At the time of its release in November 2007, Mass Effect took the Xbox world by storm. It quickly rose to become one of the best rated RPG titles in history. IGN gave the game the #1 slot it in its Top 25 list, while on the GameRankings aggregate ratings site the game came in only behind 2006's Oblivion. Mass Effect quickly became known for an immersive, engaging story, fun game play, and of course beautiful graphics. It wasn't long before the gaming community was champing at the bit for a sequel.
That expectation became realized just a couple months ago with Mass Effect 2. Released on January 26, 2010, ME2 was not just a financial success but a critical one as well. Where IGN gave the original title a score of 9.4 (out of 10), the sequel hit 9.6. As for GameRankings, Mass Effect 2 achieved 95.70% which impressively outdoes the original game's 91.07%. In fact, it has unseated Oblivion to become the best reviewed RPG of all time, on any platform.
Even though the art team is already heavily immersed in work for downloadable content to enhance this great game even farther, we were able to sneak in some time with them to sit down and talk about the character art of the Mass Effect universe and especially ME2. It's a pretty hefty interview, so settle into a comfortable chair, and enjoy!
What was the main artistic inspiration for the Mass Effect franchise, which you've described as being "80's sci-fi"?
Derek Watts - Art Director:
For Mass Effect we definitely had a few artists that we looked at for inspiration: Syd Mead, John Berkey and John Harris were just a few. We wanted this universe to have a more sophisticated look. The design had to be clean with strong geometric shapes. The structures and clothing needed a purpose. We didn’t want it to feel clunky, dirty and heavy. We wanted this future to feel like someplace that still had hope.
I’m not sure I can really describe 80’s sci-fi. I’m very content with any kind of sci-fi as long as it looks good. I can put up with a poor script as long as the visuals are cool. I’m not a huge fan of low budget sci-fi; it's very painful for me to watch.
What were your main goals for the sequel?
Jaemus Wurzbach – Lead Character Artist:
We had a pretty good start with ME1. We had a new IP and we were using a new engine for the first time: Unreal3. There were a lot of growing pains, but by the end of ME1 we had learned a lot. One of the main things we learned in the character art team was how to make assets more efficiently, so that became a big focus. This encompasses how we create textures and meshes, as well as how we build our shaders. This helped to make our assets run a bit better in game, which helped out performance a lot.
Another key focus was in creating much higher quality art that was given an iconic look. In ME1, armors you would find in the game were all based off of a few meshes. We had a ton of texture variations, but everything pretty much looked the same. NPC's that you would fight against always looked like they were loners, wearing rag tag equipment. The characters really didn’t have a cohesive look to them. The same could be said for the members of your party. The player and the party members could be wearing the same outfits and this made it difficult, from a visual standpoint, to make them stand out from one another. At the same time, all characters from a specific race/species had to be able to share the same clothing and armor pieces. This really hampered us from making unique assets.
In ME2 we sought to change this. All of the party members were given an iconic look that helped define them. This allowed us to create very highly detailed unique work that carries a lot of personality with it. ME2 was loosely based off of the idea of “The Dirty Dozen” so the party members in ME2 were the primary focus for character art. We also made the NPC's you fight against in the game part of factions; mercenary teams and gangs had similar-colored suits and logos, which helped tie them into the world.
Was the texture popping effect that bugged a lot of people in the first game something that was solved through programming, or did it require changes in the graphics themselves?
Jaemus Wurzbach – Lead Character Artist:
The texture pop-in problem was a huge issue in ME1. There was no one easy solution to fix the issue; it was a more of a team effort. For character art in particular in involved us managing our texture resources and material shaders a bit better. The player and party members were given slightly larger texture sizes, whereas the NPC’s you fight or talk with generally have lower res textures. We also tried to limit the amount of textures a character uses to no more than 3, although there are some exceptions. In Unreal we try to share the same materials across a large amount of individual characters and keep them as streamlined as possible while still having them look great. This, in addition to the way that the Level art team built streaming load sections into the levels and the efforts of the programming team helped to significantly minimize the texture pop-in.
Looking back at the first game, what was the one thing that you really thought, "we SO need to change that!"?
Derek Watts – Art Director:
Beyond the performance and pipeline issues, we never really looked at ME 1 and said “my God, that needs to be fixed”. It’s not until you are further into production on the second game that you look back and see all the issues that you should be fixing. Lighting was one of the big jumps we made in ME 2. If you look back at the first game you‘ll see how washed out everything looked compared to ME 2. We needed to bring back the contrast and saturation in the game. That was a key feature we liked in many of the early concepts we had referenced.
Jaemus Wurzbach – Lead Character Artist:
Yeah there were a lot of things that needed an overhaul. Like I said previously, ME1 gave us a great base to start from, but it needed a lot of refinement. The lighting system needed some love. This was one of the biggest problems we had with ME1; the lighting setup brought in a large amount of ambient light so shadow values would appear very faint. In addition to this we used a post-process effect that added a lot of bloom to the camera. The drawback to this meant that even the smallest amount of specular highlight would blowout to full white. This effectively nullified any effect we would have received from our normal maps. The lighting we did have was also not dynamic and would not interact with your character like you would expect it to. In ME2 the programming and level art team completely revamped the lighting system. The new system was dynamic and would affect your character in a way you would expect it to: when you walk in front of a light, you get brighter. The old post-process effect was also toned down, so the artwork is able to read much better than before.
In ME1 we created a large amount of our assets to be used within a tintable system. This was used on our armors, guns and weapons to help give our designers a large amount of variety to choose from at a cheap cost. This helped them when populating some of the large city areas in the game. The downside to this system is that it makes the artwork look rather bland. Very few things looked unique or stood out as amazing. In ME2 we still use this system for our crowd characters, however we use much more unique assets as well. The blend of these two make the game look much better than before.
How did you go about turning the super high resolution detailing that you were getting out of ZBrush into something that wouldn't make the Xbox puke?
Jaemus Wurzbach – Lead Character Artist:
Well we made the artwork to serve different purposes. We knew when we were making the work that we would not be able to have it look as good on the Xbox as it did in ZBrush. We make our promotional assets using 2k texture maps and then down-res those for in game which used a combination of a 1024 and two 512’s. Often times that meant rethinking some of the designs; a very ornate clasp might look great in ZBrush and on the promotional artwork but look bad in the in-game version. In cases like that we would go all the way back to our detail meshes and simplify things to keep them consistent. We also would have to go into our lower res textures and clean things up a bit, particularly with our specular maps.
Speaking of platforms, do the Xbox and PC versions use the same assets across the board, or does the PC version have extra elements such as higher resolution textures? What were the main challenges developing for both platforms?
Jaemus Wurzbach – Lead Character Artist:
The PC and Xbox both use the exact same assets, however PC's can potentially utilize higher resolution textures that the Xbox can. When we put our textures in game we import them at the 1024 level, however we need to set them into LoD (Level of Detail) groups which scale their sizes down to be used in game. For a character we basically use a color diffuse map at a 1024 LoD and a Normal and Specular map at a 512 LoD -- we do this for performance issues. If you had a nice PC though you could load the textures in at full res. There weren’t too many issues developing for the two different platforms. They are similar enough that they don’t make too many headaches. Making assets on the Xbox does have a few very minor extra steps, but nothing notable that we ran into.
How many texture maps did a model like Grunt or Mordin end up using in-game? That includes any color, bump, normal maps, etc.
Jaemus Wurzbach – Lead Character Artist:
We break our characters into different chunks. The vast majority -- especially the aliens -- are made up of two chunks: a head and a body. We do this because our game has such a huge focus on conversations. We need to get as much resolution as possible in the faces because our designers love to do extreme close-ups. If we make creatures that are never intended to be in conversations, like the Collectors, then we keep them to one chunk. We try to keep each chunk to around three textures. We have a color diffuse, a normal map and a specular stack. We break the specular stack into 4 sections: the R,G,B and Alpha channels and then slot 4 grayscale images into that. Typically these maps might include a specular map, a specular power map, an emissive map, and various masks we might need to control special parameters we create in our shaders. So a character like Grunt or Mordin would run about 6 textures each; 3 for the head and 3 for the body.
