A Dream Of Mel

This probably will take more time, and I’ll throw various stuff related to the matter in here …

Preface: currently, I’m collecting material for my first reel (targetting character animation). To do so, I pick some scenes from the comic novel I’m working on, and try to set them in motion. I’m not strictly sticking to the 2D drafts, but re-create this and that in the progress, elaborate some detail, and drop others, just as I feel like.

As the female main character I use, is based on rigging- and modelling skills already outdated in my own measures, I remodel and re-rig her, while I use the old model for some animations and scenes, to avoid getting stalled in preparations for eternity.

The comic contains nudity and targets grown-up audience, therefore I model my characters ‘complete’, and some imagery I post here, will display nudity.

Character rigs use muscle systems.

Main char intro (deprecated version, nudity warning): http://www.thalion-graphics.de/portfolio/anim/telaya-intro-v3.avi

Comic pages (a dream sequence) serving as loose drafts:



Some animation frames:




More details, and finished 1st animated scene later (currently rendering, taking about 20-30min/frame on i7/2600k, 16GB argh) …

All hints plus C&C welcome.

1st scene rendered: http://www.thalion-graphics.de/portfolio/anim/draakhrider-wip.avi

Unfortunately, I forgot to double-check compositing, too much glare … her head movement and hand position change is animated too jerky, too. Meaning re-render …

Btw., merry X-mas all. :slight_smile:

Hi, man. Super environment in your 1st scene, but animation is still draft.
Don’t render until it will look good in preview.
If you do char. anim. reel, you can place obly previews in it, render will only slow you down.
I suggest you to share you anim. previews and wait for some critiques on them.
When it will look not too jerky - render them out.

The models were stunning, I found nothing wrong with the animation, though I did think that the dragon’s movement was a bit sluggish. :smiley: food for thought.

Thanks a lot for your feedback, guys!

I should probably write “animated character” instead of “character animation” as main topic of my future reel, as it is unrealistic for me to achieve really polished animation at that point - I hardly did four or five animation sequences so far, for testing purposes, and, for that reason, will have to practice a lot, until my animations will be fluent. The sequence here has a bit of old stop-motion-animation l&f in my own eyes … there’s probably a lot of smoothing-the-arcs required. So, of course you’re right, Shader2 - should have waited with final rendering. ^^

My first two years of 3D have almost completely been dedicated to rigging and skinning - my main goal has been to find out how to make characters which are able to move naturally, so, currently, this would be the main topic of my future reel.
Here are some of the under-the-hood-details for the animation:

http://www.thalion-graphics.de/portfolio/anim/0112.png (click on image for animation)

As this learning process is far from being complete, and I have made many errors and unnecessarily complicated things (have put much too much emphasis on rigging hacks to get good deformation most of the time - now I am trying to solve much more by using optimal topology), I started to re-model and re-rig the female character (again, loosely, version 3.0 will not be an exact copy of 2.0, neither from technical, nor visual aspects).

Currently, I’m re-adding her muscles. I’m hoping to integrate a nice softbody-jiggle-solution for the muscle system with animatable influence this time. Imagery and details will follow. :slight_smile:

Can we see a still pose of her, in a T pose, maybe the neutral posse you used
to bind her to the armature for example? and some wires?,

After looking at the animation I think there’s still some problems with how the shoulders deform and on the turn table part at the beginning pay close attention at her feet, one of them is not touching the ground by maybe 2 or three centimeters, this brakes the illusion of weight, focus on sec 7 on the left foot, you can clearly see the projected shadow and a gap between the feet and the ground.

I also feel the deformation of the shoulder feels forced, check sec 4, some how it doesn’t look as natural as it could, the arm feels a bit disjointed and since your going for very realistic style little details like this stand out a lot.

The comic pages are great as well, the style is strong and you know your anatomy.

Overall Great work, its shows dedication and lots of skill, looking forward seeing more.

The dragon animation looks Great, the weight of the beast is near perfect! good work!

You have very good flowing style, artistically.
Attention to details is marvelous - excellent work.

Thanks a lot!

Here is the mesh for the female character:

As you can see, the topology is all but ideal, particularly in the shoulder area.
Anyway, I will not fiddle around with more fixes here, but try to avoid these mistakes in the next version …

@Helmut, you low-poly female seems to be designed to be used in game engine, not in film. Yes it is easy to animate, but she has to be smoothed while rendering.

Yeah shoulder topology could use some work, also the legs look a bit too short, usually to make a woman as attractive as posible people tend to make the legs longer and more slender, I dont know if this is intentional but right now the legs look unnaturally short.

