find_photog find_assist join_asmp
 

Harnessing Technology with Computer Generated Imaging (CGI)


This Q&A with ASMP member and CGI master Walt Jones offers resources and insights for navigating the wide range of software options and the steep learning curve involved in working with Computer Generated Imaging.

 

Judy Herrmann: What steps would you recommend that a photographer interested in getting into CGI take right now?

 

Walt Jones: CGI can carry a pretty heavy learning curve, even for those who are technically-minded. The software is, in many cases, very mature and can quite literally offer up endless possibilities. Before you dive in, know what you want. Designing in three dimensions within a computer can often be a very time-consuming task. Those who can work quickly know what they want before they ever pop open Maya or ZBrush. That’s not to say that you can’t explore various options or refine your concept once you’ve begun the process of implementation, but having a clear concept at the start will make all the difference in the world. This is something I try to instill in everyone I teach or work with, and something that’s surprisingly all-too-rare in today’s industry.

 

For those interested in getting their feet wet in CGI, here’s where I’d start:

 

Find like-minded people. That would be the first thing. Attend user group meetings-organizations like the various member groups for Autodesk software (formerly the Alias Global Users Association), ACM SIGGRAPH (ACM Special Interest Group on Computer Graphics) and the ASMP (you’d be surprised how many ASMP people you can run into at group meetings who have their hands in CG) where you’ll network quickly with other artists at all skill levels and get great first-hand information on how to achieve the looks you want. Sign up for a few of the listservs on Highend3D (mentioned further below) and get insight into how to approach various problems and how to get the most out of specific software features.

 

The second thing is to start playing with the software. You can get trial versions of most of the major software packages — Maya, 3ds Max, Lightwave 3D, etc — that allow you to get your feet wet and play around a bit before you decide you’re ready to buy a full license. Different software packages have different interfaces and workflows that can make all the difference in how intuitive they are and how quickly you can accomplish certain tasks. For nearly all of my work I use Maya, primarily because it’s what I’m most familiar with and because it’s more or less the de facto standard in the visual effects industry for generalist work. That means that if I find myself in a tight spot, there’s a large network of people I can reach out to for knowledge or help meeting a deadline. However, applications like 3ds Max and Houdini offer up huge benefits depending on what you’re trying to accomplish. So, I like to at least have a basic understanding of the various strengths and weaknesses of the different packages. That comes from taking at least a little time to experiment in each and really feel out how I think each of these “tools” will work for me. From a personal standpoint, I also find that simply playing around in an application will introduce me to features and workflows that I didn’t even know were there. Those who enjoy more formalized training can pick up some of this from online tutorials, training DVDs, lectures, etc… but you’re not really going to know your way around an application or have a bag full of back-pocket tricks until you’ve gotten your hands dirty.

 

It’s also important to really understand how to “tune” a look using the various pieces of software. Nearly every piece of software now has a set of “buttons” that will create various looks and effects for you. These are a great starting point, but these will never get you much beyond that point. Every yahoo in their basement has access to those same buttons, so pushing them to create your final look won’t do anything to set you apart from the herd. The reality is also that the vast majority of the time, your client is not going to want anything that looks like the default settings in Maya for “fire,” so you’re going to need to know how to tweak every available parameter to really nail the look that’s required. There’s never going to be a “make it look cool” button, so invest in really understanding what’s going on behind the settings you apply.

 

At the end of the day, my advice is to be familiar with a bunch of packages, but know one piece of software really well. Having in-depth knowledge of how to attain a variety of looks and tweak them within your favorite interface will mean you’re able to not only give the client what they want, but you’ll be able to do it quickly without messing around in a package with which you’re not intimately familiar.

 

When I first started getting my feet wet in CGI, I literally gave myself personal assignments that would force me to explore different aspects the software. The timelines were always open-ended, giving me the opportunity to really focus on the look and learning a solid workflow rather than force something out under deadline. Once I started doing CGI work for hire, I had plenty of nasty deadlines, but my initial forays under my own steam had allowed me to understand what was needed to get things out quickly while still looking really polished. You will always wish you had more time to complete a project, so learning how to focus your energy in the right places and which “broad strokes” should be placed first are key.

