ONLINE EXCLUSIVE
 The Nokia Theatre decked out for the SIGGRAPH computer animation festival. photo: ACM SIGGRAPH 2008 |
Geeking Out at SIGGRAPH 08
by Michael Kunkes
The 35th SIGGRAPH conference, held in mid-August at the Los Angeles Convention Center, is what it always has been since it was first held for a small group of CG practitioners back in 1974—a professional conference, not geared for the public. On the other hand, there was much to love at this show for the CG geeks who weren’t invited to the E3 video gaming show. Haptics were hot––technology which interfaces with the user’s sense of touch by applying forces, vibrations and/or motions. There were gloves that track the motion of your hands in 3-D space and give you tactile feedback when you collide with surfaces; magnetically levitating handles with no friction and strong feedback forces; and even a glove that gave the sensation of ants or roaches crawling up your extremities.
 The exhibit floor at the LA Convention Center for SIGGRAPH. photo: ACM SIGGRAPH 2008 |
Then there was “Rome Reborn,” an exhibit from the Institute for Advanced Technology in the Humanities at University of Virginia that used a variety of techniques to take the ruins of Rome during the reign of Constantine and reconstruct them in virtual 3-D, using handheld GPS-enabled devices to view the virtual city. Intel presented a paper on its emerging Larabee GPU architecture, soon to be unveiled in high-performance graphics cards. And Microsoft demo’d a new research technology called “Unwrap Mosaic,” which it predicts will someday allow video users––professional or prosumer––to edit artifacts into video as easily as they do with digital photographs today.
A lot of focus at SIGGRAPH 2008 was on stereoscopic 3-D––acquisition, processing, projection, viewing, theory and practice––and its emergence as a new language of filmmaking. One area that has always benefited from emerging technologies has been media storage, and now, in these large, high-performance 3-D multi-purposing data systems, many storage manufacturers are starting to move away from standard RAID-based data protection and recovery into variants designed to improve performance, especially during data recovery and storage addition.
Aberdeen LLC of Santa Fe Springs, CA ( www.aberdeeninc.com) showed its line of flexible Stirling storage servers that expand from a basic storage server of up to 40 TB, expandable through additional daisy-chained JBOD and XDAS expansion boxes to a total of 328 TB, all on one self-contained unit controlled by dual Quad-Core Intel Xeon processors and without the need for additional controllers. “We offer scalability, “ says Aberdeen senior account executive Tim Staus. “You can keep adding 16 Tb external JBOD servers and keep on going. Our servers are being used in numerous post- production and TV studios, in addition to surveillance and medical image scanning uses.”
Isilon Systems (www.isilon.com) featured its new X-Series clustered storage system, introduced early this year. Achieving 100 times the scalability and 20 times the performance of traditional SAN and NAS storage, Isilon is he first storage system that scales to greater than 1.6 petabytes and provides 10 gb/second performance in a single system and volume. The X-series is the newest evolution in the Seattle-based company’s IQ series of solutions, all powered by their OneFS OS, functional with Windows, Mac and Linux. Among Isilon’s entertainment clients are Sony, NBC Universal (including the recent Beijing Summer Olympics), and key visual effects firms such as the Orphanage, Intelligent Creatures, and James Cameron’s Lightstorm Entertainment, which is using Isilon to power the workflow of Cameron’s upcoming 3-D feature Avatar.
The core of Isilon’s clustered storage systems is the operating system volume manager, which basically combines OneFS software with a scalable number of storage nodes to create what appears to the user as one giant, but simple volume. Then, a suite of proprietary Isilon applications such as Sync IQ, Snapshot IQ, Migration IQ, SmartConnect, SmartQuotas, etc., can be licensed on a per-node basis to provide a range of different functionalities. “Because we stack data vertically across nodes instead of in a standard array setup, a minimum of three nodes is the smallest configuration that can be purchased, and 96 nodes is the largest, providing an overall range of 4TB to 1.6 petabytes,” says Isilon’s Paul Barr. “We can add a node into the cluster and scale it to a volume in about 50 seconds,” he says. “No one else can do things that easily; it takes other manufacturers much longer to rack, stack and cable the product.”
According to Barr, each node can be individually managed from the Isilon web interface. “The software contains a performance monitor so you can look at how much space each node is using and what I/O is running,” he says. “In a production environment, it’s important for a client to know how many scenes are rendering and editors editing and know that they aren’t dropping any frames and that things are running smoothly. This interface gives you the complete ability to manage and monitor that information. There’s also a protection level, which means that if a node fails, data is still available across multiple layers. You can also mix Mac, Windows and Linux files and be able to access the same information without any special software requirements.”
Omneon ( www.omneon.com) showcased its MediaGrid active storage system, designed to aggregate the performance of multiple nodes to combine petabytes of storage with multiple gigs/second of bandwidth. Omneon Regional Sales Manager Dwayne Brady reports that MediaGrid utilizes what it calls “Media Intellignece” to optimize performance for leading edit platforms and enable an “edit-in-place’ functionality that allows content to be accessed and modified while it is being transferred into a specific platform, putting it more squarely into the active workflow. MediaGrid uses the company’s concept of Dynamic Data Redundancy (DDR), which is immediately replicated one or multiple times once content is written to the system. MediaGrid also utilizes Omneon’s own ProXchange, a transcoding system for a variety of DV, MPEG-2, SD and HD files formats contained within MediGrid.
