development of proprietary tools, dozens of facilities are also integrating existing technologies to make it easier to view, share, manipulate, color correct, store, and remotely comment on high-end digital dailies within the context of a larger digital intermediate workflow.

As Hart points out, such tools have been readily available for some time now, and he suggests it's important for major facilities to mix and match them to offer multiple solutions. “[Facilities should do this] so that filmmakers have options, depending on what is right for their project, rather than forcing them into one, single, proprietary method of working,” he says.

“Our view is that you don't need a big consortium or virtual private network to load dailies at some faraway destination and deliver them to whomever on the project needs to receive them anywhere in the world,” says Hart. “There are a range of solutions for moving images as data around the world for those that require it. Right now, we are using the Copper service from [Markham, Ontario's] Digital Rapids for this work. It's a software product that lets you create folders with your show name on it, encode dailies with MPEG-2 or VC-1 or whatever format you want to use, put that media in the folder, and have the software automatically encrypt it to mathematical data symbols for safe travel over the Internet. We started using it a couple years ago during our work on I, Robot to receive and send digital effects frames back and forth to Weta Digital in New Zealand. Now, we are using Copper for dailies and DI collaboration, as well. It's a very secure, encrypted type of transmission, and it has no Internet latency issues whatsoever — the data will travel at whatever speed your connection happens to be. That's essentially a virtual private network over the Internet.”

Hart hastens to add that any Internet-based system obviously depends on connection speeds on both ends, meaning that transmission time requirements may limit the viewing quality of images. Higher-quality images mean larger files, and a low-bandwidth receiving connection may require the use of smaller files. But he says virtually everything is improving so rapidly that these concerns will lessen or disappear altogether in the near future. Likewise, he suggests remote monitoring technology is rapidly evolving to the point at which color decisions will eventually be made in the field.

“The device used to view files may make it difficult to make hardcore color decisions right now, since most monitors in the field do not have the quality of monitors here in our facility,” he says. “But even that is changing. The new [HL-R5688W] DLP 1080p monitors from Samsung [high-end, HD televisions being marketed to consumers for home viewing], for instance, have incredible quality. Blacks are black, reds are generally accurate, and greens are very good. So, at least, they can do reference viewing right now, and pretty soon, they probably will be good enough to make some color decisions, as well.”

He adds that facilities like Modern can also encode digital dailies for streaming or DVD viewing far more efficiently these days with the advent of a variety of new technologies. In Modern's case, Hart says, the company recently started using a realtime Windows Media 9 encoder for HD dailies — Inlet Technologies' Fathom solution.

“That technology is capable of putting an hour of HD onto a standard DVD, and players are available for viewing off the Internet for under $300, so that is a very effective tool that was not available to us just a year or two ago,” he says. “You can have 1080i dailies at about 8Mbps that are very good.”

In terms of technology to improve DP-colorist interaction by simulating a film look for digital dailies, companies like Modern and most other major players in the industry still prefer to offer multiple options to clients. Among the popular options in use these days is everything from the 3cP program from Gamma & Density and Kodak's Look Manager system to the DP personally color correcting digital still photos in Photoshop as a template for colorists.

Hart suggests that, while it's always a good idea for facilities that are large and sophisticated enough to have their own color science department and a staff of LUT builders, off-the-shelf solutions like 3cP and Look Manager remain a more feasible choice for most productions.

“We use Gamma and Density and Look Manager, and they work very well for most applications,” says Hart. “On other occasions, it remains more of a situation where the DP and the colorist work directly together using still photos. Such approaches can get filmmakers and our colorist into the zone, so to speak, and we can get the imagery pretty close to what the filmmaker envisioned. We are working to smooth the workflow and make it a more consistent process by being proactive and not reactionary as we develop techniques.”

Dailies to DI to DCI

All of these developments are about “seeking a bridge between the dailies process and the DI process,” according to John Nicolard, head of digital production at FotoKem, Burbank, Calif. That venerable film laboratory has offered all sorts of dailies services in recent years, and also recently jumped into feature film DI work, making its own strides in improving the synergy between dailies and the DI process.

FotoKem also has periodically used Look Manager and 3cP, among other products, to color correct digital dailies for clients to illustrate how they might look on film, and the company is looking into a variety of other approaches to the problem.

“We are also exploring ways to incorporate [FilmLight's] Truelight color management system [for previsualizing film images on various types of electronic display devices] into the digital dailies process to provide at an earlier stage the same kind of color control that we offer during the DI process,” says Paul Chapman, senior VP of technology at FotoKem. “We've been using Truelight for DIs, and now we are starting to use it for telecine, as well. Our next stage would be to extend it into the field for dailies. We're analyzing how we could deploy the system in the field with Truelight filmout simulations pre-baked into the imagery, or by using the 3D color cube [generated by Truelight software to emulate the filmout path] in the field.”

FotoKem officials say the company is also closely watching and supporting the ASC Technical Committee's ongoing work to create a standardized color decision list (CDL), which would eventually permit the consistent movement of CDLs from one platform to another.

That work, says Chapman, is crucial because it starts to address another of the key problems in achieving that seamless data path for image management that industry insiders yearn for. That problem: the difficulty in moving data directly from one platform to another as production moves forward.

“We talk to manufacturers about that all the time,” Chapman says. “The translation from one software program used for dailies to another for the DI is problematic because they don't talk to each other. If the DP uses Photoshop for dailies and we use a [Quantel] iQ for DI color correction, it's tough to have them talk to each other with the same files. Discreet has one way to color correct, Kodak has another, Da Vinci another — and, right now, you can't just slide the data directly from one platform to another. Therefore, everything is a re-creation of what came before. Besides the storage issue, that is the other big thing that needs to be addressed, so we're happy to see the [ASC Technical Committee] working on this problem.

“Their work is intended to address this issue of cross-platform compatibility so that we can make relatively simple color corrections out in the field that can be interpreted by whatever system you choose for your DI. The idea is that we would ship out to the dailies location the image files in some kind of proxy form, probably, and the DP would adjust them and then ship back to us a recipe we can interpret and use to guide the color correction down the road.”

FotoKem officials point out that these solutions are part of a larger paradigm shift designed to bridge the gap not only between digital dailies and the DI, but also between the DI and whatever might come next, including ensuing digital cinema releases. FotoKem, for instance, in collaboration with Doremi Labs and Christie Digital Systems, recently participated in a special screening of Universal's Serenity at USC's Entertainment Technology Center in Hollywood. That screening represented the first time a feature film was displayed to an audience directly from a master that was compliant with the DCI digital cinema spec.

The movie was played through a Christie digital projector as stipulated by the DCI spec, using the JPEG 2000 (4:4:4, 12-bit color) codec. FotoKem officials point out that the demonstration proved, among other things, that it could move 2K data files directly from the finished DI process and conform them for the DCI spec.

“We applied the color management used on the film's DI originally to the digital cinema package, with the only difference being that we also converted it to XYZ space to translate [the files] into what was necessary to show them on the digital cinema screen,” says Chapman. “But the important thing from a DI perspective was, here was a situation where the DI was done inhouse and all files were accessible on a network that made it easy to get the files and do relatively straightforward manipulation on them to create a digital cinema distribution master. Once the images are in data form and accessible, it becomes easier to repurpose them in different forms. The ultimate goal is to start that process once they develop the look in the field, and then apply that to subsequent versions of the data as it travels downstream through our network for whatever the next application might be. That's what we are all working toward.”

© 2008 Penton Media, Inc.