Digital Intermediate 2005
Feb 15, 2005 5:22 PM, By Michael Goldman
More Millimeter articles from the Digital Intermediate archive
PDF of Digital Intermediate 2005
Colorist Crossover
Digital Post Processes for Feature Films Migrate to the Commercial World
Few artists have personally experi- enced the intersection between the worlds of feature fi lm and commercial color correction in recent years more directly than Stefan Sonnenfeld, president, managing director, and colorist at Ascent Media's Company 3 in Santa Monica and New York. Sonnenfeld has become extensively involved in digital intermediate work on major features in recent years, while still remaining one of the top high-end commercial colorists in the country.
Sonnenfeld suggests the rapid acceptance of the DI process in the movie world has had a simultaneous impact on commercial color correction. Specifi cally, he says, some high-end commercial telecine facilities are now offering high-resolution data scanning, assembly, and color correction. This data-fi nishing alternative, Sonnenfeld suggests, is being introduced to commercial clients at a time in history when HD market demand is starting to escalate.
Digital Post Processes for Feature Films Migrate to the Commercial World
"I've already noticed a signifi cant rise in high-def fi nishing, meaning we are coloring, fi nishing, and delivering more HD spots than ever before," says Sonnenfeld. "The increase is actually very dramatic on high-end commercials. I'm sure a big reason for this is the awareness on the part of the advertising agencies that HD formats can be very useful for the multiple, ancillary delivery requirements they are now seeing, like cinema spots or PAL or 16x9 anamorphic, for use in broadcast, inhouse, and in other venues. This notion of one pass, and you can deliver all formats from an HD master, makes a lot of sense in those circumstances.
"The DI process offers yet another way for customers to finish their spots. It is probably not appropriate for all spots, but it is certainly viable for the ones that have a little more time to finish, and which have already been accepted by the client. When you have client approval and you have prepared properly, a DI process makes sense. You scan the fi lm at 2K resolution, conform it at 2K, color correct it at 2K, and then down-rez the material into multiple formats for delivery. In other words: color correct the fi nished spot in the highest resolution you can. This is something that will permeate our market more in the future. It doesn't make sense as an ultimate workfl ow yet for all spots because the majority of them are still standard def and can't afford all these luxuries right now, but it works for certain commercials and will become more viable as time goes on."
Among other things, Sonnenfeld suggests that more high-end commercial facilities will follow the approach Company 3 has adopted—to invest in scanner technology and build a pipeline fl exible enough to handle short-form projects of different resolutions and requirements, while also amortizing that same equipment to push into the feature fi lm DI marketplace.
"We started on commercials and then got into feature DIs, but now, our feature experience has given us a tremendous advantage when working on commercials," he explains. "For features, you have to handle all kinds of high-end formats with opticals and speed changes, develop proprietary [look-up tables], and all those issues. We have that capability now, and that permeates back into the commercial market.
"Many clients now already have theatrical delivery requirements, and there are lots of other reasons why HD delivery makes sense. We have refi ned our color calibration system, and we can produce the same quality of color from a high-resolution master for theatrical, HD, or standard-def delivery. It's a crossover situation, and I expect more companies will try going in this direction."
Hi-Wire, Minneapolis, color-corrected this Target commercial at HD resolution using its Spirit Datacine-da Vinci 2k pipeline. The company is working to transition to a complete nonlinear DI approach for all its commercial work.
Sonnenfeld routinely uses his da Vinci 2K Plus color corrector in tandem with a Spirit Datacine for both commercials and features, but adds he is studying da Vinci's new software-based color corrector, Resolve, as well as others from various manufacturers, including Discreet's Lustre.
Company 3 is also promoting Ascent Media's UP Satellite "virtual telecine" technology extensively—making the facility's tools and talent remotely available to clients in markets far from the company's main locations. (At press time, Company 3 officially announced UP Satellite is now available through its New York facility, in addition to Santa Monica.) He concedes that many boutique commercial facilities cannot offer such high-end technological innovations, but even in those situations, he suggests a slow evolution toward some sort of a DI infrastructure is inevitable, sooner or later.
