PDA

View Full Version : Rigging questions



Siber
10-28-2009, 11:27 AM
I'm taking a 3d modeling an animation class that uses Lightwave, and I've run into some issues where it's not behaving how I expect it to based on what I've been taught so far.

My issue is mostly with weightless. I was led to believe that if you used weightmaps, the areas assigned fully and exclusively to one bone would exactly match that bone's movements, and no other bone's. My model has weightmapped regions that aren't following their bone's movements exactly, and are being affected by other bone's movements.

Are there any common pitfalls I may be encountering, or am I suffering from a fundamental misunderstanding that needs to be corrected?

Additionally, my model has several identical limbs in multiple positions. Is it in any way possible to set up the weights, bones, constraints, and so forth for just one limb and copy them to the other limbs, or is my professor correct in his statements that every one must be done individually?

h2oStudios
10-28-2009, 12:44 PM
Um, let's see. I believe that if you want one bone to exclusively control certain points/vertices those vertices must have %100 of the weight applied to them. For instance if you have an upper arm bone & a lower arm bone in an arm(of course), and let's say that 75% of the upper arm geometry has %100 of the upper arm weightmap and the rest has %0 then it's liable to be influenced by other bones. I'll post an example. As for setting up multiple weights, I'd say do one half of your entire skeleton then mirror your weights with a Mirror Weights Plug.

h2oStudios
10-28-2009, 01:03 PM
Weights

One arm - Upper %100, Lower %100.

One arm - Upper -> %100, 75, 50, 25 - 25, 50, 75, 100% <- Lower.

If any point is not assigned to a weight then it's open to deformation from any bone.

78786

And as for the rest, mirror stuff when rigging.

SplineGod
10-28-2009, 02:10 PM
I tend to avoid using weight maps or use them as little as possible because the tend add another layer of complexity.
Bones work without the need for weight maps. Weight maps modify the bones natural influence. I generally add the bones, rest them and see how they work. I ch eck for good and bad deformations then I determine if extra hold bones are the way to solve the bad deformations or going with weight maps.

Siber
10-28-2009, 02:21 PM
h2o: I don't have lightwave at home, so I can't examine your example, but I will when I can. Your explanation sounds like how I thought things work, but not how things seem to be actually working. But perhaps I screwed up at some stage and didn't notice.

As for the mirroring, I'm afraid that won't be quite exactly sufficient(And my professor told us mirroring constraints and such is impossible.) A picture of my model might be helpful.

http://i262.photobucket.com/albums/ii112/EatThePath/ClassProject.png

Six identical limbs, arrayed in two sets of three at either end of the creature. The limbs are functionally identical, so should have the same constraints and so forth. I'm just not looking forward to setting up the same structure six times.

Incidentally, to give some idea what I'm trying to achieve with my weights, the purple suctions are intended to be rigid and inflexible, while the flesh-toned areas are to be flexible.

SplineGod: That sounds reasonable, but the professor wants us to use weight maps, so avoiding them would be problematic.

h2oStudios
10-28-2009, 02:22 PM
Right, Larry. But now I have a question:D. What would slow down Open GL Peformance more significantly when operating a rig - The presence of Weightmaps or extra Hold Bones? Would be nice to know.

SplineGod
10-28-2009, 02:31 PM
Neither slow down Opengl. Both can slow down the overall performance when calculating deformations. Weight and Morph maps will slow things down esp with denser meshes.

The way around that is to not combine animating with deforming. Deal with them separately. When animating you dont want to deal with deformations so that your rig is fast and responsive. You can do that by setting the bounding box threshold lower and/or using lower rez standin geometry. I can also easily temporarily turn off the extra hold bones etc. You can also have separate rigs for animating and the full blown one for deforming and copy the motions from one to the other.
Deformations should be tweaked to work with the animation thats been done.

Siber,
The purpple areas should have 100% weight while the flesh colored areas should have that same weight falling off. The flesh area should also have overlapping weight maps from the other purple area that borders it.
I like to use a free plugin called Weight Map Blur that works nicely for those flesh colored areas.

h2oStudios
10-28-2009, 02:35 PM
I see exactly what you are trying to do now. This would be done most likely using %100 Weights on each control bone (which is pretty easy compared to blended weights). If I were you I'd theoretically look at this object with the arms parallel to each other as being side Right & side Left, place the bones in there appropriate places on say side Right and name them something like R_Arm_1_Array_01,02,03, etc. R_Arm_2_Array_01,02,etc. Then when you're done with side Right, mirror those bones to side Left and rename them with L_ , then apply Bone Weights(Map> Weight> Bone Weights _ Clip 0m. Now you'll have Weights for all the bones with a % of 0. Now start selecting the geometry on the object for each corresponding bone and apply Weight Values of %100. I hope this makes any sense, lol. If not when you get a chance post your object and I'll be more than glad to show you what I mean. phew.