In other exhibitor news, Stockholm-based Digital Vision (www.digitalvision.se), a leading provider of picture enhancement systems for film and television post-production, showed the latest version of its color grading and finishing system, Film Master, as well as Phoenix, the company’s restoration, remastering and repurposing asset system for creating finished deliverables for digital cinema, high-definition transmission, Blu-ray and VOD services. The company also announced that all its software systems now support footage captured with RED One cameras. On RED’s SDK, users of Film Master will be able to grade native Redcode Raw files.
Eyeon Software, Inc. (www.eyeonline.com) presented Fusion 6, a major upgrade to their flagship compositing application, celebrating its 20th year. With this release, Fusion features a more consistent environment for seamless image quality across multiple platforms, and is available for a single license fee on Linux, Windows x64, Windows 32bit desktop PCs, and Intel-based Macs, with upgrades available at no charge for existing customers. New features include the ability to view stereoscopic pairs in 3-D, support for stereo monitors with eye separation and convergence controls, multi-layer imaging for composition node flow to process multiple images, an all-new 3-D system (being released as part of Fusion’s SDK), production-quality Open GL resolution-independent and floating point rendering, a host of 3-D tools that support both 3-D effects plug-ins as well as 2-D blend modes, and a ROIDS (Region of Interest Domain System), that allows users to limit image composition to specific areas of the image.
The company also gave a comprehensive introduction to its brand-new Generation system (first shown at NAB and shipped in September), a fully-featured collaboration, conforming, 3-D stereo compositing, versioning and editing environment, designed to be an end-to-end solution for post-production houses and to integrate with all of eyeon's current visual effects applications, especially Fusion.
The fully scriptable Generation line has three iterations: Suite, Studio and Player. Generation Suite ($9,995) can manage and administer projects that range from a few filmmakers in a small facility, or spread across thousands of participants in multiple international sites. Suite contains all of the compositing tools in Fusion, with seamless interoperability, 4K playback, EDL import and limited timeline editing functions, and the ability to bring together and annotate shots from the entire project. Generation Studio, priced at $5,995, contains all of the editing, versioning, annotation and collaboration tools found in Suite, but omits Suite’s 3-D stereo compositing abilities. In either case, the linkage with Fusion allows artists to load EDLs from various editing applications and then move shots back and forth to Fusion and have the updated versions automatically appear in Generation, pointing back to the original compositions as well.
At $695, Generation Player is the product line’s everyday, bare-bones 4K stand-alone player; it can open any version of a shot’s history from a Fusion or non-Fusion workflow and insert the current version for viewing, giving Fusion artists and other personnel, the ability to play back the latest version of the project, annotate, then bring it back to whoever is supervising the original workflow.
The prices on the Generation family (plus subscriptions) are probably set low enough to help drive Fusion sales at studios, as well as to provide studio shoppers with a lower-priced alternative to systems such as Imagineer’s Mogul and Assimilate’s Scratch. “Generation is designed to take away the bottlenecks at visual effects facilities that prevent efficient communication between 3-D, animation, effects and editing departments, ” says eyeon’s senior director of sales Michael Bailey. “From storyboards to animatic to final, as a shot progresses, you can see the progression on the timeline of any shot tied to each piece of the timeline. Generation allows multiple Fusion users to collaborate visually or through text, using all the metadata tied to a particular shot.”
Finally, AJA Video Systems (www.aja.com) featured the Io HD, the latest version of the company’s light (7 lb.) transportable stand-alone hardware plug-in solution for capture and conversion over FireWire 800, to Apple’s ProRes 422 codec, enabling true full-raster 10-bit video editing with a MacPro desktop or MacBook Pro laptop. The Io HD uses many features from the company’s KONA card lineup, including broadcast-quality SD to HD up-conversion, HD to HD cross-conversion, HD to SD down-conversion and HD/SD component analogue output.
“All the I/O is separate, so you configure your conversions however you see fit,” says AJA product manager Jon Thorn. “Our user interface is set up so that if you want, for instance, to do a down-conversion, you can do it for any kind of separate output––letterbox, cropped or anamorphic. And all that processing is done in hardware.”
The Io has added some free software driver updates since NAB, including the ability to use the box in complete stand-alone mode. “Basically, you tell it what kind of a conversion you would like to do, then disconnect it from the computer––and it remembers the changes until the next time you plug in to a Mac.” Also new is a function called audio slewing, which will intentionally delay audio to match the problem of plasma monitor video delay, something that is not an issue on CRTs. “We are really trying to make a flexible product with intuitive controls and as much interface information as we can,” says Thorn. We rev our products to address user needs, not to make a product obsolete for the sake of marketing.”
Michael Kunkes is a freelance editor and writer specializing in animation, production and post-production. He can be reached at writermk@sbcglobal.net.