"With commercials these days having multiple delivery needs, like Super Bowl spots, for instance, it has made sense in our facility to have this kind of infrastructure now," he says. "It might not make sense, though, for a boutique or a company in a smaller market right now. But ultimately, we are all working toward the same goal—to have fi lm or HD material loaded into a big server that multiple workstations can access and grab the images to color correct, dirt fi x, conform, or whatever else might be necessary. It's currently time-consuming and expensive to do that, but we're investing in our technology and skills so we can to take advantage of these new technologies for our customers, and eventually, I think it will be customary to work on commercials in this way."
New York
Officials at Nice Shoes, New York, say the commercial facility has selectively transitioned to a DI process on certain high-profile commercials since acquiring a Thomson Specter FS Virtual DataCine in order to evolve its traditional film-transfer approach. Colorist Chris Ryan uses a da Vinci 2k Plus system in combination with the Specter virtual telecine, and claims this new approach has taken Nice Shoes into the next generation of fi lm-transfer workflow due to the improved speed available in digitizing 35mm fi lm images into the 2K data world or a wide range of video formats for color correction.
Pixel Farm used Discreet’s Lustre technology to color-correct this 35mm-originated commercial for Rapala fi shing lures.
"We've also developed a process that will allow us to do a DI job with the Specter in about half the time of a normal transfer," says Ryan. "But the film still has to be scanned into our server so that it is available to all of the colorists and graphic artists on our network. So it's a fine balance for our producers to manage all aspects of the project from beginning to end, and it's also important to get involved with a project at the very beginning. We've been doing lots of R&D with every major manufacturer involved with DI in order to customize our new workfl ow, and also to advise them on what we really need to do in the real world. Engineering has therefore become a major part of this process."
Once the film is scanned into Nice Shoes' network, however, Ryan says his job as a colorist now allows him to offer clients more creative options than in the past. The end result of this: the company's technical abilities to scan and deliver material at whatever resolution might be required for the job is greatly improved, according to Ryan.
"That's what's great about having the Specter—we scan all footage at the highest resolution the telecine can offer, and then complete the job using the DI workflow we have developed here at Nice Shoes. I also have the ability to show the client the end resultin any format or aspect ratio, without having to do alternate passes. But if the client wanted an HD version or a 4x3 NTSC version, it would only take the additional time involved in recording the color corrected material to that format. No re-transfer would be needed like it would have been in the past, and that's very attractive to agency clients."
Particularly attractive, says Ryan, is how this nonlinear process has the potential to improve the fi nal product creatively, allowing clients and vendors alike to explore more options in a compact timeframe.
"You can do a rough conform, and the client can see the spot played as we color correct it in our suite," he says. "That gets your mind into a more creative workfl ow, which means the client will get more time to fi nd little glitches and things they want to change. Also, you can spend time focusing on the look, and not technical issues, with no approach being right or wrong. You might want to do the effects work fi rst, and then conform and color correct at the very end, or the other way around. We've had such a good reaction, we have to book the Specter weeks in advance, and we are now in the process of buying a second one."
Midwest DI
Outside the coastal markets, the overlap between long- and short-form color correction has also started to inform technology purchases and workfl ow design. Dave Sweet has worked on commercial campaigns at Pixel Farm, Minneapolis, for years, working in the standard-def realm using a da Vinci Renaissance 8:8:8 system and a Rank URSA telecine. In recent months, however, he has seen his company and his workfl ow evolve as Pixel Farm initiated the early stages of its conversion to an HD workflow.
Company 3 president/colorist Stefan Sonnenfeld, who sees the high-resolution data scanning, assembly, and color correction DI workfl ow as the eventual standard approach for major commercial work at large facilities.
Photo by Duncan Stewart
That transformation began with the company's decision to include color correction in its Discreet-based digital pipeline, with the addition of Discreet's Lustre color-grading system this past June as a primary tool for commercial work. The company has already made extensive use of other Discreet tools in recent years for high-end graphics and effects work, and Sweet says it therefore made sense to add Lustre for color correction. "The next thing we are looking at is getting a scanner and moving right up to having the ability to do 2K scans."