EDIT: oh and yeah like SG said blurring or blending weights on the flesh toned areas is 'probably' ( but not 'definitely' :D ) going to need to occur

RebelHill
10-28-2009, 03:00 PM
Easiest way to think of the weight issue is how LW compares to other packages...

In other packages, give a bone to a mesh, and it has no effect, you need to weight certain vertices to said bone to tell ti what to influence... so to that extent, the weights could be described as additive.

LW does things the exact opposite, add a bone to a mesh, and it will affect ALL vertices of that mesh, giving the bone a weight, will allow it to ONLY effect the verts with that weight, so it can be thought of as a subtractive approach to influence.

As for the cloning/mirroring of hierarchies, including constraints, targets, goals, etc... its perfectly possible, but some of the newer additions in the most recent version wont clone over, but once a single limb is setup, its a very quick process to just it X times, and reapply the non cloned constraint areas...

The tuts in my sig cover all of these issues in great detail, weigting, constraining, cloning and reusing setups, etc, etc...

sounds like ur prof could use em too (;

Rove
10-28-2009, 04:18 PM
I tend to avoid using weight maps or use them as little as possible because the tend add another layer of complexity.
Bones work without the need for weight maps. Weight maps modify the bones natural influence. I generally add the bones, rest them and see how they work. I ch eck for good and bad deformations then I determine if extra hold bones are the way to solve the bad deformations or going with weight maps.

I never looked at it this way Larry. I've recently read Jonny Gordon's books Cartoon Character Creation where he goes through setting up cartoon characters using weightmaps amongst other things. I'm in the process of designing my first character ever. I'll give your technique a try as well as weightmapping and see what fits me best.

h2oStudios
10-28-2009, 04:33 PM
I think it's really a matter of preference & control when it comes to Weight or not to Weight. I've always used weights in the past and now am exploring different things. Let's hope Siber has grasped his issue, and that his proffessor will get RebelHill's Rigging Tutorials, lol ;).

SplineGod
10-28-2009, 04:39 PM
Since LW bones ARE deformers already and weight maps are only modifying that influence then it makes sense to make sure that the bones are doing their job as good as possible before adding weight maps to the mix. Many times its easier to model the character in a pose to help reduce the possibility of getting cross influence from bones (which is generally why many resort to using weight maps). Its easier in my experience to predict hwo bones will effect a mesh then to simply jump into weight maps immediately.
If I do need weight maps I can generally keep them quite simple. For example I may have one weight map for the arms and legs and another for the body. Multiple bones can share the same weight map.
Another reason why I like hold bones is that they can be animated unlike weight maps. This gives another level of control as
in the following examples:
http://www.vfxcast.com/media/326/Simple_Muscle_Bones_pt1/
http://www.vfxcast.com/media/335/Simple_Muscle_Bones_pt2/
BTW these are things Ive always covered in great detail in my Character Course as well as other DVDs I have on KURV. :)

Siber
10-28-2009, 04:58 PM
If anyone is curious, the last version of my model is up at http://dl.getdropbox.com/u/896078/Triskellion_v006.lwo

As to weights vs not weights, or getting the professor tutorials, I've generally found that trying to convince the teacher of a course that they're wrong about something doesn't end well. I'll see how things go, though.

SplineGod
10-28-2009, 05:47 PM
Thats understandable. The best way to convince anyone is to show them.
Another aspect of rigging that isnt touched upon as much as it should be as it should
is that fact that modeling is also a big aspect of rigging.
Ive seen many deformation problems that are solved by improving the flow
of the models in particular areas. If a model has bad flow no amount of weight mapping
etc will solve it. Geometry can only bend or fold along edges.
This is something that effects not only how bones or weight maps influence a mesh but also in the creation of endomorphs etc. This aspect, especially with regards to flow is something I also cover heavily in my tutorials.

h2oStudios
10-28-2009, 05:53 PM
Okay, I see what the problem is here. You've used Weights for your object but have areas where there are bones with no weights applied like those finger thingies. You'll have to weight map those fingers aswell. What happens is when using weights EVERY bone must have a weight even if it's value is zero unless their Strength is set to 0, becasue as said before bones in LW are deformers by themselves. but I sure you want those fingers to be controlled, so give those geometries weights and you should be fine with those, but there are still some issues with other things in this rig. I'll check into it further later tonite. Hop that helped, ;)

SplineGod
10-28-2009, 06:09 PM
Thats an important concept with weights:
Every vertex should have a weight map and every bone should be assigned to a weight map.
Another thing to remember is that a vertex with no weight map is not the same as a vertex with a weightmap or zero strength.

