Time, Time, Time…. see what’s become of me.

Ah most companies and people will talk about their successes and things they did right on a project but here, and now, I’m going to share some of my mistakes with you. This is what I’ve figured out in my short tenure in the game industry.

1. Find out what resolution the programmer needs from you graphics and do it. Then double check that you are actually working in that resolution before you finish the graphic. In fact, get a blank template and then you won’t have to worry and it will solve your next problem…

2. Keep the graphic/picture  ”tiles” in multiples that will fit easily onto a game board. For instance we have been working to develop for the iphone which gives me limited screen room to work with. The pieces on the board need to fit as snugly as possible so that I don’t end up with wasted space on the screen. By keeping things in powers of 2 like 256 x 256 or512 x 512, I can do this.

3.  Drawing and then scanning your work in is a great idea, but it works much better if you do a digital version on top of the original work. I drew my pretty pictures and then scanned them in hoping to clean them up to use in game. I ended up spending a lot of time cleaning them up while I could have just done my  line work digitally. Most of this came from artifacts that we kept finding as we tried to use the art in game. or every time I thought that I had caught all the specks and grime I would find more. Ink it in photoshop over your original work, it WILL go easier.

4. Keep “tiles” the same size. If you are designing a bunch of farmers (like I have been doing) it is much easier to start with them at the same size and resolution. Using templates is always a good idea in a production environment. It makes for less mistakes and for and easier creation process. Anything you can do to make your pipeline move easier the better you will be.

5. Remember your platform and design accordingly. For us, with the iphone, I have to consider the size of the graphics that I’m creating.  At first I spent a lot of time on my drawings and made them fairly detailed. After the first few test runs of the game I realized that most of that detail is getting lost when it is viewed at actual screen size and resolution. I then switched to less detailed pieces although I did need to keep track which ones may need to be made bigger later.

6. Find a palette that works for what you are doing. This sounds simple enough but it hasn’t been for me. At first I found myself going with colors that worked for each characters but then, when I put them together, they didn’t work. I then sat back and set up most of the palette that I felt I would need. I also limited some of the tones and forced myself to stay within the palette that I have set. As a results the characters work better together and again, the production sped up.

7. Concept, Concept, Concept. Once you get going you may try to get around it but you still need to do it. After designing a number of the animals I found myself trying to skip the concept phase of design and moving straight to what I thought would work. But in the long run I found that I didn’t like the end product and I headed back to the concept board to do it right.

So there you have it. Some of the mistakes that I have found myself making so far in this game making process. I know that I’m bound to make more and that I hope to keep learning and sharing what I’ve found. Good luck to you and if you see something that I’ve missed that was a big lesson to you, please share!

–David

When I left it, I had some feedback on my basic mesh, and before I get into the second part of this post, I’d like to say a thing or two about feedback.

Feedback, is one of those things that helps an artist, or maybe even programmers get better. When someone asks for feedback on anything they’ve done, they’re asking for your perspective, your expertise. If you dont have any expertise in the area, and the request still comes in, then feel free to offer just your unique perspective.

As an artist, feedback is one of the first things I learned to give, and more importantly take. Its that crucial to development. Heres how this lesson unfolded.

First day of Figure Construction (basically anatomy.). I heard the rumors about the class, having live models. Naked ones.  Not only that but I think i also heard that my instructor was a complete bastard about criticism, dishing out harsh criticism, and also being a slave driver when it comes to drawing, 20 pages, front, back, no spaces between drawings.

Minimum.

Do this, and receive a D.

Walked into class, was thrown off balance by the nude male model, and the fact that the only drawing position was right in front of him, and on a drawing horse! Anyways, after about 30 minutes of posing, drawing, and our instructor barking orders “Draw like you mean it! Stop masturbating the pencil! ” etc, we have to put up our works on the wall, walk around the room, then give other fellow students criticism.  Then we got feedback from the instructor.

My dread was so thick it was running down my pants-leg. As an artist, I’m fairly humble. I don’t think my work is any good, I see all my mistakes. Foreshortening, perspective, gesture, balance, each of this mistakes on the page mocking me…and now the entire class is going to see them, and I’m going to be a running joke. In my first trip around the class, I was thunder-struck. All these guys and girls drew so well. They didn’t need this class! Oh crap.

So I was worried, humbled, embarrassed when it came to my drawings, but the instructor gave clear instruction. “What do you like about this drawing? What do you think doesn’t work?”  2 very clear questions, and almost a structure for feedback. Amazing. Its important for people to hear that you can see something good, or ok, or that you like about the work you’re giving feedback on. Even if its minor, otherwise your feedback comes off as petty, inconsiderate, and has the opposite effect that good feedback has. Then you can state what you dont like. And the key here is to communicate that in a respectful way that doesn’t alienate. Giving poor feedback to anyone never helps. Don’t sugar-coat it, and don’t wield it like you would a chainsaw in a zombie infested mall.  I could probably write another post on it, but I’ve already written too much. Lets dive into the changes I made, based on the feedback from my team.