The reason for this transformation was, according to Sweet, the company's desire to find an affordable way to be able to han- dle HD- and 2K-resolution projects as the commercial world changes, without jeopardizing its ability to handle traditional standard-def work. "In the near future, everyone will be working at 2K in a DI type of process [for commercials], so we wanted to upgrade in that direction, but with a software-based tool that was afford- able given our modest size," Sweet explains. "To be able to scan a film negative at 2K and work at that resolution is a big upgrade for a company like ours. What this is all about, essentially, is bringing my world—the world of color correction—into the nonlinear world. For years, my room was the most linear room in the facility, but now we are beginning to change that, which I think is justified, given what's happening in our industry."
Sweet adds that Pixel Farm recently installed an Apple Xserve Raid system with 5.6TB of storage dedicated exclusively to color correction work on commercials. He expects the company will eventually add DLP projection technology now that it's moving toward being able to perform DI work at 2K resolution, but says there is no immediate need.
A Geek Squad computer support commercial--typical of the type of spots that Hi-Wire, Minneapolis, routinely puts through its evolving HD pipeline.
Hi-Wire, Minneapolis, is also actively pursuing the so-called "nonlinear DI approach" for color correcting and fi nishing commercials. Hi-Wire senior colorist Oscar Oboza normally uses his da Vinci 2k Plus system for spots, networked with a Spirit Datacine. The company, however, is studying both Lustre and da Vinci's Resolve as it contemplates how to revise its pipeline to make the DI model affordable.
Tammy Kimbler Weber, managing producer at Hi-Wire, feels that the transition her company and many others are undergoing is similar to the conversion to high-defi nition infrastructures in recent years.
"That took a good three years for most of us to get up to speed, and another couple of years for everybody to speak the same language," she says. "Final delivery in our market remains standard-def, but many of those clients are still interested in the highest possible resolution. Many of them have their own inhouse and in-store private high-definition networks for displays and things, and they have multiple uses for the material we produce for them. Therefore, they are already open to fi nishing in HD. Probably 10 percent of our clients right now want to work that way, and that number will only increase. Also, more advertisers are creating branded content, like the BMW fi lm series or the Buddy Lee series from Lee Jeans. Our fi rst phase in this process is already done—we've been working at HD resolution since Hi-Wire fi rst opened in 1999, even if our delivery is still standard def. We already possess the ability to output 2K scans via the Spirit. It's just that the last piece of the process has been lacking—a nonlinear data-based (2K or greater) color correction system. Upon implementing that fi nal piece, we'll be able to do commercials as DIs whenever it makes sense to do so."
Oboza adds that he expects this conversion to help Hi-Wire get more into the feature fi lm world, since "you can't be pigeonholed anymore. To compete in a small market, you have to do more than just commercials or whatever. With features being shot in Canada, it might make sense to fi nish here on some of those shows if we can demonstrate the ability to work at 2K all the way through the process. When we get to that point, we'll be using the same tools for both features and commercials."
Creatively, Oboza feels this evolution will only benefit the high-end commercial colorist. "After all, if we work in nonlinear fashion, material can come through the pipe three to four times, and we can make more changes to please the client. That's exciting because the colorist will get more control or at least the opportunity to put a stamp on the look. It will make us a bigger part of the collaborative process. There will be a merging of the colorist's interface with the graphic artist's interface. They won't be the same process, but the toolset will cross over and the colorist will be more involved than in the past."
Hotel Rwanda: An Unplanned DI
Going into production on Terry George's Hotel Rwanda, producers had no plans to perform a digital intermediate on the film. The critically acclaimed film about the Rwandan genocide was color-timed traditionally at Technicolor London just in time to get it to the 2004 Toronto International Film Festival, where it won the festival's People's Choice Award. Despite the positive reception to the film, however, producers were not happy with how the imagery translated on the big screen, and they began to ponder what their options were—if any, given their limited budget and time frame—as they prepared the movie for a wide theatrical release.
Left, a shot from Hotel Rwanda from its original, photo-chemically color-timed master. Right, the same shot after the film went through the DI process. The DI reduced grain on the anamorphic version, and smoothed out skin tones, among other things.
