Zounds and Hands!

Friday, October 7th, 2011

No, just hands.

I’ve been working on these hands for the past couple of days trying out a new way to model things like hands, and really most organic forms. A lot of people favor a cylinder or box extrusion method for making organic shapes, but for a base mesh like this, I find the best way is to actually make small elongated boxes for each main shape, and then form all that together over time. The advantage of this is that you can focus only on getting the form and proportion right (based on reference) without having to worry about poly count and topology.

I also learned a neat trick from somebody in regards to hands, and that is, don’t model them all straightened out, instead model that in a relaxed pose, and not only will you get better looking results, you will also end up with a hand that is much more useable for animation as a relaxed pose with the thumb in kind of a relaxed angle off to the side is much more practical for things like gun poses and gripping poses than a stretched out hand. This also prevents major UV stretching when the hand is in the standard “90%-of-the-time” pose.

You may not know this, but character models in games like Unreal Tournament 3 are often modeled, UV mapped, and textured with the arms bent at 90 degree angles. The reason being that since the characters will run around with their arms bent holding guns and such nearly the entire time (with the exception of cutscenes and deaths), it makes much more sense to just build them that way, and then have stretching only during moments where you’re less likely to take notice.

Combat Rifle and Marmoset

Thursday, October 6th, 2011

Hey all. Trying to do daily stuff now. This is an assault rifle I’m working on. It’s basically an “M16-ish” rifle, not real unique or anything, but it was a great learning experience for me.

First, I’ve started learning recently that the actual poly count of the final model will be greatly influenced by the original concept art because the actual shape of the gun in a concept will actually require certain poly count thresholds to look decent in final form. Now the M16 (if faithfully modeled) will actually have a pretty high poly count, so what I did was set a goal of 1000 polys, and them simplified the basic silhouette of the gun until I was sure it would not exceed my budget. As it turned out, the final gun was about 950 polys, so I’m quite happy with how that turned out. As 3D artists often say, use large shapes, and don’t try to slap artificial detail in those areas, otherwise it’s just noise.

These renders are actually in real-time using the ‘Marmoset Toolbag’ available here:http://www.8monkeylabs.com/toolbag

The Marmoset Toolbag is amazingly useful and has completely changed my workflow. I’ve been able to work much much faster now that I can preview all my changes in real time including the spec map which needs real-time interaction the most. The texture on this model is only about 70% complete, which is why the paint scratches are mostly just straight lines. The base version is the tan, and the other versions were just for fun. :3


Also, I’ve started a new system of texturing where I very slowly build up the material definition of each part, while fully integrating basic normal and spec at the same time. Never cut directly to 1px brush scratches, you will just end up disappointed in the final product. For those who don’t understand the concept of material definition, it is basically the process of identifying each material you want on your model (hard plastic, rubber, painted steel, ect) and then building up hand-painted and photo layers until those areas actually look like that material.

The easiest way to go wrong with this is with metal textures. I can’t tell you how many time I have seen others (including me BIG TIME) make a “metal” texture that looks far more like concrete than anything else. This comes from not really examining what makes metal look like metal, and also poor knowledge of normal and specular maps Always pay attention to material definition.

Sketching Guns

Thursday, October 6th, 2011

Holy cow does time fly. I’m so sorry for the downtime everyone!

I’ve been doing a lot of gun sketching lately to improve in creativity and structure of my gun sketches. I made a whole bunch, and they’re posted here. If you draw guns for modeling as well, something you should remember is this: Never TRY to be creative. If you just draw lots and lots of guns, you’ll find yourself making cool things up on the fly, but if you try to make each and every new sketch cool and amazing, you’ll fail because you will spending 3x as much time on each drawing. If you want a revolver for your game, grab some ref pics, look them over, and start drawing. Draw like 20 different revolvers. You’d be surprised how often you’ll end up with something very cool and good looking.

Also all of these were drawn without an eraser available. I don’t normally draw without an eraser nearby, but I was glad this time because it forced me to use each mistaken line as real detail instead of just erasing and being overly-specific.

