robcat2075

Those animations look very fine. I hope the museum paid you well!

I think Ken touched on it above. I don't recall the original project file, but since you had 12(?) steps in your walk action, A:M was trying to pack them all into the space of two steps (what it was expecting to find in a walk action) when it played your animation back in the Chor. Something like that. I think Ken made it work by stretching out the time that the action was happening over. A walk cycle should just have two steps in it and the "stride length" should be set to the distance covered by one step (heel to heel, for example). BTW, you'll find all of this in "Take a Walk" in your TAoA:M manual. Have fun animating and next time... remember the T-pose.

I really like the shape. The surface texture looks like cement, though. Ok, it could happen.

I took a look at this... here are the initial problems I see: -The character is not modeled in T pose which make makes successful rigging pretty iffy. I realize you probably have some artwork that shows him in that stance he has, but you really should re-imagine him in the t-pose and build him like that. His legs will need to be the same length, at least. If he just needs to carry that sign and walk, just fixing the legs would get you by. - None of the IK is working on the rig, that's going to make a walk cycle hard to do. Some constraints have been left out, perhaps intentionally to save time. That's costing you time in animation. -The walk cycle has more than one cycle in it. Just do one loop of the walk movement, and let A:M deal with repeating it in the chor. But that's a good looking character! These problems are not hard to solve really. I think you could do it in about two or three evenings. PS... Just to see if I could make something out of what we had, here's a walk cycle I did quickly. The walk cycle actually resizes some of his leg bones so they're the same length (sort of) which causes some mesh problems. It's really chunky and stiff, but it is a "walk cycle". Rebel_WalkFK.act

It's a wonderful clip. I'll admit I don't understand the sticky wrist appearance either. I was expecting he'd be digging his fingertips in for traction as he pulled back.

Are these new images intended to be similar in illumination level to the non-radiosity version? The new ones look very, very dark. It could be that my monitor isn't set right, but I have to really crank the gamma to see the details in the new versions, whereas almost any other image I view on the web is fine. ?

As i look at his still image I see that the two lights overlap fairly well while they have the model (the Nautilus? cool!) behind them, but when they have only the camera background color behind them, the rearmost lightbeam takes on a very hard edge. I think this is a rendering problem. The first thing I would try is to put one huge patch in the background and make it the same color as your camera background color. Then the light beams would be rendering over "a model" and maybe they would render as well as they do when they have the sub behind them. Rendering time may go up. I had this problem in a previous version of A:M. I think at the time my solution was to render each light beam separately and composite the passes in After Effects. This would also solve the problem of the lightbeams making the plate seams show up behind them. After Effects may not be a practical solution for you, however. And if neither of the above fixes it... I'm stumped. BTW what's making the dark halo around everything?

I think this is a successful short as is. I can easily imagine seeing this on some festivals I've been to. Short and snappy is always good. Personally, I'm still not sold on the second camera angle (the one looking at her left side). It's quite awkward in the context of the whole movie. But probably too late to reblock your opening. Floaty bit: Her pivot at 00:36-00:37 is more of a slide in place than an actual step and turn. But I like this! The character reminds me very much of an "Artist" I know, so your character animation must be working well over-all to bring up such a specific association.

I'm converting those into links so they will be easier for people to download. www.bee-s.com/~gggg/avi/g3.avi www.bee-s.com/~gggg/avi/g2.avi I always enjoy your commercials! I hope you're submitting quicktime versions of them to A:M films for the "As seen on TV" category.

Fabulous image! Brings to mind a line in an old Billy Ocean song ... "She brushed by me in painted-on jeans"

That looks like a good job of economical modeling! Meaning: not too many splines. I suppose if you wanted doors that open you'd still have some work to do, but with the right textures on it this could look pretty sharp as it is. Hard to tell if you'll have creases in this wireframe view, but I don't see any obvious splining mishaps yet. I think the big sign on the top is like the ones you see on Manhattan or Las Vegas cabs advertising shows. They're practically billboards.