Here’s my second attempt:

So a few changes. Believe it or not, as easy as this was to put together, I learned quite a bit about how to pull this stuff together.

First of all, the skybox. This was rendered out of a 3d package Vue, which, is expensive but luckily I have friends who don’t have the financial constraints that I do, and I was able to poke around for a little while and get this lovely skybox rendered. After a few tweaks and getting this imported as the unity3d skybox, I noticed that my geometry felt very out of place, and that was due to the specular or “teh shiney” wasn’t in the range of the rest of the environment, and I needed to figure out how to make the scene feel more cohesive. Thats when I added the fog color.

The fog color I pulled directly from the skybox. That gave the impression that it was dusk, or…dawn…which is the setting one?  It also had the added benefit of extending the visible light into the horizon, which gives the scene a bit more atmosphere. Theres only 1 light in this scene, and its sun…but not the one you see in the pic.

The other thing I did to this mockup, was add the hills in the distance, but the terrain there felt very empty, and I wasn’t feeling it. So how to spice it up? I did a very subtle thing, I added trees just on the other side of the hill to breakup the horizon line, and hint that theres a different scene just over the ridge. The figure in the foreground is there for scale reference.

There are other touches, like the sand texture change on the terrain where the rock wall is. There’s also some shrubbs that i placed for a bit of interest. Some final sculpting of the terrain really makes a nice scene.

Heres the process that I used in creating the screenshots below, and hopefully along the way, I can share info on my process, and maybe even learn something.

Lets go!

First thing I wanted to do,  was test out some geometry in both Max, and Zbrush. One of the problems we seem to have had in our previous efforts in designing is the” iterative-loop”. Thats a loop where an artist or programmer can see his / her work quickly in the game engine (in this case, Unity), and quickly make changes to that work, to iterate, again and again. Its how things get progressively better.

So first, being an artist, I wanted to try out the pipeline moving from one package, to another, and finally into the engine. So I quickly make some geometry in 3ds Max, making sure to keep only polygons (4 verts!) as we discussed in previous posts.  My Initial attempt left me with softer edges, and area’s where I didn’t get the resolution i wanted once I got into Zbrush, so I went back into Max and made the right adjustments to address those problems.

The exporting from Max to Zbrush was a bit tricky this time, I exported a mesh that didnt have a “bottom”. Its a poly saving technique thats becoming increasingly unecessary, but what happens in Zbrush when you import “open” geometry is your normals get all inverted. No problem, I sealed the mesh off again, and went back into z-town. I could have inverted the normals within Zbrush, but I didn’t want to create any further problems down the line.

In Zbrush, the thing you want to do with any rocky type surface is to make sure you turn off smooth edges for the first couple of subdivisions. This will give you some harder, faceted looking faces to work with. I subdivided 3 times in this mesh before turning smooth on, then subdivided even more. This mesh got up to about 4.5 million polys before I finished.

I used a collection of alphas from my own collection, and several of the resouces found on pixologics site to start working the rocky cliff faces, and layer upon layer, slowly working up the detail chain until my result was thus. I dropped down to the lowest subdivision level, and exported that as a OBJ, and generated a normal map with z-mapper.

This is all fine and dandy. If youve worked in this pipeline, this is nothing new.

Unity Time.

Now I took the .obj exported from Zbrush, into unity. Drag and drop. “It just works.” It comes in with a material assigned to it, untextured, and looking just as I intend. No fancy exporters. No weird intermediary formats, just an OBJ that nearly every 3d program can create, and unity eats it and asks for seconds. So i get to work on getting the normal map assigned, hook up the specular channels, and proceed to play with lighting a simple scene:

2 Lights:

-1 Directional light, which throws light in a direction for all objects in your scene. Think of it as the sun.

-1 Spot light. This is to highlight a specific area of interest. Also gives the materials specular highlights, and makes normal maps happy.

Looks ok. I replicated the geometry a couple of times, scaled and rotated the piece to be a bit more random. Not bad for an iterative start. I sent this screenshot to my coworkers, got some feedback.

-Looks a bit dark.

-Looks a bit shiney.

-Not much to go on.

-Questions about art direction.

All very valid points, and this being an iterative test, I take that feedback, and begin to work on a brighter, outdoor scene, with scale, and life, and some sense of scale. I’ll cover my follow-ups in the next post.

Check back soon!