HOME | DD

vasilnatalie In reply to SteelDollS [2016-09-05 15:37:23 +0000 UTC]

I haven't fixed it yet.  I have other issues still, and I'm not sure that the amount of torso bones I've built into my latest version is enough to justify worrying about it.  I'll wait until I have a bad animation.  However, the last technique provided in vasilnatalie.deviantart.com/ar… should allow me to collapse the average rotation from LB and UB onto center, then only apply the correct difference to LB and UB.  Should, because I never know for sure until I implement it....

One of the ways to quickly modify motion is to use multiply of bone frame position angle to adjust "levels"-- levels, because the math is basically the same as it is for adjusting levels in an image editor.  You add a value, then you multiply the whole thing.  However, this relies on bones being measured from a parented bone from which to measure their effective angle.  As an example, let's say that I want to exaggerate negative Y axis motion.  I want subtle effects to be exaggerated, but I don't want extreme effects exaggerated.  In order to affect the torso above, I would do something like subtract 9, then multiply by 0.9.  If I do this on the bones as animated, I end up with 56.4Y/75.6, for a net twist of 19.2  But if the redundant rotation is instead applied at the center, I end up with the very different 17.7Y/1.5Y, for a net twist of 16.2.

With regards to rotation, look at what happens to the above situation when you decide that the twist difference between the UB and LB is too great.  One of the first things you might do is reset the UB.  However, doing so snaps it to 0Y!  If 21.3 degrees of difference is too much for you, wait until you see 75 degrees of difference!

Your eye issue may be similar (or it may not be), because the eyes are another place where there's a strong possibility for redundant rotation, thanks to the ability to rotate the eyes with either the both eyes bone or with each individual eye bone.  If both of these techniques are used, you can end up with the same problem.  Think about if your both eyes bone was at -90Y and both of your individual eye bones were at 95Y.  You'd reset the both eye bone, expecting the eyes to snap to the forward position, and instead, they'd snap back into the skull!  And if you're animating with the both eyes bone, you're best off using just the both eyes bone, or only applying additional changes to a single eye, rather than both.

With eye bones, I'm not as concerned about it, because I'm in the habit of using alternate animation techniques.  I use IK eye bones parented to the both-eyes bone for compatibility with existing motions, but one of the first things I do in any motion is register these via OP to the same bone that my camera is following (an IK dummy bone, designed expressly for camera use), and delete all individual eye motion.  So I never use eye animation as its provided.  Once you have the bones set up, this is the best, fastest way I've found of animating eyes.  It probably sounds complicated.  It's complicated to make, but it's extremely easy to use.

👍: 0 ⏩: 1

SteelDollS In reply to vasilnatalie [2016-09-05 16:58:37 +0000 UTC]

This is very interesting! The way you discuss bone structure feels like you're coming from a background of editing in graphics, which gives a very different perspective, that I find very useful. I like the precision in explanation, and some of this is new to me, but makes sense.

Ah- with the tutorial you linked, I have to say that while it's also very interesting and useful, I don't agree with the comment that "direction" of bone structure doesn't matter. Parent and child bones create direction, and do indeed affect things significantly, in how they function, so I definitely do think the direction of bones matters.

I love the multiply bone frame function; it makes things significantly easier when adjusting certain things in a motion, because everything you select and apply it to is affected at once in the same percentages, instead of painstakingly doing an imperfect job by hand, which is like torture and looks crappy when you're done despite all the manual labor. It hadn't occurred to me to ask what controls made it possible, or that it was related directly to bone measurements, though it really should have, now that I think about it. I mean, you have to adjust them when using chibi models, for instance.

In the case with the strange eye motion glitch with the motion I mentioned, deleting all the motion data for the eye bones did not correct the problem. But I didn't think to look for a secondary control. The motion that had the issue has been used now- I just made do and manually corrected what I noticed- and I probably won't go back to it for some time, but at this point, I'm much more interested in learning about the specifics of the glitch and how to fix it in the future if I come across a similar issue again, than I am in fixing the glitch just to use the motion anymore.