How did the look of Mordin evolve from the original Salarian design? What role did ZBrush play in this?
Rion Swanson – Character Artist:
We really tried to push the character of Mordin beyond what the original Salarian heads offered in terms of facial features. His face shape has been adjusted somewhat, giving him a more squared-off jaw line and bulging cheekbones. His skin is quite wrinkled and weathered with lots of small age spots and flecks of different flesh and brown tones. Overall, the look of Mordin’s face and head has become more detailed and reflective of a character that has been through a lot in his life. Although he’s a brilliant scientist, the large scars over his face immediately suggest he’s had a few run-ins with direct physical threats or combat which is very true of his character. ZBrush was a great tool for achieving the aged, weathered appearance we wanted to capture. It allowed for easy and effective sculpting of both his face and head as well as cloth areas on his body. We also wanted more visible expression to show through, seeing as how Mordin is prone to sudden outbursts of gleeful expression. We put additional detail into the eye area; deep reds and oranges and more prominent inner eyelids help give a more expressive and unique look and feel.
Garrus sees some major visual changes in this game. How did you go about creating that? What role did ZBrush play in this?
Jaemus Wurzbach – Lead Character Artist:
Poor Garrus! He takes a missile blast to the face. I had to create two versions of him; one before and after his accident. In ME1 we had a great sculpt for Garrus made in ZBrush so it was quite easy to pop that file open and add in some fresh scars to his head. I also dinged up his suit as the damaged version needed a hole blasted through it. It was a pretty easy process to take the new undamaged suit we made, make a new base mesh with the hole and pop that into ZBrush to make the metal look warped and shredded.
How did the look of Grunt evolve from the original Krogan design? He does represent the pinnacle of the Krogan race, after all!
Jaemus Wurzbach – Lead Character Artist:
Grunt is a super Krogan but he is also very much like a child. When you find him in the game he is fresh out of a laboratory test tube. We wanted him to look young, like a baby. The other Krogan you see have skin similar to what you would find on a tortoise. Grunt, however has much finer skin -- it looks softer and is scaly like a lizard. He hasn’t matured enough to get really thick skin yet. The same can be said for the horns on his head. Other Krogan have a solid crown with large spikes, whereas Grunt has many smaller segments that have not formed into a solid plate yet. This is similar to how a newborn baby’s skull is soft so the bones can move around some. Originally Grunt was supposed to look super clean, his armor was flawless with no dents or dings. As time went on we decided to dirty him up a little. His armor is still quite clean, but it looks more lived in and carries a few scratches.
The Mass Effect universe features something like 20 races so far. Where do you come up with all these ideas?
Derek Watts – Art Director:
It all starts with writing. Once they give us a brief description we can then move into the concept phase. We spend a great deal of time creating a visual "language" for each race. We want to be able to give them a hook that sticks with people long after they play the game. We usually start in phases with the first phase being whatever the concept artist thinks would look cool. This is the phase where they have the most freedom to push their own ideas. Ben Huen and Matt Rhodes did an awesome job giving us tremendous variety in a short time period. Then we start to narrow it down in the later phases and begin looking out for trouble spots: will it work correctly for combat, animation and digital acting?
You rarely nail it in the first phase. We usually go 3 or 4 phases before we get the final version.
Is it challenging to create differentiation for each character within any given race so that they all look like distinctive individuals rather than cookie-cutter creations?
Jaemus Wurzbach – Lead Character Artist:
We use a morph system for all the heads we have in game. Currently we have about 8 unique base morphs per gender. These are full head shapes. We then blend those in numerous other, smaller morphs that adjust the facial features. Usually when we go about making the heads for the game the level designers will take a first pass on generating the faces, then the character art team takes a stab to polish them up. We like to have quite a few artists work on these. It can be a fun process and each artist brings a different style to the faces, so not every area looks the same. It is very easy to create heads that tend to look very similar, though, so getting unique heads that look correct can be a bit of work. Usually the best bet -- like pretty much anything else in character art -- is to have some good references handy.
Let's talk about the Drell for a bit, since they're the new Citadel race that has been added to the game and one of the main characters is Drell. How did the design for Thane evolve?
Derek Watts – Art Director:
Thane was a great example of collaborative concepting. Both Ben and Matt worked on this race and they definitely felt some frustration in getting him completed. The part that threw everyone was the fact that he was to be a romance option for female Shepard characters; we just had a hard time getting that out of our heads. I think we over thought what female game players wanted. We explored several different avenues for the design but we basically settled on a head that Ben had done. We based him more on reptiles and aquatic reference. I’ll tell you that at this point the female staff members had heard about this and became very worried that their love interest would end up being a dude with a fish head. They ended up really liking the final character and we also had great feedback on him when he first appeared in screen shots.
What were your goals when you sat down to create Thane?
Kolby Jukes – Character Artist:
The original designs for Thane were a lot more humanoid. He was supposed to be incredibly handsome; most beautiful man in the universe sort of thing. When we made the switch to a more amphibian/fish-like design, the challenge was to still make him look attractive and sympathetic. So when I started work on the character, my primary goal was just to make a believable, badass alien hitman, that the male/female audience would genuinely want to romance.
Were there any ways in which he really pushed the limits of what's possible at this point?
Kolby Jukes – Character Artist:
Technically, I don’t think Thane pushes the envelope in any significant way. From a player perspective though, I feel like excellent writing, voice work and animation really brought Thane to life – in my opinion he’s a very striking & memorable character.
What were some of the techniques that you used for sculpting the character in ZBrush?
Kolby Jukes – Character Artist:
Nothing too exciting I’m afraid, it was a pretty standard workflow: I used the Move brush to find my volumes and proportions, and then the ClayTubes brush to mass out the facial features. Then I used then a combination of Clay, Standard, Inflate and Pinch to refine. The fine detailing on his head was done in 2D using Photoshop and Crazybump.
You rarely nail it in the first phase. We usually go 3 or 4 phases before we get the final version.
About how much time did it take to create Thane? How much of that was spent in ZBrush?
Kolby Jukes – Character Artist:
All in all, with some revisions, Thane was probably about 1.5 months of work. His head was sculpted mostly in ZBrush; I spent approximately 3 days on that in ZBrush.
The Collectors are downright creepy, and come in several different varieties. They're also all insanely detailed. Could you talk about their creation?
Leroy Chen – Character Artist:
When I was making the Trooper I was going for the insect look on the overall treatment. A lot of references were gathered to help me getting the feel in the detail mesh and also in the final shader.
Kolby Jukes – Character Artist:
The first 3 subdivisions of the Collector General model were sculpted in MudBox v1.07. The model was already somewhat heavy at the 3rd subdivision (approximately 4 million polygons distributed between a number of separate meshes) and I started to run into insufficient memory errors when I tried to subdivide some of the heavier meshes like the torso & head. In order to achieve the necessary detail I imported and reconstructed the model into ZBrush 3.1, which was capable of comfortably pushing about 10 million polys on my workstation. This transition allowed me to add the necessary tertiary forms (flesh and bone detailing) and high frequency detail to the model cleanly and efficiently.
When sculpting the anatomy I drew a lot of inspiration from the works of Japanese model-makers Yasushi Nirasawa and Takayuki Takeya; two of my favourite sculptors.
Where did you find yourselves really pushing the envelope with the Collectors?
Leroy Chen – Character Artist:
As for the Collector Trooper, since it has pretty fine concentration of details on its body the distribution of the loops is pretty essential. First I targeted areas to make cuts on the body so it minimized seams and then brought them in as SubTools, thus maximizing the polycount usage. I also made sure all the pieces would contain mostly quads at this point. With the help of custom alphas and brushes like ClayTubes, the final result was achieved.
