Grading the Graders: The Long Awaited Part 2

In reviewing Part 1 of this article, I did notice that I left out one of the more important “pros” of the Baselight system, and that is its directly integrated support for the Truelight color management system. Truelight, like Baselight, is a Filmlight product, and  Baselight is designed around the use of that technology for all color management functions, a great advantage. Now…..

Continuing right along where we left off…..

Digital Vision Film Master

First, a disclaimer. I have very little personal experience with the Digital Vision line, although I know the manufacturer well, and therefore my comments here should be taken as something of an informed observation rather than a direct, hands on report.

Born out of the original Nucoda line, Digital Vision’s product is a combination of the original Nucoda Film Cutter (originally released as a Media Composer-like conforming platform for film DI work) and a comprehensive set of color tools developed largely under the guidance of veteran colorist (and trainer) Kevin Shaw. Unlike many of the other devices being discussed here, Film Master is at its heart a conforming program with a lot of other things, including but not limited to color grading tools, added, rather than the other way around, as is largely the case with Resolve and Lustre. The conforming interface is more similar to Avid’s Media Composer than any of the competing products, which for some is a help and for others a hindrance. But is is a very effective and efficient system in the hands of a trained operator.

PROS:

1. Film Master has a timeline that supports multiple tracks, much like nonlinear editing programs, allowing very flexible conforming without an undue amount of work to collapse timelines that make use of more than one video track.

2. Its history as a conforming platform ensures that conforming is accurate, fast, and flexible, and accepts modern list formats such as AAF and Apple Final Cut Pro XML. Multilayer dissolves are handled easily and allow the colorist to adjust each layer separately.

3. Support for the Red Rocket accelerator card allows for very efficient use of that format.

4. Digital Vision is an Avid partner, and Film Master has very good support for Avid files and media formats.

5. Digital Vision’s very highly regarded image cleanup tools (descended from their DVNR and DVO products) are available as an integrated part of the toolset, such as dirt and scratch removal, geometry correction, and flicker correction.

CONS:

1. The somewhat busy interface is loved by some and, well, not loved by others.

2. It has a lower installed base, particularly in Los Angeles, than most of the other products mentioned here. Because the learning curve is reasonably substantial, this somewhat limits the available talent base in the largest market.

3. Its history causes many to think of it as a conforming tool first, and a color grading tool second. This might not be fair, but it is true.

Quantel Pablo

Unique among all of the systems here, the Pablo is basically a Quantel iQ with a rather deep color toolset grafted to it. The iQ was a compositing and finishing platform that was originally released by Quantel to be the successor to its Editbox and Henry lines, themselves descended from its original Paintbox device. They later added a color module to the iQ (known at the time as QColor) and, having gotten a good reaction from facilities and colorists, decided to do a more elaborate version that became what we now know as Pablo. At the time of its release, Pablo had something of a disadvantage compared to some of the other large scale DI systems due to a number of factors, not the least of which is its rather unique, Quantel style interface, common in Europe, but considerably less so in the U.S. It really needed something to distinguish itself, and it found it when stereoscopic 3D work began to emerge as a force in the industry. The pure real time playback power of the platform (which includes and requires its own dedicated, proprietary storage modules) allowed it to be quickly positioned as a system capable of the real time dual stream playback required for 3D work, and Quantel quickly added some basic functionality to allow for quick adjustment of things like convergence and keystone correction. It has more recently expanded its arsenal of stereoscopic specific tools and gained a foothold as a mature platform for this type of work.

PROS:

1. The configuration of the system as an integrated platform, complete with high performance proprietary storage, allows for a very high degree of interactivity, particularly with multiple picture streams, and at very high resolutions. Pablo is one of the few platforms capable of delivering real time performance with uncompressed 4K image sequences (Baselight 8 being another).

2. Because it is essentially an iQ underneath, all of the tools available for things like compositing (including 3D compositing), painting, image manipulation, and a multilayer timeline are present in Pablo. This allows for its positioning and use as a “hero” device, capable in the right hands of doing just about everything required for a digital intermediate.