Hahaha, I think it sounds more complicated than it should because of the abbreviations used. It took me a while to track what LB and UB meant (lower and upper body). When you're not on the same page in regards to abbreviations, it's easy to lose your audience to them just assuming that you're talking over their heads- when usually, I think it's really not the case at all.

I think bone structure is indeed complicated to make, because there's a lot of different aspects you need to keep in mind to make things function the way you're aiming for. Adding the bones is easy. Figuring out how you need things to connect, why they need to connect that way, and keeping all the precise data required in an organized fashion can be a headache and a half to try to figure out and keep straight. But I think it's awesome to see you're putting your brainpower into this sort of puzzle and figuring out what causes what, when, and why. It's refreshing, and since I'm a beginner to this (model editing and PMDe), certainly sparks my interest to learn, too.

👍: 0 ⏩: 1

vasilnatalie In reply to SteelDollS [2016-09-05 17:29:42 +0000 UTC]

With regards to direction, you may be hearing more than I mean to say.  But there's a very easy, useful experiment that you can do to see exactly what I mean.  Take a bone, any bone, although a leg is a good one.  Change the bone it points to.  To anything.  Point at the head.  Or point at an offset, or -1.  Now load a motion on to your edited model.  Do you notice any differences?  I don't.  That's what I mean: that line, that you think is your bone, isn't your bone.  Your bone is just the point.  Points don't have direction (although they have orientation, which is what bone rotation is measuring, and which is identical for every unposed bone, barring use of "local").  So X, Y, Z-- all of these axes work the same.  It doesn't matter that your bone's line is aligned with the X axis or the Y axis, because that line doesn't actually mean anything.  But!  That is extremely hard to wrap one's head around.  I think I'm good at this stuff, working on it for a while, and it's still hard for me to wrap my head around it.  Which is why I provided the alternate "twist" measurement that's easier to visualize.

Sorry for the abbreviations.  A lot of the time, I'm trying to talk about complicated stuff, as quickly as possible, discarding as much as I can.  That's how I'd like to read it, so that's how I write it.  I know that it's going to sail over some people's heads.  I can accept that.  I'm not interested in followers or watchers or likes, none of that matters, and if what I say isn't useful to some particular person, that's fine with me.  But I also know that there are a lot of very young, very bright people involved in MMD.  One of the advantages young people tend to have with regards to learning is that they don't stop when they don't understand something, they just steamroll past it.  This isn't really relevant to abbreviations, but for some of words that I use that are more on the jargon side of things, I think using them helps these people learn those words.  I mean, that's how we learn language, by hearing or reading it, not by using a dictionary.  And knowing these words is really useful for anyone that wants to read more.  For example, if you try and learn about modelling by Googling "points," you're not going to get anywhere.  You have to Google "vertices."

But if there's anything you ever don't understand, don't hesitate to ask!  I can't know how much my audience already knows.  But I'm always happy to get into more depth whenever anybody wants.

👍: 0 ⏩: 1

SteelDollS In reply to vasilnatalie [2016-09-06 02:44:02 +0000 UTC]

Thanks so much! You are so friendly and I really appreciate that! ^^ I feel like I could definitely learn a lot from your experience, if I could just figure out the right questions to ask. Which is the hard part, when you don't already know what to ask, since you don't know yet what you don't know, haha x'D

Honestly, I often feel like it's more frustrating to learn this software than necessary, simply because most tutorials I find are often missing critical pieces that would make the puzzle be "explained" by reason. Maybe it's because I want to understand WHY something works and not just, "click on the button that looks like this picture, and do the thing, and you're done." But you don't get told what the button means, what it stands for, or the specifics of what's actually happening or being affected when you do the thing you're being instructed on, it can be difficult to form a clear understanding.

It's probably one reason there aren't a lot more editors and model creators for MMD; certain critical and basic things are just missing in the education that's available, and there's no official manual for the software to help explain parts that aren't able to be figured out by poking things. It's my firm belief that if you don't know WHY something is messing up, and WHY a certain way of fixing it works, it's a completely inefficient method of learning problem-solving skills in the program you are utilizing. 3D skill is not something that is ingrained in most human beings: it requires more than just creative thought to problem solve it.. It involves a comprehension of why things work.