How did the design for the Reaper Baby come about? Just how much can you say without giving away secrets that are waiting for 3 to be revealed?
Derek Watts – Art Director:
The Reaper Baby was another creature that became my nemesis on this game. I’m not sure if it’s possible to have every creature be a nemesis, but I came close. Eventually you just want them done and you lose track of what you are trying to make. The Reaper Baby was the end product of the humans getting processed and mixed with Reaper technology to make one Super Reaper. We had no idea what that thing should look like. Casey Hudson (Project Director) had an idea of what he wanted; it was just hard for us to get his idea down on paper. We had a bunch of concepts and even after choosing the final one we still needed to make some last minute changes. It eventually looked pretty cool and the final model was awesome.
Leroy Chen – Character Artist:
The Reaper Baby is basically a proto-form human Reaper. So we referenced Sovereign from the first Mass effect to tie it into the same Reaper family. Since it's still in its semi-finished stage, we exposed a lot more of the interior components and structures to reinforce that idea. From start to finish, the design went through a few iteration changes to reach the level of creepiness we wanted for the final fight.
The Baby features a huge amount of hard surface modeling and really sharp non-organic details. How was ZBrush used for this?
Leroy Chen – Character Artist:
The usual way of doing hard surfaces in applications like 3ds Max can get you a pretty solid result, however I find this method lacked the robustness that allows quick and fast changes. Thus I chose ZBrush to do a lot of the Reaper Baby’s fine details. ZBrush allows quick reaction time for me to respond to the art director's inputs. But before the introduction of ZBrush 3.5, to make hard surfaces in ZBrush was a bit challenging. So I prepared my mesh in Max first and instead of adding extra loops in areas like sharp edges to keep the form under mesh smooth, I picked smoothing group as the method to separate the surfaces. Through this method I made sure the mesh had the necessary clean shapes and the proper loop distribution for detailing in ZBrush. Once in ZBrush, I used brushes like Ram, Slash 2, and Pinch to do the panel lines. In addition, masks and projection were also heavily applied in the process to achieve the desired effects. However, with brushes like Planar Cut and functions like redistribution of edges in 3.5, the workflow can now be even more efficient.
The Illusive Man bears a pretty strong physical resemblance to Martin Sheen, who voices him. Was this intentional?
Ken Finlayson – Character Artist:
From my perspective it wasn’t a factor at all. The asset was done, and it was probably November or December when I found out Martin Sheen was going to be the voice. A few different actors were thrown around when the Illusive Man was just a drawing in the conceptual stages. Eventually a model (not Martin Sheen) was chosen and then scanned as a base for the head. Generally the hero human characters start from a scan, but we do a lot of alterations to that data to find the character and get the look we want. You would be hard pressed to find a guy that looked like Zaeed on the street. I personally deleted and rebuilt the scan back from about the 100K poly level, putting in weight, form and accentuating details that just aren’t in the raw data.
Sometimes the scan has a cooked-in expression we need to take out. I remember Zaeed looked like he was sucking on a lemon. I assume this was from stage direction to play up the wrinkles around his mouth. Then of course there were elements of the design like a scar we needed to add in that just weren’t present at all.
The Illusive man was a pretty interesting asset to build because he is essentially a real-time cinematics asset. He had a few design features and requirements that meant he broke the mold a bit in regards to a normal human in ME. One of the first things I did was talk to Cinematics about the types of shots he would need to hold up to while not being a ridiculously expensive asset. He needed hands that would hold up to close ups. I also made a special set of teeth and tongue for him. He has 3 UV sets for just the head to get the resolution I wanted on the chest, hair and face. The Illusive man also had an open shirt which extended far below where our head meshes usually end and well below the scan data we got. I ended up having to add geometry to the scan by sewing in a specific amount of geometry at the 100K poly level. That way I could keep what I liked in the scan and I could then rebuild hierarchy on the mesh and restore all my SUBD levels. This allowed me to get back to the in-game head mesh again and have it propagate to my changes. After that it was a straightforward sculpt.
How did the idea of his cybernetic eyes come about?
Derek Watts – Art Director:
The Illusive Man was based on a concept Casey had found. This became a good starting point for us to work from. He wanted a 40-50 year old man with almost a department store catalogue model look; a good looking man that had the whole world in his hands. One of those Fortune 500 guys that just seems to have everything figured out. After we had that, he still needed a hook that people would remember. The original reference had the man wearing glasses. We realized he needed something in his eyes that would resonate with the player. We decided to add lights into his irises for interest. A simple addition like that can go a long way to increase people’s interest in the character. It leaves a lot of room for interpretation. It gets the forums talking.
Ken Finlayson – Character Artist:
We try to have a reason behind most things in the Mass Effect universe so that it feels real and authentic. There is a back story to the Illusive man’s glowing eyes, but it has not been revealed yet.
One of the big changes with 2 is the character armor. Rather than having dozens of different armor sets that Shepard can equip, he now (with the exception of the special armor sets like the Collector armor) has a base armor that can be heavily modified and is added to over time. What challenges did that present? Was the decision to go to this new approach primarily motivated by gameplay or artistic reasons?
Jaemus Wurzbach – Lead Character Artist:
The idea to change the way we handle the character armor was art driven. This was one of the first things I proposed when we started ME2 very early in production. The idea behind this was to give the player far more control over how their character looked. In addition to customizing your face, you could now have some control over your body as well by being able to change what pieces you were wearing and color them whatever you wanted. This was also an attempt to add in far more variety to our game without doing any more work. If you take ME1 for example, we had 3 armor types: Light Medium and Heavy. With about 14 to 17 variations each, this added up to a huge amount of time spent making texture variations while the armors themselves generally looked the same. In ME2 we made 3 full suits, so we made the same amount we made for ME1. But by breaking these into separate parts, that allows the player to craft 16 unique-looking suits. The ME2 suits only have 1 texture set as well, so we didn’t have to spend a large amount of time on texture variations. In the end what we ended up with was better looking assets that took far less time to make, and that give the players much more control over their characters.
The system does have a few quirks that can make it difficult to work with. The suits work with our tinting system, so in our shader the colors are all sorted out and then our diffuse map in multiplied into it. This tends to wash out the texture slightly and can be difficult to compensate for. We also had to make sure that all the suits were modeled and textured in similar styles. Since you could wear the arms and legs of different sets we had to make sure that they all synched up together and looked like a cohesive set of armor no matter what selection you made.
Speaking of the special armor, how did the Collector armor evolve from the design of the Collectors themselves?
Kolby Jukes – Character Artist:
The Collector armor is essentially an amalgamation of the iconic commander Shepard armor and the Collector Trooper. ZBrush 3.5 R3’s ability to import geometry with different topology/point-orders was a real lifesaver in the creation of the Collector armor. I was able to ‘slice’ surfaces off of the Collector Trooper, then thicken and wrap these plates to the iconic armor rapidly – as opposed to having to create new plating geometry and scrub the detail off the original model using ZProject.
Did you have to give this armor extra attention because the player would generally be seeing it in greater detail than the enemies? Or were even the Collectors as detailed as the player armor?
Jaemus Wurzbach – Lead Character Artist:
Both artists (Leroy Chen and Kolby Jukes) involved in the creation on these assets did an amazing job, and the workmanship looks great. Texture resolution and poly counts in game where very similar, although the player does see the Shepard armor much more often and up close we still use the Collector Troopers in numerous cinematics so they need to look just as good.
How many polygons did the Collector armor come to in ZBrush? Was it built with SubTools and if so, how many were used?
Kolby Jukes – Character Artist:
The Collector armor looks a lot heavier than it is. It’s approximately 4.5 million polys and is composed of 16 SubTools.
