So what’s it like being an artist working with a handheld like PSP? Nathan Phail-Liff, Lead 3D Artist at Ready at Dawn Studios, knows just what goes into making a successful and visually pleasing game within the limitations of the handheld world from his work on God of War: Chains of Olympus. And if you ask me, the graphics are so stunning that it looks comparable to anything you’d find on console.

Q: God of War: Chains of Olympus has been called the best game the PSP has to offer, based on its graphics quality in both cutscenes and gameplay. Could you tell us a bit about your role on the project?

I came off of Daxter as an environment artist, but at the start GOW we knew we had to drastically improve our tools pipeline and I had some ideas I wanted to try. So oddly enough I spent the first 8 months or so programming tools. I helped to revamp the entire maya-side of our tools pipeline and worked closely with the engine and tools programmers to get all the systems we would need in place.

With this out of the way I moved back to the art department in the role of Lead 3D Artist. With a technical as well as 3D illustration background, my main task was to work with the art team to establish and maintain a cohesive art style and get an efficient pipeline running smoothly across all disciplines, from modeling, to texturing, to lighting and FX. In addition to supporting the art team with direction and feedback, I picked up a good bit of my own art production work as well. I floated around as needed, doing a little environment art and in-game lighting, but for the last 4-5 months of production I focused mainly on FX (particles, sky domes, water, fog, blood, etc), as well as painting rough color keys for environmental lighting guides and helping out with cinematic background art.

Q: As an artist, what is it like working with existing IP? How does your team keep the style consistent with the original game?

A: Working with an existing IP is always both a blessing and a curse. On one hand you have an enormous amount of reference to start off with as a style guide, but on the other you are always torn between things you would have liked to do differently as an artist. So it was always a tight rope act, balancing between making sure the game was strongly identifiable to the fans as a “God of War” game and being faithful to the franchise, while still adding our own touch to things from time to time.

Another mixed blessing was that the caliber of the artistic quality established by the previous God of War games set the bar for us ridiculously high from the start. We were striving to equal the quality of the most technically impressive end-of-life PS2 games, all with much less power and far earlier in the life cycle of the PSP. It was a terrifying way to begin a project, but at the same time it helped to push ourselves even further than we though was possible.

With these caveats in mind we started off in pre-production by poring over all the assets the Santa Monica team was kind enough to provide us from the first 2 games. In addition to getting valuable insight to all sorts of tricks and techniques they used to make the GOW games look as amazing as they did, we tried to boil the God of War “look” down into a number of components, from modeling details, to lighting and texturing contrast, saturation, palette, etc. Coming from Daxter there were enormous style changes we had to go through, but we established a number of visual rules that were very clear from the start. One example was shifting towards a modeling style that made a much stronger use of faceted, hard polygonal edges to catch strong glints of light off of surface details and cracks.

Other stylistic guidelines proved less obvious though. For example, it was interesting at the start of the project trying to resolve some big visual differences between the lighting and texturing look from the GOW 1 with early content we saw from GOW 2. GOW 1 tended to be more monochromatic, with strong contrast between light and shadow, while GOW 2 was shaping up to have a lot more fill light in shadows and a much more polychromatic color scheme. In the end we came up with our own compromise between the two, settling on a narrower range in color palette and a little higher contrast than GOW 2, but using stronger of fill light and colored shadows than GOW 1.

Q: What would you say is unique about working with graphics for the PSP? Did you face any challenges or limitations?

A: A single crippled system bus, no hardware clipping, lower texture and poly resolution, and competing with particles / FX against the fill-rate monster that is the PS2 required a lot of creative problem solving on both the art and tech side. We had already pushed the hardware pretty hard with Daxter, but God of War proved to be an order of magnitude more demanding.

There are far too many challenges to list, but we constantly profiled our engine and one concept that was hammered home repeatedly was that the system bus and cpu are the PSP’s Achilles heel. Because of this we tried as much as possible to stay off the bus and keep data localized on the GPU / CPU, reducing RAM access as much as possible. Some PSP specific art solutions we came up often seem counter-intuitive, but proved extremely effective. For example, we found it was actually far more efficient to subdivide geometry more, making individual polys smaller in relation to the PSP screen. This allowed us to avoid ever having to clip in software and render too large polygons and helped to workload more localized on the GPU as much as possible. It can be rather surreal as an artist to hear from an engine programmer, “That’s too expensive, use MORE polys!”, but we were hardly going to complain!