 

Another thing — learn the lingo. The more you know the language, the better the questions you can ask and the more understanding you’ll have of what’s going on under the hood. Read up on radiosity. Understand the difference between true global illumination and final gather lighting. Learn why texture bit depth is so important in physical displacement for geometry. Find out why inverse kinematics are so key to proper character rigging. What the hell is FFT and why is it difficult to produce a decent fluid simulation without it? These examples are all more advanced topics than those I’d throw at someone who’s just starting out, but they highlight some of the language people are likely to encounter as they become further entrenched. This is another place where simply being immersed in things through listservs, online forums, user groups and organizations like SIGGRAPH will do wonders.

 

But remember… underneath it all we’re just talking about tools — tools that allow you to create imagery. You need to understand the tools in order to produce, but knowing how to use a hammer won’t necessarily allow you to design a house. You have to keep a big picture view with all of this stuff — it’s incredibly easy to become mired in the details and loose track of where you were going in the first place.

 

JH: Are there any web resources that you recommend for learning CGI tools and techniques?

 

WJ: There are a number of great online resources for CGI. Above them all, Highend3D (www.highend3d.com) is the one that I use time and again. Not only do they offer a huge number of tutorials and other learning resources, but they also offer some really good downloadable assets posted by other users — full scene files ready to render for various software packages, textures, scripts, plugins — it’s pretty exhaustive. Highend3D also has some great listservs, though most are specific to a particular software package (Maya, Shake, etc). They do have a general “rendering theory” list that’s great for people who are hard-core computer scientists or who want to know more about the math “under the hood” of computer graphics. Most photographers I know would find their heads exploding with that one.

 

There’s also www.cgtutorials.com, which is more of a “collector” with links to tutorials scattered across the web. What you’ll find there is quite a bit less technical and hard-core than highend3d.com. However, it covers a larger gamut, including tutorials for Photoshop, Corel DRAW, C4D and other applications that aren’t on Highend3D but can still be quite useful in attaining a certain look.

 

For some tutorials specific to software produced by Autodesk (such as Maya, 3dmax, etc), check out www.area.autodesk.com. Some very cool stuff can be found here. In fact, there’s a great tutorial posted just last month by Tomás Müller that walks through some of the process of a CG environment created for a Harper’s Bazaar shoot.

 

For those up for more formalized training, The Gnomon Workshop (www.thegnomonworkshop.com) offers pretty solid training DVDs for purchase as well as a smattering of free online tutorials.

 

Finally, for an overall roundup of online “learning resources,” check out ACM SIGGRAPH’s “Learning Resources on the Web” directory at www.siggraph.org/cgresources/displaycategory.php3?category_id=1.

 

The list is pretty exhaustive, with resources for nearly every level from those who are just getting their feet wet in CGI to those who are writing their own software. It’s a great place to start when you’re trying to find references on particular topics and aren’t getting much love from Google.

 

JH: How do you stay on top of new techniques and technologies? Where do you hear about them? How do you approach mastering them?

 

WJ: I’m very lucky due to the fact that, in addition to my photography, I work at one of the top five visual effects studios in the world. We’re constantly working on the bleeding edge of technology — a necessity to stay competitive in the market. Clients are always asking us for the next great thing, so we end up creating a lot of techniques and technologies as we go. It’s part of the process. For someone who isn’t right smack in the middle of the crucible, you’ll still find that challenges brought due to a particular aesthetic or schedule will require venturing into new territory. To stay current on things, I refer back to many of the resources we’ve already touched on-listservs, online communities and articles about what others are doing with CGI. Even just lurking on the listservs will expose you to the questions being asked by others and the various solutions being presented. More often than not, you’ll catch wind of a number of techniques and technologies that you’ll quickly work to implement in your own pipeline. Or, if there’s a particular challenge you know is coming down the pipe, you can throw out a “how would you handle this?” question and see what people have to say on the subject. There’s also the annual SIGGRAPH conference where you’ll find nearly a full week filled with lectures, classes, demos, exhibitors and a computer animation festival that showcases the latest and greatest in technology and techniques. You can attend a class on how Industrial Light and Magic developed realistic-looking skin for a particular film project, or how a group of scientists found a way to render realistic, physically-based lighting 1,000 times faster than anyone else had before. For anyone who hasn’t been before, it’s a pretty intense experience, with upwards of 30,000 attendees from more than 80 countries being quite common. Your brain will hurt by the time you leave, but your eyes will have been opened to a vast array of exciting new things.