Normandy II is a big step up from the original. How did you go about evolving the design while also staying so very true to the spirit of the first?
Derek Watts – Art Director:
I can still remember the day that Casey told me he wanted to blow up the Normandy: instant rage. I screamed, "Are you crazy? The first one took months to design and build!” I was overreacting on this subject and went on many rants about why we blow up everything in video games. Once the reality set in that they weren’t going to change their minds, I convinced Casey to make only minor changes to the original design. We defined the cockpit more and made it smaller to create the illusion the ship was larger. We streamlined the shape and gave it more of a delta wing appearance. We added another level of windows on the side to bulk up the size. We referenced the B1 bomber and the Concorde which still look aggressive for such large planes. In the end the Normandy 2 came out looking pretty damned awesome.
Rion Swanson – Character Artist:
The design was very well planned out in the concept stage, but after the base model was built there were quite a few tweaks and adjustments made to further refine it. Slight adjustments of subtle angles and curves were the major factor that eventually led to the final approved model. It did take some time to get this right, as it sometimes does when translating objects from 2D into 3D. We basically considered the Normandy II a character in itself and had to make sure we spent the necessary time refining both the exterior and the interior spaces. The original Normandy was referenced heavily throughout the process in order to preserve and carry forward elements of its unique character.
What is your favorite ZBrush feature and why?
Ken Finlayson – Character Artist:
I’ve been using ZBrush since 1.55b, so my favorite feature of the software has changed a lot as the toolset has grown and grown. When I first started with it I loved not thinking about the polygons and line flow. I only cared about the form in ZBrush and I could make something look good in ¾ view, which I had always found challenging back then coming from the days of modeling in a quad view. You never worked in perspective, you just looked at the result there.
In a general way I really love that there are often multiple ways in ZBrush to get to wherever you’re trying to go. I think that says a lot. The ability to use an arbitrary mesh with ZMapper was a big deal at the time, and I still find new uses for the rebuild subdivisions function in ZBrush. These days my favorite feature is ZSpheres 2. I’ve dreamt of the ability to sketch in 3D for years, even to the point where I had mocked up an interface for how you would represent your brush position in 3D with a monitor and mouse/Wacom that only has x and y inputs. I always wanted the ability to design in 3D as fast as you could in 2D, or at least close the gap some.
Kolby Jukes – Character Artist:
The biggest step forward for me has been the ability to import models with different topology, point-orders, etc. and have ZBrush propagate those changes through the existing subdivision levels. Being able to add an extra pair of arms to a character composed of a single, contiguous mesh or reroute the edge loops on a face quickly and efficiently is awesome.
Rion Swanson – Character Artist:
I do really enjoy the SubTools functionality, as well as the possibilities that Layers offers. ZMapper has in the past been a very helpful tool also, and I look forward to trying out UV Master within the latest version.
Normandy II is a big step up from the original. How did you go about evolving the design while also staying so very true to the spirit of the first?
Derek Watts – Art Director:
I can still remember the day that Casey told me he wanted to blow up the Normandy: instant rage. I screamed, "Are you crazy? The first one took months to design and build!” I was overreacting on this subject and went on many rants about why we blow up everything in video games. Once the reality set in that they weren’t going to change their minds, I convinced Casey to make only minor changes to the original design. We defined the cockpit more and made it smaller to create the illusion the ship was larger. We streamlined the shape and gave it more of a delta wing appearance. We added another level of windows on the side to bulk up the size. We referenced the B1 bomber and the Concorde which still look aggressive for such large planes. In the end the Normandy 2 came out looking pretty damned awesome.
Rion Swanson – Character Artist:
The design was very well planned out in the concept stage, but after the base model was built there were quite a few tweaks and adjustments made to further refine it. Slight adjustments of subtle angles and curves were the major factor that eventually led to the final approved model. It did take some time to get this right, as it sometimes does when translating objects from 2D into 3D. We basically considered the Normandy II a character in itself and had to make sure we spent the necessary time refining both the exterior and the interior spaces. The original Normandy was referenced heavily throughout the process in order to preserve and carry forward elements of its unique character.
What is your favorite ZBrush feature and why?
Ken Finlayson – Character Artist:
I’ve been using ZBrush since 1.55b, so my favorite feature of the software has changed a lot as the toolset has grown and grown. When I first started with it I loved not thinking about the polygons and line flow. I only cared about the form in ZBrush and I could make something look good in ¾ view, which I had always found challenging back then coming from the days of modeling in a quad view. You never worked in perspective, you just looked at the result there.
In a general way I really love that there are often multiple ways in ZBrush to get to wherever you’re trying to go. I think that says a lot. The ability to use an arbitrary mesh with ZMapper was a big deal at the time, and I still find new uses for the rebuild subdivisions function in ZBrush. These days my favorite feature is ZSpheres 2. I’ve dreamt of the ability to sketch in 3D for years, even to the point where I had mocked up an interface for how you would represent your brush position in 3D with a monitor and mouse/Wacom that only has x and y inputs. I always wanted the ability to design in 3D as fast as you could in 2D, or at least close the gap some.
Kolby Jukes – Character Artist:
The biggest step forward for me has been the ability to import models with different topology, point-orders, etc. and have ZBrush propagate those changes through the existing subdivision levels. Being able to add an extra pair of arms to a character composed of a single, contiguous mesh or reroute the edge loops on a face quickly and efficiently is awesome.
Rion Swanson – Character Artist:
I do really enjoy the SubTools functionality, as well as the possibilities that Layers offers. ZMapper has in the past been a very helpful tool also, and I look forward to trying out UV Master within the latest version.
What is your favorite element of ME2? What is the one thing that we should watch for as we play the game?
Ken Finlayson – Character Artist:
ME2 has really beautiful level art that always blows me away. In our bi-weekly art meetings there are lots of great details like this little gem on Krogan Hub.
Kolby Jukes – Character Artist:
The party members are by far my favorite part of the game. It was a lot of fun watching them come to life, from the initial concepts all the way to hearing them speak for the first time. Garrus, Mordin and Grunt are some of my favorite game characters of all time.
Rion Swanson – Character Artist:
I love the idea of tracking down and recruiting all sorts of unique characters with different abilities and specialties… essentially your own band of mercs. The fact that each character really has their own flavor is great and watching the interactions between them with their sometimes clashing personalities really creates for a more immersive experience. I like the mix of new characters as well as characters from Mass Effect 1 that you run into once again.
Keep an eye out for Mordin’s little number!
(Interviewer's note: Mordin's perfectly-sung take on Gilbert and Sullivan was one of the funniest moments I've ever seen in an RPG.)
Jaemus Wurzbach – Lead Character Artist:
I find the experience of playing the game to be my favorite thing of ME2. It feels much more like a movie that you are directing than a game at times.
Now that ME2 is out, are you breathing a sigh of relief or are your noses already to the grindstone on 3?
Jaemus Wurzbach – Lead Character Artist:
While it’s definitely a sigh of relief to be done with ME2, we have our noses back to the grindstone. No rest for the wicked, I guess. We have a fairly good amount of post-release content to get started on and there are a few new things we need to get prototyped before we start up production on ME3.
What are your main goals for 3? Where do you feel you'll have to focus in order to top what you achieved with 2?
Jaemus Wurzbach – Lead Character Artist:
It’s hard to say right now. Until we get the general story bolted down it’s hard to say where to focus our attention. I’m sure the party will be a huge focus, and I’d like to see as big of a jump in quality from 2 to 3 as we made from 1 to 2.
Let's see -- that would put the GameReview score at 100.33%. That would be quite an achievement, indeed!
We at Pixologic would like to express our sincere gratitude to Derek, Jaemus, Rion, Kolby, Leroy and Ken for pulling back the curtain and revealing these insights into the Mass Effect universe. There's a lot of great information here for everyone from hobbyists to experts. Many thanks!