Much better!. There may be something going on under the nose, but we'll sidestep that. The next big issue is the flatness of the face as a whole. real faces bend WAY back. Go to a mirror and tilt your head way up and way down and see the line that your jaw or your mouth or your cheeks make as they reach along the side of your face. Run your hand along them. Are you able to follow them just by moving your hand from side to side? No! Are they flat? No way! We look at our front view in a mirror every day but never stop to consider the real shape of it. And after you do this first face, you'll probably take what you've learned and make a whole new bette one next time out.

Anytime you make spline intersections like the ones Hutch has circled a warning light should go off. Stop and find another way to connect your splines. I drew some new splines in the problem areas that Hutch identified. a "5" shows a new five point patch. where you see a spline dead end into another spline without a red CP... that's a hook. The "?" might be a legitimate 5 point patch but you should really break it into two 4-pointers by extending the nose spline i circled in yellow through the middle of it.

Well, it's getting better... but there's still weird stuff going on, like the way the bottom left box moves in after the collision. But if you covered this up with a big Bamm-crumple-crumple sound effect you might get away with it.

That's a fine first jump in! -Where's the background in the first shot? -The difference in camera view between the first shot and second shot is so slight that it's almost a jump cut. -I like the blue flames -Nice steam effect. It has a jump in it. Maybe that's from rendering it in two parts. Some turbulence might make even more steamy. -The lipsynch seem to be working well. -The arm waving is rather generic acting, but for a first time out I have no business complaining. Are the antennae keyframed or a dynamic constraint? They seem to react sometimes and freeze at other times. There's more working here than not working. I think you should be pleased.

Looks wonderful! What is the little bump/crease/shadow at her elbow? That was the only odd thing that struck me.

Nice looking head! I believe a decal such as that can have an alpha channel in it to fade out the edges. You would add that alpha channel in an image editing app such as Photoshop. Basically most of that alpha channel would be white(255,255,255) (to let the decal show) but it would fade to black(0,0,0) around the edge to gradually hide the decal. EDIT: Shaders? People use that term differently, but in general it would be the colors and textures you can apply to a model that aren't made of bitmaps. If you're not making a body yourself, yes you might weld this on to one and then add some bones to control the eyes and mouth. Mouth rigging is a bit complex, Other people know more than I. And once you have your character rigged you can start to pose and animate it.

I like the snout area. The side profile looks good. However, the inside edges of the ears aren't nearly so close together on my two cats, even when they're listening "forward".

Looks good... hope it builds lots of buzz for your production!

Ah? It doesn't look better yet.

The first test was manually animated, but with so many boxes it became a bit complex.. Let's see... you tried manual keyframing for all of 15 minutes, but how many hours trying to get dynamics set just right?

And sometimes it's easier to animate them manually... A few keyframes per box... you're done. And the boxes are exactly where you want them. I experimented with dynamics for this shot but I think I saved time by keyframing it. Get down from there QT

It is a smooth head, and if it works for him, it works for him. I did the Animation Showdown 30+ times last year and almost every week, there'd be entries done in 3DS. Very chunky looking. I'm sure that's mostly operator error, but it seems that even at the professional level people struggle with animating in 3DS. For example, the troubling animation in Kaena: The Prophecy (a feature done in 3DS Max). I gotta say I'm not tempted to switch, even for smoother heads.

I think there's a passage in "The Illusion of Life" where someone asks how to draw Mickey Mouse's head from the top. The answer was "Don't draw his head from the top!" Toon characters have a way of defying 3D. But in CG you have to consider it, i guess. That's probably more rigorous than most rotoscopes. You'll find the few details that don't match when you start modeling. One thing you might do is make a copy of your rotoscopes and start planning where you'll put your splines. And where do you take a 3D robot for repairs these days?

Cute piece! It's the timing and motion of the hook. The hook should be thicker and it's motions need to be more obvious, otherwise we don't notice it in time. When the n comes back in it should overshoot its spot then loop back to the right place. Or maybe it could even enter from the opposite side. Or maybe the hook could still be chasing it. More work to do that, yes.