Tonight you get a bit of advice on the contrary to the previous lesson. But not until the third paragraph. For now, you have to suffer through how I went making this masterpiece of modern art. Well, first I thought it'd be cool to mess with Corel Draw to make the explosion. Indeed, special effects are as important to JF as the sub animations. Without the animations, nothing is happening. Without the sub animations, you don't know what exactly is happening. Well, in an action scene, without special effects, you don't know what is going on, either. If I just show a bunch of people with guns, I'm going to get the question arising: "Why don't these idiots fire their guns?" Well, to fix that little problem, I've started work on explosion props. When a person is supposed to be firing a gun, a translucent explosion similar to the one shown here will show up. This explosion is from a grenade hitting the stairs. How did I make this translucent? It's not as simple as you think. In Corel Draw, I made a squiggly cloud (symmetric and smooth bezier curves), copied 4 times, rotating each of them radially. Then I added a star. I made a radial gradient with color changes at the 0, 20,40,60,80, and 100 marks to make the thing work. Then I copied the star and rotated it. Then I exported it and scaled it using anti-aliasing (a mixed blessing, you'll see why later). Then I copied that image and made it greyscale. I inverted it and I filled the stars in with white (completely opaque). Then I made a plane in LithUnWrap and textured it. It was very quick. I merged it with sensei's building and moved the plane into place. It worked wonderfully. I really can't say enough about LithUnwrap. One problem I had was saving it as an MS3D file. It saved, but MS3D wouldn't do the transparency right. It just didn't show the explosion at all. The picture here was made in LithUnWrap. Not only that, but there's something wrong with Lith UnWrap's exporting bones into MS3D. I dunno what is going on, but there are work-arounds.
Tonight's lesson is one of futility vs. endurance. Our question is: "Is it worth it to spend a lot of time trying to get something right instead of just doing it half-arsed?" The long answer is that those who set their minds on a subject will ultimately prevail in getting what they want. The short answer is no. Hehe. As you see here, we have the shotty guy from the other night and we have a 3d face with fully usable mouth. In fact, a movie would probably show you that this is better than it seems. But note that Scene 5 will have two of these guys with closed mouth in one 200x200 box and one 200x410 box. I'll also probably have one of these guys in a 200x200 box with mouth open (just to make use of my hard work). Check out the two smallest pictures. That is what it will look like. Anti-aliased, all my work is for naught. I could've put about five minutes into this and gotten what took me an hour and a half. What kind of crap is that? It's the fact that if you want to compile data lossy (anti-aliasing is lossy, of course), you only need a bit of data to make it look right. Putting a lot of time and effort into a few pixels before loss, you'll end up wasting your time entirely. Unless of course, you plan to use it later. I plan to make this an anime, so a movable mouth is good practice at least if not good material for use in the anime. I actually like this model. It started out looking like a guy and then I put the skin on it. I found that the 3d mouth didn't line up with the mouth skin. What's that about? Well, it's easy to mess up position of stuff when you're in 3d. The perspective, the third dimension, etc can mess up by a few inches. When I applied the skin, I looked and decided to move the mouth down. Note how far the nose is from the lip. If you're an artist, you can look in a mirror and look how far your nose is from your lip. It's anime, don't go nuts. The mouth was a major point for me. I wanted to make it look right when it's flapping. That means the lips have to go onto the top and bottom of the mouth hole. Not easy. Secondly, I wanted the mouth to be as rounded as possible. It couldn't be square or even pentagonal, it had to be hexagonal or octogonal. I went for hex and ended up with a weird octagon. It really made me mad more than once. You see, the mouth goes like an misshapen O when open, but then goes flat when it's closed. A grin brings the corners up, a frown brings the corners down, but the middle needs to be in the line of the outer lips. Well, mine didn't work out that way. I couldn't seem to get it right. I know that the lower lip goes down with the jaw and the upper lip curls up, but that didn't help much. It wouldn't do anything that looked right, so I made it into a thin = for closed and an anime-style open mouth. Looking at it, you'd probably say no, but I was going to use LithUnWrap to make the inside of the mouth red, but I am tired. Tomorrow is classes, ya know. You won't see it in Scene 5, but if you buy JF Final or definitely JF Final in 3d Medium, you'll see it big time. JF Anime will make it look really good.
Ah, you didn't think I'd make it. But I did. This is the lesson for 2/2/2002 or 2/3/2002 if you think about relative time and viewership habits. I'll get to that later. For now, let's marvel at this little puke of a picture. It started out as a geosphere. Remember how I told you last time that sometimes extruding and manipulating of normal objects can make good meshes? Well, this is another example. It looks like heck, right? Well, I didn't do much to it other than splitting four faces, flipping a few edges, and moving almost every vertex. The idea is that you have a sphere that has vertices at regular intervals. It works well for construction, especially 6 mile diameter domes, but heads? Not very well. I messed around with it and got a face type thing going. Then the eyes went wacky. So I tried to fix it and it ended up with crazy looking bugeyes. So I did a bit of messing with it and it turned into glasses. So I selected the faces and regrouped and gave it a 90% black material. Not too bad, right? The mouth looks awful, but oh well. 58 vertices and 100 faces. Pretty decent for such a complex face. The back of the head is much better than usual. It's completely ready for realistic looking hairdo.
Aha, another one right when you didn't expect it. This isn't Scene 5 material because I'm in a bad mood. Scene 6 it is and it's all you'll get from me tonight. So, what is it about? Well, you might remember me saying that vertex-by-vertex is the way to go for good models. Well, this is a contradiction. To make this model, I extruded a 1 stack, 6 slick cylinder from the hand to wrist, to elbow, to mid bicep, to shoulder, and to neck. Then I messed around a bunch with the faces and added another row for the chest. Then I duplicated, mirrored, regrouped, and welded each vertex. Not too bad, eh? Well, then I just started pulling vertices each which way. For curvier breasts I did a few "Split 3" operations. That's a nifty function along with the "Flip Edge" operation. It can get annoying from time to time and hard to fix when I mess up, but it works 95% of the time. I messed around with the specular and emissive color on the model. You might notice if you use it yourself. Most people don't use it because it's uses insane amount of processing to do it dynamically. But I say that's a bunch of baloney. You see, if we use fewer polygons and fewer vertices, the lighting calculations are quicker. Thus, specular calculations makes up for a lack of polygons and has a cool effect all it's own.