Each bone in LW has an effect on every vertex in a mesh(before weight maps).
This influence is controlled by the falloff , bone strength, a bones rest length, orientation etc. If you use the Faster Bones setting then each vertex is only effected by the 4 closest bones. That can help speed things up but can also give better or worse results with deformations. All this means is that any vertices NOT assigned to a weight map can be influenced by any bones NOT assigned to a weight map.

Remember that multple bones can be assigned to a single weight map. You could have one weight map or the hands and all the hand bones share it.

Theres a couple of ways to check to make sure all your vertices have a weight map assigned:
1. Click on a point in modeler and hit the i key to get information on that point.
2. In modeler select the move tool. Hit the n key and change the falloff to weight map.
Select a weight map and use the move tool. If any points are left behind then those points are not assigned to that weight map. Use Undo to move the geometry back.

Siber
11-04-2009, 03:05 PM
Thanks for all the help so far, it has been a great boon. If I may impose further, could someone please walk my through the steps for copying the rigging from one limb to another, and/or mirroring it? I've looked for myself but haven't found anything yet. Pointing me at a (hopefully free) tutorial that covers it would work just fine as well.

SplineGod
11-04-2009, 04:16 PM
You can mirror bones under the setup tab in layout. I believe it has the option to also mirror goals etc.

Siber
11-04-2009, 07:06 PM
Simple mirroring isn't quite enough. As shown previously, the model has six limbs, three on each end of a central trunk. Ideally I'd be able to bone, constrain, and goal one limb, and copy all that work over onto all five other limbs. I have worked this way when doing the mesh, but can't see a simple way to do it for rigging past the skelegon phase.

SplineGod
11-04-2009, 09:52 PM
My first suggestion is to not use skelegons.
If you do it in layout you have a l ot more options and
a lot less gotchas.

Siber
11-04-2009, 10:22 PM
Oddly enough, I was taught it was the other way around.

EDIT: Something more concrete, I have in the tutorials I have watched so far and here in this thread heard of "resting" bones. This hasn't been mentioned at all in class and seems quite important. Could someone explain it?

evenflcw
11-04-2009, 11:50 PM
Resting means selecting one or more bones and executing the Record Bone Rest Position command. Put simply (and somewhat incorrectly) this command binds the mesh vertices to the selected bones. Each vertex will be bound the strongest to the bone closest to it. In other words, the command just tells LW to calculate the influence each bone should have on each vertex given the bones current positions, thus "rest position" (aka bind pose; the process is commonly called skinning, perhaps because you're sortof wrapping the skin over the bones).

Just try it. Load a mesh. Create two bones inside the mesh. Rest both of them. Move the child bone and observe the deformation. Now rerest that child bone somewhere outside the mesh. (Just move it there and rerest; the mesh will snap back into place, ie go back to it's rest state). Now when you move the child bone it will not deform the mesh as much, because it was not as close to the vertices as last time. Move it back within the mesh and rerest and you're good to go again.

SplineGod
11-05-2009, 02:12 AM
Siber,
Having use both skelegons and bones in production work Ive learn to do as much to simplify the workflow as much as I can.
Skelegons force back and forth between modeler and layout to tweak things.
If you dont watch for the gotchas with skelegons youll spend a great deal of time fixing the issues in modeler or layout which again isnt efficient.
The orientation of bones is alot less of an issue when drawing bones in layout and thats where all the tuning happens anyways. You really cant see how the placement of skelegons etc will effect things like you can in layout just as evenflw mentioned. You can also adjust the rest length of bones after theyve been rested to see how they will grab or release vertices etc...again not possible with skelegons.
Also if skelegons are not welded to other skelegons then they are not parented, which is another gotcha.
As Megalodon pointed out it sometimes boils down to personal preference but for me I try and do as much as I can to reduce the complexity of my workflow when possible. This is especially true when it comes to rigging.

jasonwestmas
11-06-2009, 07:54 AM
The setup/rigging tools in 9.6 are much improved, take advantage of them. You can't always do that when using skelegons. When you get comfortable with bones, you should try Joints but I guess Joints might be adding things you don't necessarily need. I just enjoyed setting up the IK better when using joints, it was easier. Of course the new snapping feature helped for adding in controllers a lot!

Siber
11-11-2009, 06:17 PM
Rigging is due tomorrow, and I'm closing in on having one limb rigged completely, hoping to get everything else done in a big push tomorrow, a matter of pretty much duplicating the work over on every other limb. I'll put a link at the end of this post, if anyone wants to take a look at it and see if I've commited any major sins that need fixing, I'd greatly appreciate it. Your comments and advice have already been a great help in showing me what is possible.