3. Image “fixes” during a color grading session are common and very easily accommodated.

4. The stereoscopic toolset is quite evolved and being constantly updated with features like depth isolation for color corrections, allowing “near” and “far” objects in a 3D stereoscopic project to be color graded separately, even if they are the same color, based on their calculated “distance” from the “camera.”

5. Pablo is one of the few “high end” devices that directly supports the Cineform codec.

CONS:

1. The use of proprietary storage provides a great deal of the source of Pablo’s power, but it is also an additional investment that is not available to other devices in a facility. It also requires moving projects on and off the local storage, which is done rather quickly, but is still an extra step when compared with direct access of files via a SAN or other network attached devices.

2. Since Pablo is basically iQ, it was not designed specifically for colorists, although a lot of that weakness is obviated by its control panel design, which is quite colorist friendly. However, to access all of its power and make it a true “hero” device requires either multiple operators (i.e., a colorist, a conforming editor, and a visual effects or finishing artist) or a very, very highly skilled and versatile individual. Having all of this power in one box means less opportunity for concurrent workflows, and a bit of a bottleneck when revisions are required. Quantel does, of course, offer smaller systems for this purpose, but they are not as quick or nearly as powerful.

3. Pablo is, for better or worse, probably the most expensive of all of the systems evaluated here. Now, it should be remembered that the pure cost of any device is really only relevant to what it can produce in terms of revenue. But the rather high cost of Pablo means that in order for it to be profitable, it must be in operation for a lot of hours at a rather substantial hourly rate. This is, in general, only possible in the industry’s major markets, such as Los Angeles, New York, London, and to a slightly lesser degree, Vancouver, Toronto, Sydney, and some Asian cities.

Assimilate Scratch

It is somehow strangely appropriate that we move from the most expensive of all of these systems (Pablo) to one of the least expensive in Assimilate Scratch. But cost is far from the only or even the dominant determinant for an appropriate equipment choice in a rapidly changing technically driven industry.

Like the Digital Vision Film Master, Scratch was originally introduced as a “data management” tool whose primary usefulness was twofold: first, playing back image sequences reliably enough for real time evaluation, and second, conforming those image sequences for finishing in a digital intermediate environment. With a solid playback engine, and a good suite of conforming tools, it gained a foothold before anyone considered using it as a general purpose DI platform. Over the years, Assimilate has added a pretty complete set of color tools that have allowed them to reposition Scratch as a “DI in a box.” In the last 3 years or so, they have also forged a very close relationship with Red (they were one of their original partners, and designed the original Red “Redcine” program, based on Scratch code), which has allowed them to become something of a system of choice for many newer entrants into the DI game that are, for lack of a better description, somewhat Red-centric. The basic program design also allowed them to rather simply incorporate basic support for 3D stereoscopic work at a time when that was not widely available. In general, Scratch is a very comprehensive toolset that does a number of things pretty darned well, at a very reasonable price, and running on commodity hardware. A nice combination for today’s market.

PROS:

1. Scratch runs on commodity hardware under Windows 7 (64 bit). This allows some flexibility in terms of the hardware necessary for different levels of performance.

2. Early support of Red put Assimilate in a very good position, both in terms of being able to handle the newest file formats and color settings during Red’s development, but also in terms of supporting their Red Rocket hardware. Scratch systems can use the Rocket as an integrated accelerator, yielding real time decoding and debayering of 4K R3d files while color grading. But they can also use multiple Rocket cards as a transcoding accelerator, allowing faster than real time transcoding to formats such as DNxHD for dailies purposes.

3. Scratch supports reading and writing of Avid DNxHD MXF files.

4. Conforming is quick and simple, with either side by side or overlay playback of multiple streams, allowing real time matching of the conformed sequence to the offline reference.

5. Assimilate has demonstrated real time support for Arri’s ArriRAW file format as implemented for the Alexa camera.

6. Scratch has nicely implemented video style scope displays that can operate in real time, somewhat obviating the need for external scopes in some situations.