We also wish to thank BioWare for continuing to push the genre forward, and Electronic Arts for giving the consent to do this interview in the first place.
See you around the galaxy!
That expectation became realized just a couple months ago with Mass Effect 2. Released on January 26, 2010, ME2 was not just a financial success but a critical one as well. Where IGN gave the original title a score of 9.4 (out of 10), the sequel hit 9.6. As for GameRankings, Mass Effect 2 achieved 95.70% which impressively outdoes the original game's 91.07%. In fact, it has unseated Oblivion to become the best reviewed RPG of all time, on any platform.
Even though the art team is already heavily immersed in work for downloadable content to enhance this great game even farther, we were able to sneak in some time with them to sit down and talk about the character art of the Mass Effect universe and especially ME2. It's a pretty hefty interview, so settle into a comfortable chair, and enjoy!
What was the main artistic inspiration for the Mass Effect franchise, which you've described as being "80's sci-fi"?
Derek Watts - Art Director:
For Mass Effect we definitely had a few artists that we looked at for inspiration: Syd Mead, John Berkey and John Harris were just a few. We wanted this universe to have a more sophisticated look. The design had to be clean with strong geometric shapes. The structures and clothing needed a purpose. We didn’t want it to feel clunky, dirty and heavy. We wanted this future to feel like someplace that still had hope.
I’m not sure I can really describe 80’s sci-fi. I’m very content with any kind of sci-fi as long as it looks good. I can put up with a poor script as long as the visuals are cool. I’m not a huge fan of low budget sci-fi; it's very painful for me to watch.
What were your main goals for the sequel?
Jaemus Wurzbach – Lead Character Artist:
We had a pretty good start with ME1. We had a new IP and we were using a new engine for the first time: Unreal3. There were a lot of growing pains, but by the end of ME1 we had learned a lot. One of the main things we learned in the character art team was how to make assets more efficiently, so that became a big focus. This encompasses how we create textures and meshes, as well as how we build our shaders. This helped to make our assets run a bit better in game, which helped out performance a lot.
Another key focus was in creating much higher quality art that was given an iconic look. In ME1, armors you would find in the game were all based off of a few meshes. We had a ton of texture variations, but everything pretty much looked the same. NPC's that you would fight against always looked like they were loners, wearing rag tag equipment. The characters really didn’t have a cohesive look to them. The same could be said for the members of your party. The player and the party members could be wearing the same outfits and this made it difficult, from a visual standpoint, to make them stand out from one another. At the same time, all characters from a specific race/species had to be able to share the same clothing and armor pieces. This really hampered us from making unique assets.
In ME2 we sought to change this. All of the party members were given an iconic look that helped define them. This allowed us to create very highly detailed unique work that carries a lot of personality with it. ME2 was loosely based off of the idea of “The Dirty Dozen” so the party members in ME2 were the primary focus for character art. We also made the NPC's you fight against in the game part of factions; mercenary teams and gangs had similar-colored suits and logos, which helped tie them into the world.
Was the texture popping effect that bugged a lot of people in the first game something that was solved through programming, or did it require changes in the graphics themselves?
Jaemus Wurzbach – Lead Character Artist:
The texture pop-in problem was a huge issue in ME1. There was no one easy solution to fix the issue; it was a more of a team effort. For character art in particular in involved us managing our texture resources and material shaders a bit better. The player and party members were given slightly larger texture sizes, whereas the NPC’s you fight or talk with generally have lower res textures. We also tried to limit the amount of textures a character uses to no more than 3, although there are some exceptions. In Unreal we try to share the same materials across a large amount of individual characters and keep them as streamlined as possible while still having them look great. This, in addition to the way that the Level art team built streaming load sections into the levels and the efforts of the programming team helped to significantly minimize the texture pop-in.
Looking back at the first game, what was the one thing that you really thought, "we SO need to change that!"?
Derek Watts – Art Director:
Beyond the performance and pipeline issues, we never really looked at ME 1 and said “my God, that needs to be fixed”. It’s not until you are further into production on the second game that you look back and see all the issues that you should be fixing. Lighting was one of the big jumps we made in ME 2. If you look back at the first game you‘ll see how washed out everything looked compared to ME 2. We needed to bring back the contrast and saturation in the game. That was a key feature we liked in many of the early concepts we had referenced.
Jaemus Wurzbach – Lead Character Artist:
Yeah there were a lot of things that needed an overhaul. Like I said previously, ME1 gave us a great base to start from, but it needed a lot of refinement. The lighting system needed some love. This was one of the biggest problems we had with ME1; the lighting setup brought in a large amount of ambient light so shadow values would appear very faint. In addition to this we used a post-process effect that added a lot of bloom to the camera. The drawback to this meant that even the smallest amount of specular highlight would blowout to full white. This effectively nullified any effect we would have received from our normal maps. The lighting we did have was also not dynamic and would not interact with your character like you would expect it to. In ME2 the programming and level art team completely revamped the lighting system. The new system was dynamic and would affect your character in a way you would expect it to: when you walk in front of a light, you get brighter. The old post-process effect was also toned down, so the artwork is able to read much better than before.
In ME1 we created a large amount of our assets to be used within a tintable system. This was used on our armors, guns and weapons to help give our designers a large amount of variety to choose from at a cheap cost. This helped them when populating some of the large city areas in the game. The downside to this system is that it makes the artwork look rather bland. Very few things looked unique or stood out as amazing. In ME2 we still use this system for our crowd characters, however we use much more unique assets as well. The blend of these two make the game look much better than before.
How did you go about turning the super high resolution detailing that you were getting out of ZBrush into something that wouldn't make the Xbox puke?
Jaemus Wurzbach – Lead Character Artist:
Well we made the artwork to serve different purposes. We knew when we were making the work that we would not be able to have it look as good on the Xbox as it did in ZBrush. We make our promotional assets using 2k texture maps and then down-res those for in game which used a combination of a 1024 and two 512’s. Often times that meant rethinking some of the designs; a very ornate clasp might look great in ZBrush and on the promotional artwork but look bad in the in-game version. In cases like that we would go all the way back to our detail meshes and simplify things to keep them consistent. We also would have to go into our lower res textures and clean things up a bit, particularly with our specular maps.
Speaking of platforms, do the Xbox and PC versions use the same assets across the board, or does the PC version have extra elements such as higher resolution textures? What were the main challenges developing for both platforms?
Jaemus Wurzbach – Lead Character Artist:
The PC and Xbox both use the exact same assets, however PC's can potentially utilize higher resolution textures that the Xbox can. When we put our textures in game we import them at the 1024 level, however we need to set them into LoD (Level of Detail) groups which scale their sizes down to be used in game. For a character we basically use a color diffuse map at a 1024 LoD and a Normal and Specular map at a 512 LoD -- we do this for performance issues. If you had a nice PC though you could load the textures in at full res. There weren’t too many issues developing for the two different platforms. They are similar enough that they don’t make too many headaches. Making assets on the Xbox does have a few very minor extra steps, but nothing notable that we ran into.
How many texture maps did a model like Grunt or Mordin end up using in-game? That includes any color, bump, normal maps, etc.
Jaemus Wurzbach – Lead Character Artist:
We break our characters into different chunks. The vast majority -- especially the aliens -- are made up of two chunks: a head and a body. We do this because our game has such a huge focus on conversations. We need to get as much resolution as possible in the faces because our designers love to do extreme close-ups. If we make creatures that are never intended to be in conversations, like the Collectors, then we keep them to one chunk. We try to keep each chunk to around three textures. We have a color diffuse, a normal map and a specular stack. We break the specular stack into 4 sections: the R,G,B and Alpha channels and then slot 4 grayscale images into that. Typically these maps might include a specular map, a specular power map, an emissive map, and various masks we might need to control special parameters we create in our shaders. So a character like Grunt or Mordin would run about 6 textures each; 3 for the head and 3 for the body.