If not, I'll hardly be offended. I think I'm on the right track, even if it's a bit of a slog ahead of me.

http://dl.dropbox.com/u/896078/triskellion%20s_v004.lws has the scene file, not sure if that's necessary or not, and http://dl.dropbox.com/u/896078/Triskellion_v025.lwo is the object.

h2oStudios
11-11-2009, 10:37 PM
Oh dear.

h2oStudios
11-11-2009, 10:37 PM
what time tomorrow/today?

evenflcw
11-11-2009, 10:50 PM
Hot Tip: When you upload a rig for review, animate the controllers so we know what should be moved and what should be left alone, what is already setup and what is not. It's not always apparent as everybody rigs differently. Not all nulls are always controllers.

Biggest boo boos!? You turned on IK everywhere. On every bone and every channel; You didn't terminate the IK chain properly; And you tried to end or restrict IK by setting extremely tight limits.

An IK chain starts at the pivot of the item where you set the Goal Item (for the leg this would be the ankle or foot bone), and ends at the parent where you check 'Unaffected by IK of descendants' (for the leg this would be the parent to the thigh bone, ie somekind of hip bone). All items inbetween these two have the potential to be animated by IK (the thigh and shin bone). To do that, you set one or more of their controllers to 'Inverse Kinematics', as you have done. However, you do not have to turn on IK for all these inbetween items on all channels. Leaving some on 'Key frame' will not break the IK chain, but rather just make those joints completely (all channels on keyframe) or partially rigid (some channel on keyframe).

You also should never turn on IK on all rotational channels because that gives the solver too much to work with and it ends up finding multiple solutions for every pose. An LW won't know which you prefer. So you get unstable jittery motion as LW jumps from one to the other. You're also forsaking any chance for control, leaving everything up to the solver. It's not the computer that is supposed to decide how your character moves. It's you. So pick at most two channels that should be controlled by IK. Commonly it's H and P, leaving B for keyframing a twisting motion. But other combos aren't unthinkable. Just not all three. Often it's just one. (for the leg, the knee doesn't bend in all 6 directions, just 2 (pitch+ and pitch-), so leave the others on keyframe).

Instead of setting limits to min=0, max=0, just set that channel to Keyframe. As explained above, the IK chain won't break or end until it finds 'Unaffected by IK of descendants' on an item.

PS. What school is this and did your teacher go through basic IK setup before giving you the assignment?

Siber
11-11-2009, 10:51 PM
due pretty much 3pm eastern US time. I'm going to make one last check around 11 am here to see if there are any bombshells, and probably brush up on a few key tut vids, and then will sequester myself away in a tiny computer lab with naught but the hum of fans as my company.

Though really, just having even one fully functional limb is probably sufficient for a pass, which I pretty much have, so. I'm not worried.

EDIT:
evenflcw: Thank you, those were factors I was unaware of. It seems like in some cases, for example the leftcore->leftouter joint in that model, rotation in all channels is desirable or required for proper motion. Is this a hard and fast rule that this can never work, and what would you do in a situation where you need, or think you need, that kind of motion?

evenflcw
11-11-2009, 11:33 PM
It seems like in some cases, for example the leftcore->leftouter joint in that model, rotation in all channels is desirable or required for proper motion. Is this a hard and fast rule that this can never work, and what would you do in a situation where you need, or think you need, that kind of motion?

I honestly can't say for sure that rotation on all channels is desirable because I can't figure out the nature of this beast. :) But for a joint such as leftcore->leftouter it is common to set H and P to IK, and B to FK. B can still be animated by simple keyframing. FK or keyframing is the basis for animation, not IK. IK is just assistance.

If you put all channels into the IK mix, you really can't predict when it will bank rather than change the pitch or heading, because you give it free reign. The rig need to be predictable to be controllable. By taking one channel out of the mix (commonly bank) you restrict the number of options and make it predictable. If the goal goes right or left it can only change heading, up or down, only the pitch. If bank is still in the mix it might start banking and twisting at anytime, when you'd rather prefer it to pitch or change heading, or the other way around. But if you don't mind your creature twisting when it should be swinging, dance on!

h2oStudios
11-12-2009, 12:42 AM
Here's somethin i came up with with your stuff, in a quicky mind u. May help in the next hours (if your brain hasn't exploded by now) before class time. -

79194

I hope that's the right attachment.:)

Ps- there's some bone orientation issues that I went through towrds the hand thingy, there may be more, dunno. simple targets andsuhc and a... er... johnny 5, stay alive.

h2oStudios
11-12-2009, 12:46 AM
PS- there's some IK issues with that, might be solved with some poles? You've got breaking rotations so it's a bit diff from CA rigging. "No disascemble Johnny Five!" (SP?)

