Saturday, January 2, 2010

Chernobyl circa 2010.

Happy new years and all that stuff. I'm starting my new year off right by saying that Chernobyl is complete! No reworks. Case closed. It's always hard for me to walk away from projects due to the fact that I'm a perfectionist when it comes to just about everything. But, there comes a time when you have to lock the gate and move on to newer things. So here it be laddies and lads:

As you can tell, there have been even more changes. A list:

1) Four new props: paper towel dispenser, hand soap dispenser, ivy vines, and trash can. Though they're all mundane objects on their own, they help to make the level more believable. A sink without at least something to go with it just looks odd in my opinion. Not too much wear and tear on any of them. They all have normal maps but the only one done in Zbrush was the trash can. It was mainly just to keep Zbrush fresh in my mind for when I use it in the desert project later. Plus the trash bag detail and the grooves on the metal lent themselves to 3d sculpting extremely well. The vines especially make the room look older and overgrown as nature tends to retake things that haven't been kept up in awhile.

2) I desaturated some objects to give the level a more dead, abandoned feel. Some of the objects had too much chromatic pop to them, so I lowered it so they wouldn't stick out so much. The chair and wall tile are the most noticeable by far.

3) I added what is called a post process effect. It is basically a box that makes you able to make changes to stuff like color contrast, which is what I used it for. You might not be able to notice it that well, but it allows things to have more of a contrast that just looks more pleasing to the eye.

4) I darkened and desaturated the poster so that it would blend into the wall better. Stuck out like a sore thumb before.

5) Not big, but I moved the papers around on the floor to better compose the scene based on the new objects I added.

That's it! I will probably never be fully satisfied with it, but compared to how it looked before, I must say It looks pretty good for my first time using Unreal.

So, new things in the mix! I'm still waiting it out for Jake to get me some sketches of the desert, so I'm going to make one more run at the laser rifle and correct some minor issues. I hope to get the ball rolling on more detailed sketches by mid January, but I have a huge amount of things that I can do while I wait, like post production for the demo reel and reworking the website. The gun shouldn't take long to finish, so look for that soon.

Ryan out.

Thursday, December 17, 2009

End of Year Stretch

So, things aren't happening in the way that I expected them to this month. Not that anything usually does. This time it's nothing really bad, just that things will be switched around a bit. Starting off:

Desert: It's been pushed back. The holiday rush has proven to take a toll on all three of us involved (Me, Jake, and Joe). It's not certain how much time everyone can devote to this project while our jobs all take a firm grasp on our throats, so I'm kind of playing it by ear. I'm hoping that me and Jake can have a concept fleshed out by the beginning of the year, and then the more detailed sketches can come by mid January, which is when I hope to start on full production. It's a whole month behind where I thought I would be, but I don't think that's terribly bad, considering that I wasted a whole year on the TF2 project. Here's how the process goes in my head:

1) Me and Jake decide on a general concept. Stuff like where the setting is, the story behind it, a general discussion on how it looks. Not too grounded on details yet. Rough sketches from Jake will accompany this.

2) Me and Jake solidify what we want, and then him and Joe start sketching more detailed stuff, but still probably nothing too insanely detailed. I guess it kind of depends. I'm basically gonna take what I can get. If it's all super detailed, great. If not, I'll fill in the blanks.

3) I start on full production. I will probably work myself like a beast and hope that it gets done by mid March.

In the meantime, I'll be reworking Chernobyl and the laser gun. So, as you can see, I switched the order in which I'm doing things. I have roughly one month of having nothing really big to do with the desert project, so it gives me time to rework the other stuff, and if I'm feeling lucky, maybe I'll even get to rework the website.

Here's a sample of what I've done so far (right) compared to how it looked before (left). I think you'll see quite a difference.

Seriously though, it's like comparing the PS2 to the PS3. No comparison. Richer colors, better composition, better details, better materials, everything. It's mainly thanks to Unreal's new SDK (software development kit) that they released. After the gun was done, I did some relearning on my unreal editor skills, when I stumbled upon some information that said that Epic Games released their full version of the development tools they used to create Gears of War and Unreal Tournament. Naturally, I replaced my crappy version with it, and viola! More creative options! The big one is their new lighting system, called lightmass. It simulates how the lights in your scene interact with surfaces. It simulates how light bounces around, and bleeds colors onto each other, etc. All kinds of stuff that makes it look more realistically lit. Also, I redid the texturing on the ground, moving objects closer together, and completely revamped the medical lamp. Then, I added in some high contrast in Photoshop, which accurately represents how it will look when I run a contrast filter for the video.