Squeezing every drop of performance out of the PSP also required creative adjustments in artistic style to hide the limitations of the hardware. For example, to avoid hits from bus/ram access overhead, artists were responsible for making sure every single texture and mesh was broken up to fit exactly within the ideal hardware cache size. This meant that in order to have a 128×128 texture we could only use 16 colors! In response, we shifted to painting textures at far lower contrast and with much less color variation than the PS2 games, and using alpha texture overlays and vertex color to build up more interesting detail and tonal variations. Hopefully, with the clever application of these tricks, no one will every guess that virtually every texture in the game is only 16 colors, including characters.

Q: How many team members did you have, and how did you help them manage workload and flow?

A: By the peak of production we had about 15 artists, split between character, environment, animation, texturing, lighting, and cinematics.

Streamlining the art production pipeline is always a 2 pronged approach for us. First off, we constantly look at what we can do to improve our approaches towards art direction and feedback. We are extremely focused on quality and polish and the only way this mentality is maintainable in a production environment is to aggressively reduce the impact of content changes and cost of iteration on artists’ schedules. Early on we found too much time was being wasted in redoing content that wasn’t up to the standard we wanted to achieve, and we knew this was a problem we had to solve quickly.

One effective solution was we had artists shift to prototyping very small slices of content. For example, when an environment artist was starting on a new level, he or she would work on a very small section, perhaps as small as a room and a hallway, and quickly iterate over and over on it in tandem with the texture and lighting artists. This would yield a strong style guide for the rest of the level’s content that would have to be approved before any further work was done. While this small section might take a week or two, the lessons learned could be extrapolated by the artist to produce an enormous amount of high quality content, in a fraction of the time. This production approach might be scary for a producer to hear, but it saved us a huge amount of time on the back-end, as well helped to keep a very consistent, high level of quality across the entire game.

Our other big focus in work flow optimization was to constantly evaluate the pipeline from a standpoint of technology, identifying and prioritizing bottle-necks to see what improvements to the artist tools or game engine could best help to accelerate content production and iteration time for artists. We value a small, highly skilled team and have to constantly look into ways to reduce any busy-work in the pipeline to a bare minimum. We worked with the programming department to come up with tools solutions and prioritize the implementation schedule to address the largest art production bottlenecks as soon as possible.

Q: Are there any techniques or programs you would recommend for a smooth art pipeline?

A: Iteration is king. Low tech solutions with faster iteration times and more flexibility for changes beat more sophisticated, bell and whistle technology every time. In an industry as driven by technology as games, there are a ton of buzz-words being tossed around every day. Global illumination, radiosity, normal mapping, etc. are meaningless if you don’t have a strong understanding of the artistic goals behind them and the specific visual style you want to achieve.

In my opinion art style and gameplay should always drive your technology choices and never the other around. By really understanding the artistic goals behind the technology trends you can often achieve the same look far more cheaply. By using lower tech tricks we can focus programming resources instead on much more robust user interface and real-time editing tools, allowing us to iterate on art style to the point that it looks visually superior to more elaborate, but slow and unwieldy technology solutions.

Almost as important as iteration, but less commonly talked about, is the concept of instantiation. A good art pipeline should allow as few artists as necessary to seamlessly edit the largest amount of content possible. Any type of similar visual components, from lights, to shaders, to geometry, to even cameras should have robust support for templating content and attributes. With this concept, visual polish and performance optimizations can be adjusted on as few assets as necessary, allowing all content in the game to benefit from lessons learned late in development. Instantiation makes the task of last minute changes and polish far less costly and risky, without having to wade through a mountain of assets with grueling, but repetitive changes.

Q: What are you most proud of on Chains of Olympus?

A: Everyone on the team is enormously proud of the fact that with the aggressive schedule and small team that we had we were able to pull off a full-fledged God of War game on the PSP, with virtually no concessions. Frankly there is no way we could have pulled off this game without the amount of dedication and sacrifice that everyone on the team put into it and, by all rights, it really was the game we shouldn’t have been able to make!

In terms of my own personal small moment I’m most proud of, it would definitely be the Morpheus fog FX. We had this particle fluid system that was implemented on the tech side and was supposed to be an integral story component. The landscape of the over world was to be enshrouded by this ethereal fog and the player actually had to interact with it, repelling it away with flame sources. The tech worked great on paper, but we were so swamped with other things that we didn’t get around to looking at it until far too late into production. The fill rate on PSP is pretty abysmal, particularly compared to PS2, and we realized out pretty late we had no idea how to implement dynamic particle FX that would look good walking through and engulfing the player, while still run in frame rate.

I ended up giving myself an ultimatum that if I couldn’t solve it in 2 days it would have to be cut from the game. Much to my relief, I was able to pull it off with all sorts of smoke and mirrors and I’m pretty happy that the end result not only looked pretty cool, but, I hope, helps to further immerse the player and sell the narrative of the story.

This entry was posted on Wednesday, April 2nd, 2008 at 4:23 pm and is filed under Art Interviews, PEOPLE. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.