Edit: I targeted something to pole, when I shoudn't have. Ima dunce

Dodgy
11-12-2009, 12:50 AM
One thing to bear in mind is the new Motion panel is multiselect aware. This means you can select, say, all the elbows and set their IK channels and limits at the same time (if the values are to be the same). You can even select an IK goal for multiple items.

evenflcw
11-12-2009, 01:31 AM
Right!
AND you can RMB click an item to select it and all it's children (must by same item type)!

h2oStudios
11-12-2009, 05:51 PM
Just curious as to how this turned out, Siber?

evenflcw
11-12-2009, 09:35 PM
Probably got kicked out of school for telling the teacher how IK works. ;)

h2oStudios
11-12-2009, 10:21 PM
Probably got kicked out of school for telling the teacher how IK works. ;)

LMFAO!!! LOL! oh man, that would be something... Wait, what if he really did? ;)

Siber
11-13-2009, 10:21 AM
Well, the professor wasn't able to open the lab for us as early as he'd indicated he would, and ultimately decided nobody was ready for the deadline, so he moved it back. Before I left the lab I had all the legs rigged and targeted, and even played around with moving it into a standing animation. Things are a bit hinkey still, I think I need to play with the 'hip' constraints a lot and look into some other variations. That's the biggest problem with this model, I have no living example to work off from in how the joints need to move. The advice here, particularly setting the constraints of multiple bones at once, but also in having at most two channels of IK only, helped clear up a lot of issues and speed things along nicely. I still have to set up the hands, too.

Some questions on that front: From some of the tutorials I've watched it seems possible to have one null for each leg, then parented to (or following, I'm not sure the proper terminology) that null, I would have another null controlling the hoof, and another one controlling the hand as a whole, then following/parented to the hand null there would be a null for each finger In this way you could have control over the whole limb without having to move seven handles each time a leg moved, but you would also have control of individual fingers and the independant movement of the hoof and hand. Is this reasonable? Are there any pitfalls and gotchyas here?

Unrelatedly the professor recommends/demands that 'keep goal in range' always be checked for the final bone in the IK chain that is attached to the null, as well as match goal orientation. Keeping the goal in range seems to only make the handles harder to control and put where I want them. Opinions?

Tangentally, is there a way to save selection sets and recall them later? So I could have a fingerbones selection that can be recalled at any give time, or whatnot. I've looked, but I haven't found anything.

jasonwestmas
11-13-2009, 11:32 AM
I would just use nulls for moving the arms and legs IK. I also only use IK in legs and only sometimes in the arms. I would use spline IK in the spine but that's not easy to do without 3rd party plugins. You can create selection sets via the "classic" scene editor but you have to go back to the scene editor every time you want to activate the selection set. Use additive rotations for the spine, fingers and other bone chains using selection sets. Then tweak the rotations one bone at a time for fine details in your motions.

You can use the arrow keys to move up and down the bone chains and I like to use the "schematic view/viewport" to select bones that don't have a selection set.

Another tip is to keyframe your neutral "T" pose on a frame before your first frame in your animation. So if you ever want those neutral keys again for whatever the reason you can get them back using the graph editor.

RebelHill
11-13-2009, 11:32 AM
Yeah, keep goal in range isnt much help, except in a few very rare situations, and MGO isnt always necessary either, it depends on the particular needs of what you're setting up wether you want to use it or not.

Opinion... your profs not really teaching u much that sounds particularly useful, lol.

jasonwestmas
11-13-2009, 11:40 AM
Yeah, keep goal in range isnt much help, except in a few very rare situations, and MGO isnt always necessary either, it depends on the particular needs of what you're setting up wether you want to use it or not.

Opinion... your profs not really teaching u much that sounds particularly useful, lol.

Could be a Max 4 guy :>

Siber
11-13-2009, 11:51 AM
He seems to be a very large Lightwave fan, and has been using it for years upon years.

Andrewstopheles
11-13-2009, 06:48 PM
I think you would be well advised to follow your instructor's teachings. Respect his position. He is the one who will be judging your work and if you don't follow the workflow he is trying to teach you your grades will likely suffer.
I have encountered this problem in the past where my instructor seemed to be leading me down the wrong path. Sometimes the instructor was wrong, sometimes he was right, but in either case going through the experience with him is the best training you're going to get IMO.
After all that is said, and when you're done learning from this instructor, go back and do things YOUR way, using the advice you've received here. These guys are some pretty experienced fellows and know what they are talking about. But don't let that turn you against your instructor - he probably knows what he is doing as well. There are more ways then one to skin a cat.

(Probably the wrong forum to be talking about skinning a cat! hehe):D

jasonwestmas
11-13-2009, 07:14 PM
I'm a presumptuous person when it comes to personal learning needs and creative style. I never enter contests or stick to a single style of graphic, illustration or genre. I would be weary of instructors that teach you only one prescribed way of doing things. If you know what you need to learn, screw the grades. All I'm saying is that ultimately you should be looking out for yourself and try to go beyond what a school curriculum thinks you ought to know and how to learn it. I just know from experience that I learned 3D the best by really studying how my mind works and keeping account of what I already know and capitalize on my strengths. No teacher can do that for you and all too often your mind doesn't work the same as others who teach it to you; not to mention others can have the tendency to slow you down if you aren't careful. Of course if you have a scholarship on the line. . .that may require some compromise. I personally would not find myself in that kind of situation where I'm a total slave to someone's expectations, holding me to the so-called grade.