And also, the Global Game Jam is coming up again. Hopefully, it will be alot better than last year's, which those of you that read my facebook stories will know, is pretty easy to accomplish.
It's at the end of January, so I'll keep you posted on any developments.

Tuesday, December 8, 2009

Laser Rifle is Complete!

I always find this part of my portfolio work very exciting. I'm proud of myself for getting my last project done, and I'm also excited to get moving on the next one. I must say that with some victory Chipolte in my stomach, I am also excited to take the rest of the day off and catch up on some Team Fortress 2, Left 4 Dead 2, and Modern Warfare 2 (Apparently I like sequels?). As I said, the normal map is what caused the most problems as far as time goes. The diffuse map took me about half a week to do, and the specular and transparency maps literally took only about two hours. It's always discouraging when only the diffuse or only the normal map is applied to the object, because without all the textures applied, it looks crappy. When they are all working together harmoniously, it really makes me think that maybe I do have a clue with what I'm doing. :) I was a bit late on this one, but not by much considering what I learned from it. I originally wanted to have it all done by Thanksgiving, but instead got it done today, which is about a week and a half late. With all that I've taken from this project, that amount of time seems very, very trivial.

BOOM. There she is. One thing I must first of all say is that I learned one huge lesson with specular and normal maps in Unreal. Lighting is king. If your lighting sucks, then you won't see bump details or any kind of glossiness on your stuff. I feel dumb for not seeing it before in the Chernobyl project, but the lighting definitely needs to be reworked in order for more details to show. Since I didn't have any reflection map to work with on this one, i made up a fake reflection in the diffuse map. Combined with specularity, I think it works well.

Here's another closeup of the ammunition holder. I had problems with getting the batteries to look like they're behind some glass, so i faked it by lightening them in the diffuse map, and adding specularity to the transparent glass covering it. The "fan hole" or whatever you call it was added last, without any problems. It bumps nicely and has some good depth to it. I actually used my external hard drive as a reference for painting in the scuffs marks on the silver metal surface, which was a happy coincidence that it happened to be laying on my desk while I was pondering how to treat it. I really had to battle the specularity on the black wiring. It was coming across as having a way too strong shine to it, so I had to tone it down alot. Unreal definitely takes the grayscale value of the specular map differently then Maya. It seems to be much pickier.

I'm really happy with with this one. It's definitely one if the more realistic looking things I've made. Just like the wiring and many other things on the model, specularity was a battle. The dents from the normal map seemed to get swallowed up by the color detail and I was starting to get worried that it wouldn't show. Once the specular map came into play, it was no longer an issue. Just another example of how everything works together.

And the scope. I really struggled in my head with getting the wear and tear to really show on it without totally knocking you over the head with it. I cut back on some of the edge chipping compared to before and made the scratching minimal. Just so that you could JUST see it. I had to make the specularity almost not show up at all in order to make it look like metal. It seems that unreal REALLY likes to make cylindrical objects shine, so that was probably the hardest part with the spec map.

And last but not least, the improvised part. As you can tell from the conceptual picture, the knobs on the trigger were on the opposite side. After talking it over with Jake, it seemed that it would work better on the other side, so that the detail was more spread throughout the model. It worked in the conceptual sketch, but on a rotating model, well, that's different. The extrusions on the ammunition box were improvised due to there only being one side showing in the sketch. So, once again I took inspiration from my external hard drive and filled in the gaps with what you now see here.

A few specs on the model:

triangles- 2,100 ish.
triangles on the high poly model used for the normal map- roughly 16 million, or 8 million polys.
layers for the diffuse map in photoshop- 49
size of all textures used- 2048 x 2048
time spent on pre vis- a week or so
time spent on low poly modeling- a week and a half.
time spent on UVs and texturing- a week
time spent on learning and making the normal map- a little less than a month.
time spent overall- roughly a month and a half

I am doing the same thing that I did with Chernobyl. I'm holding off on any kind of camera movement and final touches for the demo reel, just in case I change things last minute, which I'm sure I'll do.

So thanks are in order:

Jake Rynkiewicz, for of course providing the foundation in which to build upon. Without his concept work, this wouldn't even exist. God knows I can't do that stuff. ;p

Jesse Sandifer, an extremely talented 3D artist. He's the guy I mentioned that tutored me with zbrush. Without him, you can just forget about me figuring out zbrush on my own.