"What happened was we shot the movie on Super 35mm and then converted it to an anamorphic presentation for release, which introduced a lot of grain when the image was blown up," explains Alex K. Ho, co-producer on the film. "We also had a big restraint in the schedule and the amount of money we could spend during production, and so, we couldn't [originally] color-time the movie in an optimum way to deal with the flaws that those limitations created in the original photography. In other words, by the time it got to Toronto, we never had it quite right, and I was concerned about what we could do to smooth it out and make improvements, as well as dealing with the grain problem."
Around that time, filmmakers brought the film to Post Logic, Hollywood, to start work on a home video master. Ho, collaborating with colorist Michael Underwood, decided to produce a short test portion of the movie, color corrected digitally from the home video release's HD master, and he then presented the results of that test to officials at MGM/United Artists, the film's theatrical distributor.
"We were already concerned because some of the original photography, shot fast and low-budget, was not as kind to the actors as it should have been after we originally color-timed the film," says Bruce Markoe, executive VP of feature postproduction at United Artists. "This was very obvious in our anamorphic release prints coming out of London. When Alex showed me the tests from the HD video master, I was really surprised. It wasn't a full 2K film scan, but their ability to balance out lighting and contrast issues was impressive. It really did improve the look of the film overall. Because they did the conversion from Super 35 to anamorphic digitally and not optically, it was a lot kinder in terms of grain problems."
Markoe and Ho then prevailed upon the studio to pay what Ho says was "a modest amount" to re-color-time the movie digitally from the HD master at Post Logic just before its theatrical release.
"We discarded the original color correction and went back to the original IP," says Ho. "This was very unusual, obviously—invention from necessity. We had the movie already scanned at Post Logic for the DVD project, and the difference in resolution was so minimal that it really didn't matter for our purposes. A couple of years ago, we never could have done this, but now it's affordable and the tools are sophisticated enough that it can be done quickly and efficiently, even in a situation like the one we were facing."
In addition to making the residual grain cost from the Super 35 to anamorphic conversion less painful, filmmakers say the DI paid off in terms of improving skin tone problems.
"There are lots of different skin tones in the film," explains Ho. "We have a red-faced looking Irishman, a pale-looking American, a dark-skinned black man, a lighter-skinned black woman. Everyone's skin tone is different, and it all changes. We needed to introduce a consistency to everybody's skin tones. This process made a big improvement in that regard."
Underwood adds that originally, "there was lots of red in the faces, and they couldn't pull as much of that out as they wanted to using a traditional lab timing process. With the DI, though, I was able to back that off with the da Vinci [2K Plus] color corrector. The ability to isolate faces in the color corrector without impacting other aspects of the frame, like green grass or whatever, is a big advantage. Also, some of the night scenes seemed too bright when timed photo-chemically. In those cases, I was able to darken the scene overall, but still isolate the main characters so that the viewer can see their eyes and expressions. The other thing we did was reframe some shots, and digitally, you can do that without losing a generation like you would get in the photo-chemical world."
For the Hotel Rwanda DI, Post Logic scanned the movie on a Spirit DataCine, color-corrected it on a da Vinci 2k Plus system, viewed the imagery in the company's digital theater on an NEC DLP projector, managed data through DVS servers, and recorded it back out to film on a Celco Fury film recorder.
Markoe suggests that the experience of performing an unplanned DI on Hotel Rwanda at HD resolution was instructive for filmmakers and the studio because it illustrated the "wider range of methods available in the DI suite" for feature films.
"This is probably the first theatrical film done this way, when it wasn't budgeted and planned up front," says Markoe. "We used a process that films originated in HD might use, but it worked out well for this project—everyone involved agreed it was a significant improvement. And I think it did contribute to the film's success because there is a more consistent tone to the imagery—it's less harsh. That means viewers can concentrate on the story.
"What I learned on this project is that a lower cost digital intermediate can save lots of money and has the potential to really improve the product, even if you started in film. The quality of the imagery really held up when we went back from HD to film, and I expect that conversion process will continue to improve as facilities improve their software and color correction tools. This process cost a lot less since we already had a digital IP from the home video project and did not have to scan from scratch. Therefore, the time spent in color correction was largely a single process for both the theatrical and home video versions, with just some additional tweaking done for the theatrical version. That's why the studio considered it cost-efficient and agreed to do this process.
Continue the discussion on “Crosstalk” the Millimeter Forum.