SplineGod
11-13-2009, 07:48 PM
i Think is possible to respect and instructor and not agree with them. The bottom line is that youre in school for YOUR education..
So in that regard I agree more with Jasons view.

Im also goiing to throw in something else. One of the problem with many instructors/tutorials/books etc is that they approach the rigging issue from a technical standpoint rather then a requirements standpoint first. IOW what makes a good rig a good rig. What are the things a good basic rig should be able to do. Then from there you go into the technical aspects of your software that allows those requirements to be accomplished.
Studying IK without those requirements is like trying to learn about home construction by studying hammers and saws first.

Siber
11-13-2009, 08:37 PM
Sadly, I care at least a little about my GPA. While being a professor in my view doesn't earn you automatic respect but rather sets a higher bar for it, it also puts you in a position to affect my future if I don't satisfy your whims and whimsy. Having a master plan that affects the current work but not revealing it, particularly when you don't check your student's progress closely and make sure they're on the right track, is just a recipie for disaster.

In any case, I'm trying to strike a balance, keeping to the requirements as closely as I can while still following the advice here, and some of the advice here has been quite simply crucial, as well as the tutorials linked. But on the subject of striking a balance... would the structure described in my last post, down to IKed fingers with nulls following nulls, work and be possible in lightwave? Even if not ideal?

jasonwestmas: The model in question, well... I suppose it could be said to have a spine, but nothing you would spline. Similarly there is no real distinction between arms and legs for it. It is in it's own version of a T pose, though. There is at least one picture earlier in the thread.

jasonwestmas
11-13-2009, 10:04 PM
Im also goiing to throw in something else. One of the problem with many instructors/tutorials/books etc is that they approach the rigging issue from a technical standpoint rather then a requirements standpoint first. IOW what makes a good rig a good rig. What are the things a good basic rig should be able to do.

That's a great point. Another example is painting and surfacing a model, I do this most of the time. So many times I had to stop myself and say is this level of detail acceptable for the 3D shot and camera in a certain context prescribed not by seeing how far I can take something but by what is needed in a shot breakdown and storyboard. Especially in video games where you can't always see those tiny details in a 1024 texture map, even in a 2048 texture things get lost. Then you have to always take into consideration where the camera will be pointing and zoomed in on most of the time during game play. If it won't be noticed by anyone, especially you when you line up your shots, don't go there. You can save yourself many of hours of work on literally a single model. This is a very subjective concept but one can be surprised what is not necessary to sell the shot.:hat: Same can be said for rigging of course.

SplineGod
11-13-2009, 10:05 PM
Siber,
The good news is that in all the years ive worked in this business Ive never once had anyone care whether or not an artist went to school or what their grades were.
They only cared about two things:
1. Could you do the work ie. a good demo reel.
2. Personal recommendations from others.
Good grades in no way insure that someone is good at this whic h is why everyone at some point wants to see a demo reel.

jasonwestmas
11-13-2009, 10:11 PM
jasonwestmas: The model in question, well... I suppose it could be said to have a spine, but nothing you would spline. Similarly there is no real distinction between arms and legs for it. It is in it's own version of a T pose, though. There is at least one picture earlier in the thread.

Yeah I was just giving you a loose example to base your mechanical looking creature on. I tend to base everything on human anatomy. Your model's rig should be fairly straight forward, don't over-complicate it, would be my advice.

h2oStudios
11-13-2009, 10:57 PM
... But on the subject of striking a balance... would the structure described in my last post, down to IKed fingers with nulls following nulls, work and be possible in lightwave? Even if not ideal?



Yes, of course it is. If your question is literal, it's just a matter of some item relationships thru parenting, same as item, contsraints, so on. Understanding how they work is where the majority of rigging is at:thumbsup. Post what you have so far, for disection. I'm actually doing alot, and I mean alot of rigging right now. So I'm in that vain.

Siber
11-13-2009, 11:09 PM
http://dl.dropbox.com/u/896078/triskellion%20s_v007.lws is the latest scene file. Nothing animated, I'd fix that but I don't have LW at home. I see mention of a time-limited demo on the main website, but can't find it if it actually exists, so. That's one reason I'm asking so many exploratory questions instead of just trying and finding out; my access and experimentation time is relatively limited.

Edit: Holy temporary-lapse-of-search-skills batman! Scratch that last. Whee!

