Rigging a simple character -- can't get IK settings right, other issues

Not too deep at all. That kind of insight is exactly what I’m looking for. To me, redoing the pelvis and legs sounds more reasonable and possibly easier. I will try to do that this weekend. I’ll look around for some examples of good topology for that area.

As far as the shoulder copy rotation constraint, I guess I was just asking if that was generally considered bad practice or anything to do that, if it violated any principles of how the bone hierarchy should be etc. But it seems to work decently for me so I’ll probably keep it.

I will indeed look around in the unity docs/forums, good call on that one.

Thanks so much for your help!

f I want to allow transitions between animations, do I omit keyframes on 0th frame so that whatever position the bones are in before the animation starts, they can interpolate toward the first pose of the new animation? Or I guess, how does that work in general – I’m wondering how you should choose a starting pose for an animation

Okay… some of the answers to this probably won’t apply when your using a game engine…

I’m not a game engine guy… so I can’t say for sure…

but in Blender… if your just trying to make movies…

there is the many animation tools…

once you start into animation… trying to build a walkcycle is the first step…

here’s a great primer for this…

There’s an example at cgtalk of spiderman that I think is considered to be about the best topology around, might want to google that. I looked at a character I am using, and it’s topology isn’t as I described, but it works well. here’s a pic:


As far as the shoulder copy rotation constraint, I guess I was just asking if that was generally considered bad practice or anything to do that, if it violated any principles of how the bone hierarchy should be etc. But it seems to work decently for me so I’ll probably keep it.

You’re the one using the rig, so if it works for you, use it. But a word of caution here… You just can’t go around slapping constraints on bones as you want. Sometimes you constraints can conflict with other constraints, or they might conflict with the bone hierarchy. This is called a ‘cyclic dependency’ and the usual symptoms of this are bones that flip wildly around, totally out of control. If you have this happen to you in pose mode, tab into edit mode and tab back out, then check blender’s console window for error messages. It will tell you what the conflict is.

Oh, and I forgot to mention, I think the pole target rotation is off on one of your legs, you can see it by jumping from pose mode to edit mode. Good general practice to keep your bones inside your mesh, your arm bones are a bit above the arm mesh. Could cause auto-weighting issues. Oh, and I think you want at least 1 more loop cut around the elbow area, min of 3 loop cuts at a joint. See above pic and look at the loop cuts on legs and elbows…

wish I had more time to spend pointing things out… but you’re making good progress…

Randy

Thanks for the reply – that’s a really good example. I’ll keep looking around for the Spiderman example too.

I’ll probably do a longer post in the morning with my issues I’m still working on, but here is my new version after all my changes today.


Once I understood the principle of what you were saying for the crotch/leg topology, it wasn’t that bad to implement it – of course assuming I haven’t just misunderstood it again in a different way! :smiley: But I ended up basically remaking the bottom of the torso as if it was a leotard, then extruding the leg out of the side loop by loop and rotating the loops to the correct orientation. I already notice a huge difference in the area when the legs are lifted high up, so thanks for the very helpful advice.

As far as the constraints go, that’s pretty much what I was wondering about. I might have to fiddle with it and test out if it causes any problems. I’ll check out the pole angles as well.

To add my random thought as I always do to the end of my post:
All I can think about as I watch all these retopology videos is that it kind of seems like a task someone should have been able to automate by now. I’m sure it’s way easier said than done, and I directly polygon modeled my character this time anyway, but I can’t help thinking of doing more complicated models in the future where I might want to sculpt it first, and then spend all that time doing the retopo – sounds like a huge pain! But high detail art projects are usually tedious.

Latest version:

Attachments

guardian.blend (1.02 MB)

Well, I looked for the spiderman thread and I think it’s been closed on cgtalk forum, and all links are broken.

Hip topology looks much better, judging from your pic. I did a paint over of the changes I would make:



The orange lines show how I would have your edges flow in that area. The pink lines are loops I would add in. But at the same time, adding more loops may hurt you later on down the road.

Here’s the deal, the more edge loops you have in joint areas, the better you can get it to deform, but it will hurt you if you have to manually fix the weighting, costing you time there.

As to your random thought on re-topo, the blender cookie folks have created 2 addons for re-topo work, one was called ‘contours’ and I forget the other one’s name. You have to purchase them to use them, but they are aids aimed at speeding up re-topo work. 3D coat, a commercial sculpting software I guess is decent at auto re-topo, but I’ve never used it.

Randy

All I can think about as I watch all these retopology videos is that it kind of seems like a task someone should have been able to automate by now. I’m sure it’s way easier said than done, and I directly polygon modeled my character this time anyway, but I can’t help thinking of doing more complicated models in the future where I might want to sculpt it first, and then spend all that time doing the retopo – sounds like a huge pain! But high detail art projects are usually tedious.

There is always the issue of Automation v.s. Control…

the more you Automate the less control you have … People have different skills…
I personally like modeling so I use as few automation tools as possible when doing it…

also for me … I had to learn Rigging from the ground up… an so I am much more comfortable with building my own Riggs from scratch than most people are… I like the control I get when I start a Rigg from zero…

plus what most people dont’ really realize is once you build one Character Rigg you and reuse that Rigg for many different characters with minor adjustments…

I stay away from auto-riggs… Like the Blender Meta-Rigg/Riggify system… simply because I already know what I need to know to build an equivalent… where as most people will tend to use the Riggify system to save time… that works for them… but for me I give up some control I’m not quire ready to give up…

there are other areas of Blender I would rather automate and not worry about… were as someone else might want the control…

I do very little compositing work… I know little to nothing about separating render layers for instance… and My work probably shows it…

3D is such a large field you have to start specializing at some point on what you like best and concentrate on those things… and you have to be careful to not let others pass judgment on you for not knowing some portion of the big picture that is 3D … keep your shoulder to the wheel on what your working on now and keep after it until you get to where you want to be with that part… then you can branch off a bit and start learning the rest…

Thank you both! I will try out those changes to the edge flow.

Trying to import my animations to Unity was a huge headache. The model deforms completely differently, and the animations only play in the preview window and won’t play on the actual model. I’m thinking I will render my animations to PNG sequences with flat lighting, and then try to light those sprites in a similar way to how I do the other sprites in the game. Hopefully doesn’t look too weird – I’ll show you guys a demo once I get it working.

For my other sprites, I manually paint a normal map for the lighting to be able to fall properly on the object as if it was slightly 3D. Is there a way I can render copies of my animations as normal maps, where the closer to the camera a pixel is, the brighter it is? I was thinking I could attach a lamp to the camera, but then it would cast shadows on the arms, right? Unless I uncheck ‘cast shadows’ maybe, but can you guys think of any better ways off the top of your heads?

again… hop over to the game forums and see is anyone has any idea why your animations are transferring over to Unity properly…

here’s a guy going over some of this very thing…

me not being a game building guy (maybe I should try making some) I don’t know if he knows what he is talking about or not… but he seems too…

I’ve watched tons of videos on bringing Blender animations into Unity, but I don’t think it’s really worth it to keep troubleshooting at this point. I think rendering to PNG sequences will do fine and might even make the main character fit in more with the rest of the game. Thanks though. I might run into more problems with my rig while I’m doing various animations and if so I’ll post again – thanks for all your help :slight_smile:

here’s another one that seems to be directly addressing getting blender animations into Unity…

Thanks, if I decide to try the direct 3D model into Unity again I’ll check it out. Edit: Took a look at that, and that’s exactly what I did to try to import the model. Really it’s fine that the model doesn’t work since the game is all 2D anyway! Having the model in blender still makes animations much easier.