How did the look of Mordin evolve from the original Salarian design? What role did ZBrush play in this?
Rion Swanson – Character Artist:
We really tried to push the character of Mordin beyond what the original Salarian heads offered in terms of facial features. His face shape has been adjusted somewhat, giving him a more squared-off jaw line and bulging cheekbones. His skin is quite wrinkled and weathered with lots of small age spots and flecks of different flesh and brown tones. Overall, the look of Mordin’s face and head has become more detailed and reflective of a character that has been through a lot in his life. Although he’s a brilliant scientist, the large scars over his face immediately suggest he’s had a few run-ins with direct physical threats or combat which is very true of his character. ZBrush was a great tool for achieving the aged, weathered appearance we wanted to capture. It allowed for easy and effective sculpting of both his face and head as well as cloth areas on his body. We also wanted more visible expression to show through, seeing as how Mordin is prone to sudden outbursts of gleeful expression. We put additional detail into the eye area; deep reds and oranges and more prominent inner eyelids help give a more expressive and unique look and feel.
Garrus sees some major visual changes in this game. How did you go about creating that? What role did ZBrush play in this?
Jaemus Wurzbach – Lead Character Artist:
Poor Garrus! He takes a missile blast to the face. I had to create two versions of him; one before and after his accident. In ME1 we had a great sculpt for Garrus made in ZBrush so it was quite easy to pop that file open and add in some fresh scars to his head. I also dinged up his suit as the damaged version needed a hole blasted through it. It was a pretty easy process to take the new undamaged suit we made, make a new base mesh with the hole and pop that into ZBrush to make the metal look warped and shredded.
How did the look of Grunt evolve from the original Krogan design? He does represent the pinnacle of the Krogan race, after all!
Jaemus Wurzbach – Lead Character Artist:
Grunt is a super Krogan but he is also very much like a child. When you find him in the game he is fresh out of a laboratory test tube. We wanted him to look young, like a baby. The other Krogan you see have skin similar to what you would find on a tortoise. Grunt, however has much finer skin -- it looks softer and is scaly like a lizard. He hasn’t matured enough to get really thick skin yet. The same can be said for the horns on his head. Other Krogan have a solid crown with large spikes, whereas Grunt has many smaller segments that have not formed into a solid plate yet. This is similar to how a newborn baby’s skull is soft so the bones can move around some. Originally Grunt was supposed to look super clean, his armor was flawless with no dents or dings. As time went on we decided to dirty him up a little. His armor is still quite clean, but it looks more lived in and carries a few scratches.
The Mass Effect universe features something like 20 races so far. Where do you come up with all these ideas?
Derek Watts – Art Director:
It all starts with writing. Once they give us a brief description we can then move into the concept phase. We spend a great deal of time creating a visual "language" for each race. We want to be able to give them a hook that sticks with people long after they play the game. We usually start in phases with the first phase being whatever the concept artist thinks would look cool. This is the phase where they have the most freedom to push their own ideas. Ben Huen and Matt Rhodes did an awesome job giving us tremendous variety in a short time period. Then we start to narrow it down in the later phases and begin looking out for trouble spots: will it work correctly for combat, animation and digital acting?
You rarely nail it in the first phase. We usually go 3 or 4 phases before we get the final version.
Is it challenging to create differentiation for each character within any given race so that they all look like distinctive individuals rather than cookie-cutter creations?
Jaemus Wurzbach – Lead Character Artist:
We use a morph system for all the heads we have in game. Currently we have about 8 unique base morphs per gender. These are full head shapes. We then blend those in numerous other, smaller morphs that adjust the facial features. Usually when we go about making the heads for the game the level designers will take a first pass on generating the faces, then the character art team takes a stab to polish them up. We like to have quite a few artists work on these. It can be a fun process and each artist brings a different style to the faces, so not every area looks the same. It is very easy to create heads that tend to look very similar, though, so getting unique heads that look correct can be a bit of work. Usually the best bet -- like pretty much anything else in character art -- is to have some good references handy.
Let's talk about the Drell for a bit, since they're the new Citadel race that has been added to the game and one of the main characters is Drell. How did the design for Thane evolve?
Derek Watts – Art Director:
Thane was a great example of collaborative concepting. Both Ben and Matt worked on this race and they definitely felt some frustration in getting him completed. The part that threw everyone was the fact that he was to be a romance option for female Shepard characters; we just had a hard time getting that out of our heads. I think we over thought what female game players wanted. We explored several different avenues for the design but we basically settled on a head that Ben had done. We based him more on reptiles and aquatic reference. I’ll tell you that at this point the female staff members had heard about this and became very worried that their love interest would end up being a dude with a fish head. They ended up really liking the final character and we also had great feedback on him when he first appeared in screen shots.
What were your goals when you sat down to create Thane?
Kolby Jukes – Character Artist:
The original designs for Thane were a lot more humanoid. He was supposed to be incredibly handsome; most beautiful man in the universe sort of thing. When we made the switch to a more amphibian/fish-like design, the challenge was to still make him look attractive and sympathetic. So when I started work on the character, my primary goal was just to make a believable, badass alien hitman, that the male/female audience would genuinely want to romance.
Were there any ways in which he really pushed the limits of what's possible at this point?
Kolby Jukes – Character Artist:
Technically, I don’t think Thane pushes the envelope in any significant way. From a player perspective though, I feel like excellent writing, voice work and animation really brought Thane to life – in my opinion he’s a very striking & memorable character.
What were some of the techniques that you used for sculpting the character in ZBrush?
Kolby Jukes – Character Artist:
Nothing too exciting I’m afraid, it was a pretty standard workflow: I used the Move brush to find my volumes and proportions, and then the ClayTubes brush to mass out the facial features. Then I used then a combination of Clay, Standard, Inflate and Pinch to refine. The fine detailing on his head was done in 2D using Photoshop and Crazybump.
You rarely nail it in the first phase. We usually go 3 or 4 phases before we get the final version.
About how much time did it take to create Thane? How much of that was spent in ZBrush?
Kolby Jukes – Character Artist:
All in all, with some revisions, Thane was probably about 1.5 months of work. His head was sculpted mostly in ZBrush; I spent approximately 3 days on that in ZBrush.
The Collectors are downright creepy, and come in several different varieties. They're also all insanely detailed. Could you talk about their creation?
Leroy Chen – Character Artist:
When I was making the Trooper I was going for the insect look on the overall treatment. A lot of references were gathered to help me getting the feel in the detail mesh and also in the final shader.
Kolby Jukes – Character Artist:
The first 3 subdivisions of the Collector General model were sculpted in MudBox v1.07. The model was already somewhat heavy at the 3rd subdivision (approximately 4 million polygons distributed between a number of separate meshes) and I started to run into insufficient memory errors when I tried to subdivide some of the heavier meshes like the torso & head. In order to achieve the necessary detail I imported and reconstructed the model into ZBrush 3.1, which was capable of comfortably pushing about 10 million polys on my workstation. This transition allowed me to add the necessary tertiary forms (flesh and bone detailing) and high frequency detail to the model cleanly and efficiently.
When sculpting the anatomy I drew a lot of inspiration from the works of Japanese model-makers Yasushi Nirasawa and Takayuki Takeya; two of my favourite sculptors.
Where did you find yourselves really pushing the envelope with the Collectors?
Leroy Chen – Character Artist:
As for the Collector Trooper, since it has pretty fine concentration of details on its body the distribution of the loops is pretty essential. First I targeted areas to make cuts on the body so it minimized seams and then brought them in as SubTools, thus maximizing the polycount usage. I also made sure all the pieces would contain mostly quads at this point. With the help of custom alphas and brushes like ClayTubes, the final result was achieved.