Siber
11-13-2009, 11:57 PM
Woooah. Okay, uhm... how backwards compatible is 9.6 with previous 9 versions? and the other way around?

h2oStudios
11-14-2009, 12:59 AM
Woooah. Okay, uhm... how backwards compatible is 9.6 with previous 9 versions? and the other way around?

woah, did you just have an epithany or something, lol? I did something with your scene a couple posts back the night before the previously dated due date of your rig, don't know if you took a look (and quite honestly don't remember exactly what I rigged up with it, lol). I believe 9.6, atleast when I try, is backwards compatible with Objects back to (i think) v8.0, but Scene files only to 9.5 (though I'm not sure what criteria are involved in terms of whether or not certain feature usage plays a role).

I'll take a break later this day and mess around with your scene to try and show you how you might be able to achieve what u want.

Siber
11-14-2009, 01:11 AM
h2o: By the time I got to the lab I was a bit stressed out and forgot entirely to look at your file. I have done now, though, in the comfort of my home, and am intruiged by what I saw. I'm just worried that what I work on here won't transfer well, but I believe the school computers use 9.5.something, or at least 9.4, so I have some hope. We'll see tuesday, I suppose.

h2oStudios
11-14-2009, 01:33 AM
dude, honestly whatever i did in that file i don't think used anything exclusive to 9.6, probably stuff that works in 7.5 (& it was a quick roughy ;). Anyway, I'll be rigging all day tommorow/today, check back.

h2oStudios
11-14-2009, 07:58 PM
Okay, here's somethin for ya, Siber.

79292

I only rigged one leg for ya, *makes it easier) so as you can disect it a bit

9.6 scene file, only due to the fact that I used Pole Item on your Wrist Flex bone. But I could def make this thing backwards to 7.5 or whatever.

FIRST- I took the Skelegons that are i think from base of that vertical bone before the "thigh" (I Left the main Body Bones alone as I do not now what their function would be) and used EDIT Skelegons to orient there Bank Axis's in a way that their pitch face forward in relation to of their direction (you should definitely check out Using EDIT Skelegon Tool, as Bone Orientation is VERY Crucial, and it's way better IMHO to get that right in Modeler before shoot off to Layout and F-ing around to get proper Orientation. Once in Layout I Record Pivot on ALL Bones, Then I worked my Nulls in there. I placed contstraints on the IK bones JUST as place-holders to adjust later for whatever motions this thing would go through, but I find that in certain cases it's better to animate things according to their range without having to throw constraints into the mix unless neccessary.

No IK Fingers yet, but the concept of relationships is there in the Leg Rig and is pretty much the way u would do the fingers as well.

I think I used 4 Control Nulls and a Few Targets Parented to eachother in a way that gives u a basic IK/Reverse IK thingy-ma-bob.

So take a look, and fire away with any questions. It's quite simple once you begin to understand Relationships (not my wife is a nag, but the 3d kind:D, wait, I'm not married, anyway).

Cheers!

h2oStudios
11-14-2009, 08:25 PM
Wow, I am so oblivious that I forgot to look at your current file, Siber, and even terminate the IK chain on L1_Thigh in my example (not that it had any affect on the parent bones, but it's always good to make sure). Pretty good so far.:)

h2oStudios
11-14-2009, 08:57 PM
Now, check this out. Load the scene and move the L1_Hoof_Reverse. You'll note that L1_Hoof_Controller follows that. But if you wanted L1_Hoof_Conroller to be Controlling L1_Hoof_Reverse while it acts as a "Tip_Pivot" then - Clone L1_Hoof_Controller> rename it L1_Hoof_Goal> Parent it to L1_Hoof_Reverse, parent L1_Hoof_Reverse to L1_Hoof_Controller, then parent L1_WristFlex_Pole to L1_Hoof_Goal.:) Lastly make L1_Hoof Bone's Goal - L1_Hoof_Goal. I hope that I wrote that right:D.

Siber
11-15-2009, 10:26 AM
That's pretty cool, h2o. Very instructive, and I've had fun playing around with the rig. Seem to be able to do some nice things with it. Not sure it'd fly with the prof or not to be so clever, so I may have to make an in-class rig and an actually-having-fun rig. Either way, thanks a bunch!

h2oStudios
11-15-2009, 01:07 PM
That's pretty cool, h2o. Very instructive, and I've had fun playing around with the rig. Seem to be able to do some nice things with it. Not sure it'd fly with the prof or not to be so clever, so I may have to make an in-class rig and an actually-having-fun rig. Either way, thanks a bunch!

Hope it helped ya in figuring how you're gonna go about it man.:thumbsup:

Cheers!

Siber
11-17-2009, 10:02 PM
Some new developments! Firstly, I have pinned down the school's version as 9.3.1, and determined that even if I avoid adding things that are new to 9.6, it won't open files saved in 9.6. Unless there's a save compatibility mode I'm missing. Additionally, it is unacceptable to have some joint controllers on keyframe, everything must be on IK, even if the limits allow no movement. Dunknow why, but that's the requirement, so please ignore that quirk if you happen to look at any of my WIP downloads. It's not by choice.