I see what you mean about the bones- I didn't re-point them except in my imagination, but I know what you mean now. I agree that ending points do not effect control over the vertices rigged to a specific bone, but I still definitely cannot agree that where a bone ends/points doesn't matter. I think it matters quite a lot, as it affects everything that's attached to it below the end point.

For instance of explaining my thoughts there, imagine you have a model's hand you need to rig. The hand has bones for the fingers, and each finger bone attaches to the one after it. The direction matters, because if you have the tip of the pinkie finger be the parent bone, and the middle bone of the pinkie be the child bone, you're going to have major issues with making motions function in a human-like way, simply because human bone structures do not work that way. Also, unless you put the tip of the pinkie's bone in this example's parent bone as the center bone, the finger will not be movable with the model as a unit at all, because it is not connected to the main bone.

The same thing with any bone that ends with an attachment to another bone; granted, it doesn't really matter if the final bone in a chain (which we usually make invisible since it is merely an endpoint and not a connecting point, so it doesn't affect function of motions in MMD) goes to something else, since it's essentially the end link in the chain... but it does matter that it points from the beginning point to the end point, and is connected to the rest of the structure. Just like branches on a tree.

Hopefully I explained that thought in a not-too-confusing way; sometimes I use too many words to describe what I'm thinking of and I am not clear enough. But that's what I mean when I say I think bone direction matters greatly, if you follow me.

👍: 0 ⏩: 1

vasilnatalie In reply to SteelDollS [2016-09-06 03:20:35 +0000 UTC]

I follow you!  Parents definitely matter.

I also agree regarding it being difficult to learn stuff.  I felt the same way.  I found lots of tutorials that amounted to, "Hit this button.  I don't know why it works but it does."  Which never helped me very much.  But people were just trying to share as much as they understood.  They're all in the same boat as you or me, trying hard to figure out a foreign program involving a complicated subject.

What helped me personally was lots and lots of experimentation.  I try to share what I figure out. You might like my tutorials at learnmmd.com/http:/learnmmd.co… , which is where I put things aimed at a more general audience.  Although, still, I get complicated sometimes.  I try to explain why as well as what.

Anyhow, I hope I can help you in some way.  Don't worry about what questions to ask-- just ask what you want to understand, to get the next thing done.  That's the best way to learn.  So send a pm, or post on my profile, or at LearnMMD, anytime you want to understand something better, and maybe I'll be able to help.

👍: 0 ⏩: 0

vasilnatalie In reply to Hogarth-MMD [2016-09-03 02:28:19 +0000 UTC]

No.  They didn't create new models, with new bones to control the distribution of UB and LB motion along the torso.  That's the significance of the right side of the image.  "...In order to have more beautifully animated torsos, with existing animations."

👍: 0 ⏩: 1

Hogarth-MMD In reply to vasilnatalie [2016-09-03 06:57:04 +0000 UTC]

Okay, I apologize for being slow to understand this issue. So the one and only 3d model which is having really ugly torso mesh deformations is a 3d model which you have created "to have more beautifully animated torsos".  Is that right?

👍: 0 ⏩: 1

vasilnatalie In reply to Hogarth-MMD [2016-09-03 14:56:27 +0000 UTC]

Yes.  Look at the exact reasons I give in that picture.  There are three of them, although yes, forward compatibility is a bigger issue than the other two.

You want to argue about everything, and arguing with you takes more time than making something would, and is much more stressful.  It feels like you're just waiting get some kind of "Aha! So you were wrong!" on me.  Stop it.  Go make something.

👍: 0 ⏩: 0

vasilnatalie In reply to Krysel1992 [2016-09-03 02:30:02 +0000 UTC]

I probably ought to try some different animation programs.  MMD is nice, in that I was able to get going in a matter of minutes.

But I have a feeling that I wouldn't bother creating any interesting structures for something like Blender.  I'd just go straight to bone scripting.

👍: 0 ⏩: 0