And finally, Joe D'amico for his support on well, everything I do. HIGH FIIIIIIIVE JOE!


And so, in a couple of days I will be posting the beginnings of the desert environment. I need to see where Joe and Jake are at with this, as I know they both work ridiculously hard with their jobs. Expect to see something soon.

Ryan out.

Thursday, December 3, 2009

New Skillsets Abound

So, I'm really happy. I mean REALLY happy. I have finally gotten a grasp on the concept of normal mapping. Not only that, but I have also gotten a grasp on actually making them (GASP!). I kinda see it like this: Imagine you are a mountain climber. You have climbed many numerous mountains throughout your life, but Everest it right in your backyard, and try as you may, you just can't muster the courage to really try to reach the top. Sure, you make tiny little baby steps here and there, but you always end up getting discouraged and going back to the other mountains. Well, that is what normal mapping has been to me until very recently. I can always get to the top now, but I can still get faster and more efficient. Okay, so my metaphors suck a bit, but you get the point. And yes, it has taken me this whole month just to get the map to work, but it's with good reason. Normal maps are ridiculously complex and hard to understand for beginners. It's also very technical, which makes those who aren't so tech savy like myself learn it a bit slower than others. It certainly doesn't help that when you google how to make them, everyone has a totally different opinion on methods. Basically in a nutshell, I thought that normal maps were a lot easier than they really were, and had to study into it more. This probably happened about 3 times throughout making it. I now have my own method that works pretty well, and future models will go a lot faster and smoother. Hey, if I have to take a month to learn one of the most important aspects of being a game artist, then so be it. Very much worth it.

So, onto the pics.

The first one is just a general view. It's pretty much how big the gun will appear in the demo reel, So the closer shots I'll be showing aren't an indication of how close you will be to it. And oh yeah, the two pieces on the bottom right are the planes used for the battery details. They will be inside the ammunition holder, behind some glass when it's done.

Here is a closeup of the ammunition holder. I'll probably add in the round metal rims for the "fan" last, just because, well, I don't know. Just because. This pic makes me realize that I need some light on the bottom of the model, now that I look at it closely. It's really amazing how easy Zbrush makes it to create these kinds of geometric details. The color mapping will make it even stronger i feel.

This view lets you see the scope in more detail. Getting the normal map details to look right on this part was probably the most difficult. Some dings and scratches here and there, as well as more geometric detail to give it more believability.

The actual wood stock of the gun changed the most of all. I originally wanted very minimal wear and tear for everything, but I thought the concept of having the gun assembled from different parts lent a cool concept of having the wood to be beat all to hell, as it's seen a different environment than the other pieces. This also gives me more of an excuse to add more eye candy detail to show off for my portfolio.

So that's it! Now it's on to the color, specular, and transparency maps, which will take probably no more than a week or two. After that it's full steam onto the desert environment.

Ah yes, the desert environment. It's the next thing that I'm going to be working on. For now, it's very early in stages of development, so there's not a lot to talk about, except that it's a desert. I'll once again have Jake Rynkiewicz on for this one, as well as my good friend Joe D'amico. They'll mainly be doing conceptual drawings to aid in getting a vision mapped out for what exact kind of desert it will be. I want this project to be something that you can actually walk around in and not worry about seeing parts that are unfinished. Simple, yet something that you will walk around in for the demo reel.

I still haven't gotten any word from Ross Scott on Civil Protection, and I don't really expect to. No hard feelings, it's probably just a mixture of him being busy with other stuff, and not having any use for someone that can't compile models for the source engine.

So hopefully this gun will be done in a week or so. Cheers!

Tuesday, October 27, 2009

UV Layout

As promised, here is the UV setup for the laser rifle. Extremely boring to those who don't know what it is or what it does. Now I'm all set to start on my first ever Zbrush normal map! I'm not sure how long this will take as it's like I said, the first one I've done. Hopefully it won't take longer than two weeks.

Also, for those who are wondering about Civil Protection, you are not alone. I emailed Scott Ross and haven't gotten a response. Basically, since I don't know how to get my models into the Source engine, I can't make stuff for him unless he finds someone who can help us with compiling them. I've been down that road before with no success, so there's not a lot I can do. He's a busy guy, so I'm not sure when he'll get back to me IF he does, that is. So it's hard to tell what will happen with that.

