Mime's Animation Sketchbook
polycounter lvl 14
Hi,
I am by profession an environment artist. But recently i am getting more and more attracted to animation.
So i decided to start learning some in my spare time.
I started with the basics , bouncing ball (3ds Max)
Then i started following some tutorials, and as i realised that what seemed the best are in Maya, i switched to that so i could follow more easily. (My main modelling package is Modo with some 3ds Max mixed in)
This is my latest work, trying to add some character to the ball.
Should i start tackling a simple walk cycle , or do the ball with the tail ?
Guidance and feedback is greatly appreciated.
I am by profession an environment artist. But recently i am getting more and more attracted to animation.
So i decided to start learning some in my spare time.
I started with the basics , bouncing ball (3ds Max)
Then i started following some tutorials, and as i realised that what seemed the best are in Maya, i switched to that so i could follow more easily. (My main modelling package is Modo with some 3ds Max mixed in)
This is my latest work, trying to add some character to the ball.
Should i start tackling a simple walk cycle , or do the ball with the tail ?
Guidance and feedback is greatly appreciated.
Replies
Bounce 1: It's a little hard to crit with a flat texture like that. Without being able to see the spin, it seems like it's sliding at the end. Also, I feel you have too much squash and stretch on that last few bounces. It feels very wobbly, which takes away from it's the belief that it's a bouncey ball, rather than say a water balloon. It also feels like it slows down too quickly. you could probably drag it out for a few more bounces.
Bounce 2: Love this! lots of subtle stuff going on. The slight bounces as it hits the ramp going up, or when it hits the ground coming down really shows the realism of the ball. Going up the ramp though, it does feel like it hits a brick wall at the top. The timing of hitting its peak is off. It also slows down a little to quickly at the end of the animation. Just slightly thought. It feels more like a car hitting the breaks at a stop sign, rather than a ball coasting to a stop. You may also be able to speed up the initial fast spin that gets it all going. just the initial one though. Play with it a bit. Push it further than you're comfortable with, then tone it down, if need be.
Bounce 3: Really good stuff! Can definitely tell the ball has personality, and that this is your most recent anim out of the three. Really, the only crit i have on this one is with the second jump to the left platform. After the ball reaches the peak of it's jump, it feels like it speeds up a bit mid air. Once an object is in the air, its' center of gravity has a linear movement. It cant speed up or slow down without extra force (and wind isn't enough!). So just check out your curves and you can probably just shift the translate y (up direction) over a few frames so it's in the middle.
With all that said, I really like these! very well done, indeed!
I would say after you do some fixes, move on to a walk using a ball and legs. http://www.cgmeetup.net/forums/files/file/95-ultimate-rig-walker/ You`re definitely ready for it. Walk and Run, if you have time!
Also, I editted your title, just to avoid confusion if we get more people doing sketchbooks.
I downloaded the rig you mentioned, got excited and tried my hand at a basic walk cycle.
Here's a rough block-out
I will revisit the bouncing balls but i was too excited to try this out.
I have a couple of questions:
Should't a rig like this have a key-able root ?
I watched a tutorial a while back and it was mentioned that a walk cycle should be animated moving and not in place ( for beginners).And i did spot a lot more mistakes that way,so he was right.
But for this rig i had to manually move most of the controls and key them moving (in the direction of travel). I would have thought it would be easier to key a root or a parent to all the controls, moving horizontally , and if needed remove that animation so it becomes a walk cycle in place.
Does this make sense?
Feedback and help appreciated as always.
It looked weird , slow and not really good.
So i remade most of it today. This is what i have so far.
Comments and feedback appreciated.
But in this case the root "had no keyable attributes". So i had to key the feet and hips on the Z axis ( in this case). And to turn this into a 'in place' walk cycle, i would have to redo most of the feet positions, since they move in Z with the hips and relative to themselves.
Anyway i worked on it a bit more and made it longer.
I think i'll upgrade now to a full body rig and do a walk cycle on that, see how that goes
I started work on a full rig today. Still WIP. (damn knees)
Comments and feedback is appreciated.
Great so far, very purposeful walk
I got some more work done on this over the weekend.
I'm gonna call this done , it seems the more i work on it, the more i break it
I'm will start working on a run next, see how that work out.
I do have a question. On the walk cycle : does the elbow or the shoulder lead the motion of the arm , i've watched a lot of reference , but i just can't tell.
As always feedback appreciated.
I refined it some and i will work on it a bit more. Hopefuly i will get it done by the end of the day, if i have a bit more free time and i will post it.
Thanks for the feedback everyone , every bit helps.
Here's what i have so far. I'm sure it's not perfect but i don't know what to improve on it. Any ideas ?
Maybe someone can chime in with more ideas for this cycle, I just want to see what you do next!
I toned down the arm swing a bit.
Here's the final animation:
I have a question about curves. i bumped into this while animating the arms.
I started by setting the key poses and refining from there. But the speed of the rotation of the arms was inconsistent. So i went into the curves editor and saw that the curve was a bit jagged so i deleted most of the keys leaving a nice curve.