CONS:

1. Scratch supports only a single video track in its timeline and does not directly import AAF or XML files (EDL only).

2. Although the product supports multiple LUT implementations, and has multiple slots at various locations in the pipeline for LUT application, it does not have any integrated color management system.

3. Unlike all of the systems discussed to this point, Assimilate does not have or supply any custom control panels for color grading in Scratch. They do support a number of third party solutions (Tangent Wave and the CP200 line of panels are both supported), but colorists will always have a limited amount of direct access to program features as a result. This has, to some degree, limited their deployment as a color grading system in large facilities where efficiency is king and such panels are expected.

Iridas Speedgrade

Iridas is unique and interesting among all of the companies I’ve discussed in that they seem to be a technology company first, and a products company second. This might sound a bit strange, but if you examine what Iridas offers the industry, their strongest contribution seems to revolve around their ability to do high quality, real time debayering of multiple RAW formats, as well as their ability to transcode those formats into other formats efficiently and repeatedly. Iridas originally made their name (at least in the US) in the visual effects industry, producing one of the first high quality, memory driven image sequence viewers (FrameCycler), thus solving a need to be able to view film resolution frame sequences in real time on the desktop computers that were being used for both CG animation and compositing. This evolved into a disk driven playback system (FrameCycler DDR), and eventually into the Speedgrade product, as well as the Speedgrade DI line. Used only sporadically in the Los Angeles market, Iridas’ post products seem to have greater traction in Europe and other markets, as well as a bit of a niche presence in the on-set tool market with their Speedgrade On-Set product. Significantly, Iridas also supplied the basic technology that powers the Silicon Imaging 2K camera, allowing for real time in-camera color management that flows right through to the post process when using Cineform RAW files, a workflow that other camera manufacturers are still trying to figure out.

PROS:

1. Iridas’ close relationship with both Silicon Imaging and Cineform allows for a very nice and comprehensive workflow, including a nicely integrated 3D stereoscopic toolset. Settings created during production are directly accessible during post when Iridas’ tools are used.

2. Speedgrade DI supports a multitrack timeline, and the design of the program allows color grades to be specific to a particular clip, a particular layer, or multiple layers. This allows for simple trims of entire timelines in one step, a very useful feature.

3. Stereoscopic support is pervasive in Iridas’ products, allowing simple matching and tracking of left and right eyes through the post process.

4. Uniquely among all the products discussed here, Iridas’ products are cross platform. They are available not only on Windows and Mac OS X, but also on Linux.

CONS:

1. Iridas’ interface is like no other. Its non-standard nature has some advantages, but it does present something of a unique learning curve.

2. Speedgrade has limited support for things like variable softness on shapes and motion tracking, common on all of the other systems mentioned here. The color toolset in general is a bit more rudimentary than much of the high end competition.

3. Like Assimilate, Iridas does not manufacture a custom control panel for use with Speedgrade. Much, if not most, of the activity in Speedgrade must be handled through direct manipulation of the GUI, not the most efficient way to handle longform color grading.

And there you have it, my first and best attempt at giving a bit of insight into what all of these systems have to offer and, to some degree, why. It should be pointed out that at no point in this comparison did I refer to the actual color manipulation capabilities of each system, and there’s a good reason for that: they can all do pretty much everything you need to do. Some are stronger at some things than others, and some bring clever tools to the table that a talented colorist can use to their advantage. But they can all do the basic work of manipulating color at a high quality level very, very well. Any of them can be used to create a high level color grading room. Nor have I mentioned the particular operating systems these programs run on, except where it was specifically relevant. But the most critical component of any room is, and has always been, the colorist. Sure, if you’ve got some specific problems to fix that one or another system has specific tools to deal with, that system might be preferable. But in general, it’s the colorist, not the system, that achieves the final result. The tools sometimes make it easier, but that’s for the colorist to deal with, not the client. The client is only interested in results, something to keep in mind the next time someone asks “should I go to a facility with a Lustre or should I go to one with a Baselight?” – because the answer is always the same. They should go to a facility with a capable colorist whose talents are appropriate for the particular project. It’s really that simple.