How did the design for the Reaper Baby come about? Just how much can you say without giving away secrets that are waiting for 3 to be revealed?
Derek Watts – Art Director:
The Reaper Baby was another creature that became my nemesis on this game. I’m not sure if it’s possible to have every creature be a nemesis, but I came close. Eventually you just want them done and you lose track of what you are trying to make. The Reaper Baby was the end product of the humans getting processed and mixed with Reaper technology to make one Super Reaper. We had no idea what that thing should look like. Casey Hudson (Project Director) had an idea of what he wanted; it was just hard for us to get his idea down on paper. We had a bunch of concepts and even after choosing the final one we still needed to make some last minute changes. It eventually looked pretty cool and the final model was awesome.
Leroy Chen – Character Artist:
The Reaper Baby is basically a proto-form human Reaper. So we referenced Sovereign from the first Mass effect to tie it into the same Reaper family. Since it's still in its semi-finished stage, we exposed a lot more of the interior components and structures to reinforce that idea. From start to finish, the design went through a few iteration changes to reach the level of creepiness we wanted for the final fight.
The Baby features a huge amount of hard surface modeling and really sharp non-organic details. How was ZBrush used for this?
Leroy Chen – Character Artist:
The usual way of doing hard surfaces in applications like 3ds Max can get you a pretty solid result, however I find this method lacked the robustness that allows quick and fast changes. Thus I chose ZBrush to do a lot of the Reaper Baby’s fine details. ZBrush allows quick reaction time for me to respond to the art director's inputs. But before the introduction of ZBrush 3.5, to make hard surfaces in ZBrush was a bit challenging. So I prepared my mesh in Max first and instead of adding extra loops in areas like sharp edges to keep the form under mesh smooth, I picked smoothing group as the method to separate the surfaces. Through this method I made sure the mesh had the necessary clean shapes and the proper loop distribution for detailing in ZBrush. Once in ZBrush, I used brushes like Ram, Slash 2, and Pinch to do the panel lines. In addition, masks and projection were also heavily applied in the process to achieve the desired effects. However, with brushes like Planar Cut and functions like redistribution of edges in 3.5, the workflow can now be even more efficient.
The Illusive Man bears a pretty strong physical resemblance to Martin Sheen, who voices him. Was this intentional?
Ken Finlayson – Character Artist:
From my perspective it wasn’t a factor at all. The asset was done, and it was probably November or December when I found out Martin Sheen was going to be the voice. A few different actors were thrown around when the Illusive Man was just a drawing in the conceptual stages. Eventually a model (not Martin Sheen) was chosen and then scanned as a base for the head. Generally the hero human characters start from a scan, but we do a lot of alterations to that data to find the character and get the look we want. You would be hard pressed to find a guy that looked like Zaeed on the street. I personally deleted and rebuilt the scan back from about the 100K poly level, putting in weight, form and accentuating details that just aren’t in the raw data.
Sometimes the scan has a cooked-in expression we need to take out. I remember Zaeed looked like he was sucking on a lemon. I assume this was from stage direction to play up the wrinkles around his mouth. Then of course there were elements of the design like a scar we needed to add in that just weren’t present at all.
The Illusive man was a pretty interesting asset to build because he is essentially a real-time cinematics asset. He had a few design features and requirements that meant he broke the mold a bit in regards to a normal human in ME. One of the first things I did was talk to Cinematics about the types of shots he would need to hold up to while not being a ridiculously expensive asset. He needed hands that would hold up to close ups. I also made a special set of teeth and tongue for him. He has 3 UV sets for just the head to get the resolution I wanted on the chest, hair and face. The Illusive man also had an open shirt which extended far below where our head meshes usually end and well below the scan data we got. I ended up having to add geometry to the scan by sewing in a specific amount of geometry at the 100K poly level. That way I could keep what I liked in the scan and I could then rebuild hierarchy on the mesh and restore all my SUBD levels. This allowed me to get back to the in-game head mesh again and have it propagate to my changes. After that it was a straightforward sculpt.
How did the idea of his cybernetic eyes come about?
Derek Watts – Art Director:
The Illusive Man was based on a concept Casey had found. This became a good starting point for us to work from. He wanted a 40-50 year old man with almost a department store catalogue model look; a good looking man that had the whole world in his hands. One of those Fortune 500 guys that just seems to have everything figured out. After we had that, he still needed a hook that people would remember. The original reference had the man wearing glasses. We realized he needed something in his eyes that would resonate with the player. We decided to add lights into his irises for interest. A simple addition like that can go a long way to increase people’s interest in the character. It leaves a lot of room for interpretation. It gets the forums talking.
Ken Finlayson – Character Artist:
We try to have a reason behind most things in the Mass Effect universe so that it feels real and authentic. There is a back story to the Illusive man’s glowing eyes, but it has not been revealed yet.
One of the big changes with 2 is the character armor. Rather than having dozens of different armor sets that Shepard can equip, he now (with the exception of the special armor sets like the Collector armor) has a base armor that can be heavily modified and is added to over time. What challenges did that present? Was the decision to go to this new approach primarily motivated by gameplay or artistic reasons?
Jaemus Wurzbach – Lead Character Artist:
The idea to change the way we handle the character armor was art driven. This was one of the first things I proposed when we started ME2 very early in production. The idea behind this was to give the player far more control over how their character looked. In addition to customizing your face, you could now have some control over your body as well by being able to change what pieces you were wearing and color them whatever you wanted. This was also an attempt to add in far more variety to our game without doing any more work. If you take ME1 for example, we had 3 armor types: Light Medium and Heavy. With about 14 to 17 variations each, this added up to a huge amount of time spent making texture variations while the armors themselves generally looked the same. In ME2 we made 3 full suits, so we made the same amount we made for ME1. But by breaking these into separate parts, that allows the player to craft 16 unique-looking suits. The ME2 suits only have 1 texture set as well, so we didn’t have to spend a large amount of time on texture variations. In the end what we ended up with was better looking assets that took far less time to make, and that give the players much more control over their characters.
The system does have a few quirks that can make it difficult to work with. The suits work with our tinting system, so in our shader the colors are all sorted out and then our diffuse map in multiplied into it. This tends to wash out the texture slightly and can be difficult to compensate for. We also had to make sure that all the suits were modeled and textured in similar styles. Since you could wear the arms and legs of different sets we had to make sure that they all synched up together and looked like a cohesive set of armor no matter what selection you made.
Speaking of the special armor, how did the Collector armor evolve from the design of the Collectors themselves?
Kolby Jukes – Character Artist:
The Collector armor is essentially an amalgamation of the iconic commander Shepard armor and the Collector Trooper. ZBrush 3.5 R3’s ability to import geometry with different topology/point-orders was a real lifesaver in the creation of the Collector armor. I was able to ‘slice’ surfaces off of the Collector Trooper, then thicken and wrap these plates to the iconic armor rapidly – as opposed to having to create new plating geometry and scrub the detail off the original model using ZProject.
Did you have to give this armor extra attention because the player would generally be seeing it in greater detail than the enemies? Or were even the Collectors as detailed as the player armor?
Jaemus Wurzbach – Lead Character Artist:
Both artists (Leroy Chen and Kolby Jukes) involved in the creation on these assets did an amazing job, and the workmanship looks great. Texture resolution and poly counts in game where very similar, although the player does see the Shepard armor much more often and up close we still use the Collector Troopers in numerous cinematics so they need to look just as good.
How many polygons did the Collector armor come to in ZBrush? Was it built with SubTools and if so, how many were used?
Kolby Jukes – Character Artist:
The Collector armor looks a lot heavier than it is. It’s approximately 4.5 million polys and is composed of 16 SubTools.
Normandy II is a big step up from the original. How did you go about evolving the design while also staying so very true to the spirit of the first?