Monday, October 26, 2009

Laser Rifle Modeling

Well now, here it is. The actual geometry for the rifle is done. I'm really happy with the way this has turned out so far, as I'm usually a little unsatisfied with my models until the normal mapping is applied. So, I'm excited to see how much better it will look with it on. Deciding what the poly count was going to be was a bit tricky, as weapons have a different philosophy to them than the regular props I usually do. There are usually three models used for a single weapon:

1) For the first person view, a higher poly model to hold up at a close distance.
2) The model that you see other players holding, which is about mid poly range. Some games will switch it to a lower poly mesh when it reaches a certain distance from you.
3)A pickup gun mesh. You know, the one that floats around or sits on the ground for you to pick up. Usually also a mid poly mesh.

So that's the dilemma. Which do I do?

I decided to do the mid poly version, because it seems like a good compromise. Doing something too high poly might make it look like I can't conserve geometry, and a low poly version for far away distances makes it look like I can't model in a complex manner.

Once I decided to go with that, I needed an example to follow from some game. I chose Team Fortress 2, because its model viewer lets you see the exact geometry display and count. I looked for a weapon similar in complexity and found that it was twice as many polys as I had on the rifle. TF2 models are usually very low poly with it being an online only game, so my model looked extremely well budgeted. I added some geo here and there, and came to a good looking model that conserved polys where it needed to.

So now, I layout the uv's, make the zbrush model, apply it as a normal map, paint the color map, and then finally the specular map. Easy right? :P

I'll update next with the finished uv layout, which will hopefully be tomorrow. After that, it'll be the normal mapped model alongside the mid geo model with the normal map applied. That will be sometime in early November if all goes to plan.

Monday, October 12, 2009

Chernobyl is Complete!!!

Hello all, that's right! It's done. It feels really good to have something finished that wasn't either A) Rushed too much for a school deadline, or B) Incomplete due to technical issues. The normal mapping and lighting gave me some troubles, but specular mapping was a breeze. I couldn't use Zbrush to sculpt out normal maps for this project due to me not realizing that you can't do some stuff in Photoshop and some in Zbrush. Apparently, when you make a model, its entire texture has to done in one program. Which is fine, that's what this next project is about (More on that later). The book I bought helped a little bit with the lighting, but not as much as I had hoped. Mainly, I just figured it out on my own. Lighting and the way normal maps react to it is still a little shaky with me, but oh well. You can't get everything right away. Anyway, here it is folks:

Sorry about the text and player model on screen once again. I will have this cut out of the version for the demo reel. Also, anti aliasing (jagged edges) are still present, which hopefully will be straightened out for the reel as well. I went online to figure out a way to remedy it, but everybody says that anti aliasing hasn't been fixed for unreal 3. I find this hard to believe. One way or another, this will be fixed, I assure you.


Now this project is done for now, but there will probably be changes for the demo reel version. There is one object in particular that might be changed, but I'm curious if someone will notice it, so I'm keeping my mouth shut about it. Two big changes that you will probably notice are the metal piece added onto the middle of the medical lamp, and the broken glass on the window. Just little things to give it more completion.

The final product will be pretty much the same as I originally planned, with a little bit of difference:

1) A moving shot from left to right of the objects, framed similar to these pics.

2)Still shot of the left side, showing a progression of the different layers. In order of final, model and lighting only view, polycount and wireframe view, color map view, normal map view, specular map view, and a shot of the final product again.

3) The same as #2, but it's the pic of the right side of the room.

The second shot I planned out in the very beginning, I decided was unnecessary. You only need one shot to establish the room I feel. So, that's that.............NEXT!
BAM! That's the gun I will be making, courtesy of the tasty rendering you see by my good pal, Jake Rynkiewicz. Over the next week, I'm gonna take it easy to celebrate Chernobyl's completion. But I'll also be planning for this one. Expect to see some stuff by the beginning of November-ish. The finalized version of this will be a turntable that rotates the object and shows the different layers of detail applied to the model. I'm going to try to make it a ZBrush only model, so we'll see how that goes. I'm going to do some online tutoring on making geometric detail on objects in ZBrush. Basically, this guy I met on a forum says he'll make a 60 minute training DVD for anyone who wants to get better with it, and it will be completely custom made to suit my learning needs. He also seems to be very professional and legit, which is a plus. Very cool stuff. He says that he'll also give homework assignments to make sure it all sinks in, so I might have that to show even before the gun. Time will tell. Now, for some well earned celebration Chipolte.

NOM NOM NOM NOM NOM!