So, the 9.6 trial is useless for actually making changes to the working model, but it is very useful for plotting out what I want to do next. I've tried to imitate some of the things I've seen on the rigs around here, but I'm not sure I've suceeded.

http://dl.dropbox.com/u/896078/Triskellion_v028.lwo http://dl.dropbox.com/u/896078/triskellion%20s_v014.lws

This seems to work okay, but in places takes more babysitting than I think it should. I think at this point it's entirely possible for me to produce the results I want, but I'm sure there are approaches that would eliminate some of the work in making sure things don't wig out when pushed. The problems are most notable if you just move the leg target up the length of the leg and watch the fingers. New rigging is only in place on one limb.

There's more I wanted to say, but I'm tired. I'll say it later if i remeber.

h2oStudios
11-18-2009, 11:46 AM
Glad to see your pushing through on your assignment. IK on all limbs? That is a bit wierd, but hey, that's he wants, so that's what you have to give him I guess. I'll gladly take a look at your attachments later on tonight.

PS- you can in fact save out from 9.6 to an earlier build, and providing that you don't use anything in scene that's exclusive to 9.5/9.6 & it should load fine, even in like 7.5.

CHeers!

Siber
11-18-2009, 12:02 PM
H2o: I tried to avoid any new features, and it simply stated "this file requires a newer version of lightwave". Do I have to save it in a special way?

h2oStudios
11-18-2009, 01:29 PM
The files I showed you or your own files?

There was and should be by default a File> Save as> 9.0, 6.0, 5.5 scene kind of options in the interface, but I think you might have to dig it up in the EDIT Menu Layout

h2oStudios
11-18-2009, 01:34 PM
Ah, no doubt I just had to do that myself because my menu setup got flushed before so, here's how ya do it.

In 9.6-

Go to Edit> Edit menu layout>, and in the Command list (on the left) there should be a file sub-category called "File" and you should find those different save scene version option thingies, just drag them over to your desired menu and viola. Now you can save scenes that can be opened in previous Versions going back to like Version 4? :eek: lol.

Siber
11-24-2009, 09:09 AM
This seems to work okay, but in places takes more babysitting than I think it should. I think at this point it's entirely possible for me to produce the results I want, but I'm sure there are approaches that would eliminate some of the work in making sure things don't wig out when pushed. The problems are most notable if you just move the leg target up the length of the leg and watch the fingers. New rigging is only in place on one limb.


You can disregard that now. I forgot to properly constrain the bank on some important bones, and since we have to have everything on IK.... some fingers were getting serriously twisted. I do have some remaining issues, but the main one is that when I turn on Keep Target In Range, as the professor recomends, everything wigs out. Which is bothersome.

Siber
12-02-2009, 09:29 PM
Things are winding down now, and I'm getting into animating the thing. The rig is far from perfect, in particular in the hands, but at this point I don't think I'd have time to animate anything complex enough to need them anyway. I've started work on a walk cycle, as best I know how, and this is what I've got so far.

Render: http://dl.dropbox.com/u/896078/triskellion%20walk3.avi
Scene: http://dl.dropbox.com/u/896078/triskellion%20walk_v003.lws
Object: http://dl.dropbox.com/u/896078/Triskellion_v028.lwo
Other object: http://dl.dropbox.com/u/896078/Floor.lwo

There are a lot of problems that come from limitations in how I built the beast in the first place, but there is little I can do about that now. Some day I may do a rebuild of it out of class. But if anybody has suggestions on ways to improve it now, or even on ways it could eventually be improved in a rebuild, I'd love to hear them.

Siber
12-07-2009, 04:35 PM
Welp. Thanks for all the help, guys. It made this (http://www.youtube.com/watch?v=gWxTTew0a3c) possible.

h2oStudios
12-07-2009, 07:46 PM
Hey, came out pretty good :thumbsup:. Wish I could've been more of a help to ya, Siber, but I've been really busy with my own stuff. So, what'd your proffessor think about it?

Cheers, & keep pushin at it!

Siber
12-07-2009, 10:20 PM
No problem, busy happens.

Haven't shown him this yet, final showing is tomorrow. We'll see what he thinks then.

Siber
12-09-2009, 09:35 PM
Showing off the animation went great. Some other people did some very simple animations, at most two or three seconds. Everyone was pretty blown away by what I had, and I was used as the example of what everyone else should have achieved in the class.

h2oStudios
12-09-2009, 10:05 PM
Showing off the animation went great. Some other people did some very simple animations, at most two or three seconds. Everyone was pretty blown away by what I had, and I was used as the example of what everyone else should have achieved in the class.

BRAVO!! Great to hear it, Siber!:thumbsup: It feels good to have accomplished something like that eh?

Keep it up, man.