Derek Watts – Art Director:
I can still remember the day that Casey told me he wanted to blow up the Normandy: instant rage. I screamed, "Are you crazy? The first one took months to design and build!” I was overreacting on this subject and went on many rants about why we blow up everything in video games. Once the reality set in that they weren’t going to change their minds, I convinced Casey to make only minor changes to the original design. We defined the cockpit more and made it smaller to create the illusion the ship was larger. We streamlined the shape and gave it more of a delta wing appearance. We added another level of windows on the side to bulk up the size. We referenced the B1 bomber and the Concorde which still look aggressive for such large planes. In the end the Normandy 2 came out looking pretty damned awesome.
Rion Swanson – Character Artist:
The design was very well planned out in the concept stage, but after the base model was built there were quite a few tweaks and adjustments made to further refine it. Slight adjustments of subtle angles and curves were the major factor that eventually led to the final approved model. It did take some time to get this right, as it sometimes does when translating objects from 2D into 3D. We basically considered the Normandy II a character in itself and had to make sure we spent the necessary time refining both the exterior and the interior spaces. The original Normandy was referenced heavily throughout the process in order to preserve and carry forward elements of its unique character.
What is your favorite ZBrush feature and why?
Ken Finlayson – Character Artist:
I’ve been using ZBrush since 1.55b, so my favorite feature of the software has changed a lot as the toolset has grown and grown. When I first started with it I loved not thinking about the polygons and line flow. I only cared about the form in ZBrush and I could make something look good in ¾ view, which I had always found challenging back then coming from the days of modeling in a quad view. You never worked in perspective, you just looked at the result there.
In a general way I really love that there are often multiple ways in ZBrush to get to wherever you’re trying to go. I think that says a lot. The ability to use an arbitrary mesh with ZMapper was a big deal at the time, and I still find new uses for the rebuild subdivisions function in ZBrush. These days my favorite feature is ZSpheres 2. I’ve dreamt of the ability to sketch in 3D for years, even to the point where I had mocked up an interface for how you would represent your brush position in 3D with a monitor and mouse/Wacom that only has x and y inputs. I always wanted the ability to design in 3D as fast as you could in 2D, or at least close the gap some.
Kolby Jukes – Character Artist:
The biggest step forward for me has been the ability to import models with different topology, point-orders, etc. and have ZBrush propagate those changes through the existing subdivision levels. Being able to add an extra pair of arms to a character composed of a single, contiguous mesh or reroute the edge loops on a face quickly and efficiently is awesome.
Rion Swanson – Character Artist:
I do really enjoy the SubTools functionality, as well as the possibilities that Layers offers. ZMapper has in the past been a very helpful tool also, and I look forward to trying out UV Master within the latest version.
Normandy II is a big step up from the original. How did you go about evolving the design while also staying so very true to the spirit of the first?
Derek Watts – Art Director:
I can still remember the day that Casey told me he wanted to blow up the Normandy: instant rage. I screamed, "Are you crazy? The first one took months to design and build!” I was overreacting on this subject and went on many rants about why we blow up everything in video games. Once the reality set in that they weren’t going to change their minds, I convinced Casey to make only minor changes to the original design. We defined the cockpit more and made it smaller to create the illusion the ship was larger. We streamlined the shape and gave it more of a delta wing appearance. We added another level of windows on the side to bulk up the size. We referenced the B1 bomber and the Concorde which still look aggressive for such large planes. In the end the Normandy 2 came out looking pretty damned awesome.
Rion Swanson – Character Artist:
The design was very well planned out in the concept stage, but after the base model was built there were quite a few tweaks and adjustments made to further refine it. Slight adjustments of subtle angles and curves were the major factor that eventually led to the final approved model. It did take some time to get this right, as it sometimes does when translating objects from 2D into 3D. We basically considered the Normandy II a character in itself and had to make sure we spent the necessary time refining both the exterior and the interior spaces. The original Normandy was referenced heavily throughout the process in order to preserve and carry forward elements of its unique character.
What is your favorite ZBrush feature and why?
Ken Finlayson – Character Artist:
I’ve been using ZBrush since 1.55b, so my favorite feature of the software has changed a lot as the toolset has grown and grown. When I first started with it I loved not thinking about the polygons and line flow. I only cared about the form in ZBrush and I could make something look good in ¾ view, which I had always found challenging back then coming from the days of modeling in a quad view. You never worked in perspective, you just looked at the result there.
In a general way I really love that there are often multiple ways in ZBrush to get to wherever you’re trying to go. I think that says a lot. The ability to use an arbitrary mesh with ZMapper was a big deal at the time, and I still find new uses for the rebuild subdivisions function in ZBrush. These days my favorite feature is ZSpheres 2. I’ve dreamt of the ability to sketch in 3D for years, even to the point where I had mocked up an interface for how you would represent your brush position in 3D with a monitor and mouse/Wacom that only has x and y inputs. I always wanted the ability to design in 3D as fast as you could in 2D, or at least close the gap some.
Kolby Jukes – Character Artist:
The biggest step forward for me has been the ability to import models with different topology, point-orders, etc. and have ZBrush propagate those changes through the existing subdivision levels. Being able to add an extra pair of arms to a character composed of a single, contiguous mesh or reroute the edge loops on a face quickly and efficiently is awesome.
Rion Swanson – Character Artist:
I do really enjoy the SubTools functionality, as well as the possibilities that Layers offers. ZMapper has in the past been a very helpful tool also, and I look forward to trying out UV Master within the latest version.
What is your favorite element of ME2? What is the one thing that we should watch for as we play the game?
Ken Finlayson – Character Artist:
ME2 has really beautiful level art that always blows me away. In our bi-weekly art meetings there are lots of great details like this little gem on Krogan Hub.
Kolby Jukes – Character Artist:
The party members are by far my favorite part of the game. It was a lot of fun watching them come to life, from the initial concepts all the way to hearing them speak for the first time. Garrus, Mordin and Grunt are some of my favorite game characters of all time.
Rion Swanson – Character Artist:
I love the idea of tracking down and recruiting all sorts of unique characters with different abilities and specialties… essentially your own band of mercs. The fact that each character really has their own flavor is great and watching the interactions between them with their sometimes clashing personalities really creates for a more immersive experience. I like the mix of new characters as well as characters from Mass Effect 1 that you run into once again.
Keep an eye out for Mordin’s little number!
(Interviewer's note: Mordin's perfectly-sung take on Gilbert and Sullivan was one of the funniest moments I've ever seen in an RPG.)
Jaemus Wurzbach – Lead Character Artist:
I find the experience of playing the game to be my favorite thing of ME2. It feels much more like a movie that you are directing than a game at times.
Now that ME2 is out, are you breathing a sigh of relief or are your noses already to the grindstone on 3?
Jaemus Wurzbach – Lead Character Artist:
While it’s definitely a sigh of relief to be done with ME2, we have our noses back to the grindstone. No rest for the wicked, I guess. We have a fairly good amount of post-release content to get started on and there are a few new things we need to get prototyped before we start up production on ME3.
What are your main goals for 3? Where do you feel you'll have to focus in order to top what you achieved with 2?
Jaemus Wurzbach – Lead Character Artist:
It’s hard to say right now. Until we get the general story bolted down it’s hard to say where to focus our attention. I’m sure the party will be a huge focus, and I’d like to see as big of a jump in quality from 2 to 3 as we made from 1 to 2.
Let's see -- that would put the GameReview score at 100.33%. That would be quite an achievement, indeed!
We at Pixologic would like to express our sincere gratitude to Derek, Jaemus, Rion, Kolby, Leroy and Ken for pulling back the curtain and revealing these insights into the Mass Effect universe. There's a lot of great information here for everyone from hobbyists to experts. Many thanks!
We also wish to thank BioWare for continuing to push the genre forward, and Electronic Arts for giving the consent to do this interview in the first place.
See you around the galaxy!
cr: pixologic.