 

In terms of mastering new techniques, it goes back to the idea of creating my own personal projects to test things out and really wrap my head around how best to tackle a problem. If I’ve found a new technology that I think is really going to make the difference to a client, the best way to sell them on it is to show it in action. If, for example, you’re hyped up on the new nParticles in Maya 2009 and think they be just the ticket for creating the effect of zero-gravity orange juice for a campaign you’re trying to land, do a test. Take something you’ve shot in the past, throw yourself into Maya, get the look you want and throw that image out to your prospective client. When they award you the contract, you’re ahead of the game because you already invested the time to figure out how to make zero-gravity OJ work. Even if you don’t land that particular client, you still have a great-looking image that shows off this new technology.

 

Much as the “what’s new” pages for new versions of software are the products of marketing departments trying to sell you on how revolutionary their latest release is, they’ll tell you a lot about what users have been clamoring for. If you see something in there that really piques your interest, bring it up in one of the online forums or at the next user group meeting. Chances are, you’re not the only one who’s chomping at the bit to understand how best to fit it into your workflow.

 

JH: What applications do you consider to be the minimum toolset for any photographer interested in incorporating CGI into their business?

 

WJ: In the world of CGI you’ll spend most of your time doing two things-creating CGI images and compositing them with something else (usually a photograph of some kind). While there are applications that can handle both of these tasks within a single interface, I’m very much of the opinion that you should use the right tool for the job, which almost always means using specialized software to deal separately with 3D data and 2D imagery.

 

For creating the CG images, you’ll want some core 3D package like Maya, 3ds Max, Lightwave 3D, C4D, Houdini or XSI in which you can model, texture, pose, light and render. Expanding from there, you’ll find packages that are really tuned to dealing with particular parts of that workflow — ZBrush for modeling, BodyPaint 3D for texturing, etc. However, most people will stave off delving into that next level until they’ve at least established a basic, working pipeline using the single, core application. As I mentioned above, Maya forms the core of my own 3D pipeline, but everyone has their own favorite and 3ds Max often ranks highly. Be sure to play with trial versions and compare features before you take the plunge — even a single license for most 3D software is not cheap.

 

For compositing those CG images together into final 2D images, as well as preparing textures and doing a variety of other tasks, Photoshop is really the only way to go for stills. For starters, it’s a package already familiar to most photographers. But beyond that, Photoshop really gives you immediate access to the tools needed for quick integration of still imagery. Specialized compositing applications such as Shake, Nuke and After Effects are really streamlined for moving footage, but incur quite a bit of overhead if you’re just dealing with a still frame. Most complex composites will require quite a bit of paint work by the time you’re done — even if it’s just to adjust exposures in certain areas or tweak the color — and Photoshop really shines in that area. Many people don’t know it, but Photoshop also now offers up some limited 3D functionality, enabling some tasks to be done completely within that singular package.

 

Typical pipelines will find people rendering imagery in their core 3D package and then creating composites in their 2D package. You can spend a lot of time jumping back and forth between the two. One tidbit on streamlining this workflow that I’ll share is to render everything in as many layers as possible. If you’re creating a CG environment that has both foreground and background elements, those should be output to separate images. Within each of those elements you’ll also want to break things down into individual layers for each light, for reflections, etc. Having everything split out like this means you can make a lot of adjustments within Photoshop without having to go back and re-render things again. By simply adjusting the color balance or levels of the layer containing the contribution of your key light, you can quickly adjust the intensity and color of that light, all interactively and without having to even go back to your 3D application.

 

JH: You wrote about working with a network of independent artists. How did you build this network?

 

WJ: The short answer is “years of being in the business.” But it’s really like building any other network. My network started out with people I met at various gatherings like Maya user group meetings, SIGGRAPH, etc. From there it expanded to people I met through those initial contacts. You try and keep in touch with everyone you think you might want to work with at some point — people who have specialized skills or those who you know can come in during a pinch and turn things around. I’ve been lucky working in visual effects because the industry tends to be pretty transient in terms of its workforce. People are brought in for a three-month project and then scatter to other studios. Typically we have several hundred project-hires working on a film during the peak of production. When the job is over, some stay on if there’s new work coming in, but many go off to the next project somewhere else, possibly not even in the same country. You keep tabs on people. And more often than not, you find yourself calling upon them for something, even if it’s years down the road.

 

For those looking to build a network of independent artists that they can lean on for work, my advice once again goes back to listservs and user groups. Ask questions, give answers, make friends. Network as much as you can — you never know when someone’s going to save your ass. Even if people aren’t in the same country as you, the Internet has made the world a small place. There’s nothing to say that you can’t lean on that amazing fluid simulation guy in London when you’re in Los Angeles trying to bang out something for a client.

 

Cheers,
Walt Jones
www.liquidimagephotography.com