“Gameplay First”

Monday, March 14th, 2011

I was watching a retrospective video about Blizzard the other day, when I heard one of the company’s founders say something that, honestly, I have very much neglected in my game development. He said, “Gameplay first.”

I’ll repeat that.¬†Gameplay First.

This is the kind of thing that should be written on the wall next to your computer, because it is that important. I have worked on quite a few indie game projects in my 6 or so years as a game developer and artist, and the two things they all have in common, is that they are unfinished or canceled, and they started with assets first, gameplay later. I think one of the most common reasons indie teams fail at making a game, is that they begin with a simple idea, and then proceed to model, and texture, and map away, forgetting that at some point, all those assets will have to come together to make a complete gameplay experience. They spend their time creating art and fx, because that’s all they can do. They have no AI, and they have no scripts to incorporate their assets.

Do you want to know why Marathon: Infested isn’t release yet? Simple. Nobody has yet been able to make a decent AI in dim3. I loved fighting the AI in Marathon, and it is one of the primary reasons I still play the game. The variety of monsters, the varying aggression, “tactics”, and the hilarious weakness to circle strafing makes for a very entertaining game experience. Marathon: Infested was supposed to be an epic “Durandal vs Tycho” fight (if you’ve never played Marathon, skip this part), and included a huge variety of locations and creatures, including the infamous “Battleroids” which were the center-piece of the entire M:I story. The levels were all planned out, the story was all written down, the monsters were designed, but even after years of¬†development, the most I had finished was a complete 3D recreation of the Marathon Infinity multiplayer. The campaign? Nowhere to be found. The reason? I had nothing to play. Sure, it’s fun to run around a cool looking campaign map, and shoot weapons and read terminals, but without monsters to fight, I had nothing. Nothing kills motivation like building a whole campaign level, and getting bored after playing it once. The reason I still have some motivation left to finish this game, is that about 2 years ago, I played a very hacked together multiplayer match in Marathon: Infested with two of my siblings. It made me so happy to see people actually playing the game I had spent so long working on. They ran around the map I had built, picking up weapons I modeled, and blasted each other apart with funny animations and big explosions. At that point, I threw the campaign out, and decided the only chance I had of finishing this game was to go for a multiplayer-only experience. Trouble is, the folks who had been following my work for that long, had seen all the monsters I had made, and I knew they would be disappointed if that was all ditched in the process. So I decided to add a “Survival” mode to the multiplayer options, whereby gamers could work together to defeat waves of alien attackers in classic Marathon “run-n-gun” fashion.

Except, oh wait, I forgot something…. I still have no AI.


The moral? do yourself a favor, when you have that awesome idea for a game, don’t write in the design document, “Story: … Enemies: …. Guns: Pistol, rifle, rocket launcher..” Don’t make your first priority be to model a gun, or design a character, or make a texture. NONE of that contributes to the immediate gameplay experience. Instead, design a map for the first part of your game story, throw in some boxes to be placeholders for cool guns and stuff, and get you some enemy AI to fight. If your game is an RPG, make some placeholder characters, and 1st map, and start adding simple dialog options that affect the game world. You should be able to enter a box marked “INN” long before you can see a beautiful 3D model of an actual inn. If it’s a puzzle game, design a simple map, and start working on puzzles to solve. When you play your game, you should be challenged, you should have to fight, or think your way through the levels.

Add all your visuals later, especially if you are the only developer. Large companies like DICE, Bungie, Nintendo, and Blizzard all have separate people who program, create art, and write story. The guy modeling an SCV for StarCraft 2 isn’t concerned about gameplay, because the programmers and designers are working on that along with the mappers. Nobody cares if Mario actually looks like a real plumber if he can’t jump or walk, and nobody would play Battlefield 2 if all you did was spawn in your base and look at pretty tanks.

Do yourself a favor, if you can’t play your game yet, you need to close that “ak47rifleawesome.wings” file, and start adding content that actually means something. I’m not just telling you this, I’m telling myself this as well. All us indie devs, lets work on what matters, and show off our cool stuff later when we have something to use it in.


In conclusion. We are not making tech demos. Gameplay First.