Is this bad practice ? How do are you "supposed" to do something like this?
I attached and image to show what i'm talking about. The curve above is the final one and the one below is the one from the block-out.
Be careful of the curve editor, it can be quite a addictive drug. Work with it, understand it, play with it, try it! Then discard it.
I hardly ever go into it, the only reasons I open them is to:
- Get rid of any keys on top of each other
- Euler filter.
I'm sure there are other good uses, for instance working with looped animation and scaling/moving keys about the timeline.
I use the dope sheet a little more then the curve editor, since I can scale my keys about but still keep them on whole frame rather then 0.234 frames and other messy decimals.
If your arcs in your animations aren't working right, practice your eyes to try and fix the problem. Delete and try again
Happy animating!
The problem with doing too much with the curve editor is this. The more time you spend in the editor shifting keys and smoothing curves, the more you destroy your initial pose. Posing is a huge step of animation, and you generally spent a lot of time finding the exact pose you want. But every click in the curve editor will change that pose ever so slightly (and not so slightly, sometimes). If you do your posing correctly the first time, and have enough poses in your block out, you wont need to do too much in the curve editor. But this really does come with time. You`ll ALWAYS need to go in and fix the loop, when doing a looping animation. And the editor can be extremely useful to diagnose a problem area though. If you see something that seems off, but cant really tell why. Whether its a pop or the timing is off on something. You can generally see it in the editor.
Now, with that said, you shouldn't be afraid to delete keys if you need to. I do all the time. Sometimes it's the best solution. Just keep in mind your posing, and if you delete a lot of stuff, make sure your pose still stands strong.
As for your run cycle. Your hips are tilted too far back, like his ass is sticking out. Bring that ass in more and it will help make it more natural looking. And is that left pic full on front view, or is it tilted a bit? It's hard to tell if it's the animation or the camera angle.
This is stuff i can't get from tutorials.
I've started work on a simple attack animation, and i'm starting to understand why poses and breakdowns are so important.
I'll try to keep in mind what you all said , about the curve editor.
@slipsius i'll fix the hips . i made the clip by importing the scene in itself 3 times. so the front and side are not truly side and front. it's all from perspective view. Is there an easy way to get true front , side and perspective, without having to comp 3 videos ?
@peterhawk i'm not use what the euler filter does. i'll read up on it.
Right click the camera icon in the upper left area of viewport, under "View" for Perspective and Ortho views.
@peterhawk : i'm sure you know this, but if you scale keys in the maya timeline, if you right click there's a snap option so you don't get frame 3.56
So i managed to block-out a simple loop-able attack was just starting to smooth it out, and realized it was all kinds of wrong.
The attack hand is the same as the forward leg, everything moves at the same time and more.
So i remade it today from scratch.
Here's what i have so far:
Like Slipsius said above it is not a tool you animate "in", you animate in the viewport, but you will also end up using the curve editor to "move" your character by tweaking keys in there while you see what it does in the viewport. Tweaking curve scales/timing/tangents/ and of course deleting untold numbers of keys. There are too many reasons why using the curve editor is important to go into but its not something you discard, as mentioned above.
Also Euler filter is to solve continuous rotation issues, not gimbal lock. Basically over 360degree rotation, software sometimes has a hard time understanding if you want to set a key at 450 degrees or 90 degrees and invariably your curves will suddenly shoot down and rotate everything backwards. The pose looks fine but the interpolation will make it rotate backwards suddenly. This tool will basically fix those keys by adding however many 360's needed to make the curve look like a slope instead of a crazy jagged mess.
Gimbal lock on the other hand is resolved by changing rotation orders (or planning correct rotation orders to begin with), or by constraining to another null temporarily with different rotations. Gimbal lock itself is when e.g. x and y are overlapped, so you can only rotate in the xy plane, or z, so you've lost one entire axis for a bit.
Mime, your second one is a lot better
I would say his left foot is rising up a lot for the movement there, barely have it come off the ground, since humans are lazy/efficient, we usually do things in the most direct way and rising your foot that high off the ground isn't natural. Instead maybe use that pose time to roll the ball of the a bit as it comes off the ground to move forward.
I would also say try adding a few more frames of the hips > torso starting to twist and drag that sword up out of anticipation, it seems to lack a visible source from which it starts moving, and with basically all of our big motions they start at the hips and work their way down.
Overlap the two arms in your blocking rather than later, they are both rotating forwards for the strike at the same time without working together to help twist that torso. Same with the followthrough, right now your posing lacks any followthrough with his left arm, head, hips etc.
This advice comes from my leads and seniors and they have been at it for +10 years. I barely see the editor on their screens. It's a waste of MY own time now I believe, only use it for technical problem solving. My animation can come out a bit faster now.
Plus the idea of happy accidents happening when you are fiddling with the tangents and moving about keys. I find more satisfaction in knowing that I did something because I saw it. I've heard quite a few big shot animators talk about working on a linear tangent. And just adding the extra keys needed that the curves would do.
As far as graph editor use, everyone will give you a different answer.