Her proportions give me a drawfed feel, and in my liminted experiece I think the problem is legs to torso ratio.

I found this image on the web, the boobs on this girls are way over the top, but
for comparison purposes its an useful image, right now the proportions of your model looks more like the girl on the left side, where as in the comic she looks more like the green girl, then again in the comic pages she is never standing so I’m not %100 sure, I would need to see her just standing to
make a better judgment.

I mention this because I’m not sure if her proportions are intentional, if they are then it looks
good, just thought to mention this, in case you could find the observation useful.

Also for the topology this site has given me a lot of ideas in the past maybe it can help you too, its by far the best repository of topology iv found on the net:




Shoulder topology, best example I know, the page might take a bit to load
because its got a heavy .gif animation showing a 3D shoulder deforming:


Did you made a orthographic sketch of Mel before you made the model?
like a drawing of her in front, back, and side views to aid you in the modeling?
if you did could you post it?, and if you didn’t since I can see your a rather
accomplished comic artist I would suggest making this orthographic drawing before
moving forward, so you can use it to check the proportions of your model
against it, this step is invaluable to model and correct a 3d model.

Looking forward seeing more

dunno why it double posted
so i deleted the text, and admin can delete this post later.

Hey Renderluz,

the proportions were off intentionally - not the typical 1-3-(>=4)-heads proportions. The idea was to have her appear a bit short, and muscular (call it an experiment; I wanted to figure out, why the typical Frazetta girl proportions, being rather off the standard “beauty” ideal, work so well - at least in my eyes).
However, I probably exaggerated too much, and in an inconsistent manner, so proportions are more according to the “standard” canon with her next incarnation. Here are some WIP images:

Just the deformed low-res mesh with mirror & subdiv:

Slowly readding anatomy details:

First shoulder deformation test - front view:

Some shots with wireframe view enabled will follow, for topology comparison.

Leagues better than the first example, she still a petite lady but the legs to torso ratio is much nicer to look at, and the anatomy in general looks more sound, I also like her eyes a lot more, mouth deformation looks really nice too.

Looking at that shoulder deformation test it’s looking much better than the examples on the previous video I saw, tho I know it still a WIP so I’m sure you will clean the deform more to get rid of artifacts, it still looks very promising.

Are you still using a muscle inner system or are you just using corrective shape keys?

Right now the only thing that I think could use improvement right way are the feet, the toes look too regular, I can’t pin point it, but something its not quite there on the way the toes are displayed.

This is a tutorial by Steven Stahlberg that I used more than once, it shows how to make a femenine feet with subd modeling and it could help you to sort out the feet:


This is his gallery, full of curvilishious women :wink:


I’m also a fan of Frazzeta and I can see his influence on your work which is a great thing, :slight_smile: coincidentally Steven Stahlberg is a Frazzeta fan as well, this is an old dicussion about deformation and women topology on CG talk that might give you ideas:

On the whole I’m really interested on seeing how your project progress,
to say the least I’m very impressed and inspired by your work,

Looking forward seeing more.

O whew, thanks for those links! :eek:
I should have posted here much earlier - these pages would have saved me months of trial and error (mostly error).
Shoulder area topology now is much closer to the example shown in the tutorial you posted (my source of information has been this thread - i guess now I know where the guy found the template for his topology).

Yes, I do use “inner” anatomy again, but I don’t want to use it as corrective mechanism, and I intend to skip inner anatomy in areas where it doesn’t really contribute a higher level of detail, like butt cheeks & gluteus muscle.

I updated the shoulder animation test, and here also is the rear view.

It shows some nice skin slide effects, which are hardly achievable via shape keys, I guess … but also some of the typical problems introduced by shrinkwrap-based muscle systems: ugly, jerky artifacts, where skin and shrinkwrap target meshes are angled towards each other such, that skin vertices are projected towards the wrong side of the target mesh.
Usually, it’s possible to get rid away of those artifacts by fine-tuning skin mesh deformation and shape+deformation of the muscle meshes. If not, well, I must seek another solution, or drop muscle sim at that particular location.
We’ll see.

Well … rigging is a painful business.
I’m still stuck in deformation optimizations …

However, things start to turn out quite promising.
Here are some animated tests: http://www.thalion-graphics.de/portfolio/anim/t3-reel.avi

This still shows nice (imo :wink: ) leg & hip deformation, achieved by the muscle simulation:

Btw, I’ve tried to improve her feet - better now?

Man, this is a very good thread! Love the style (proportions and such) and it seems to be on the right way! Keep it up!

Nice work, Is there any way that you can post front and back images of the muscle rig you created for her? What technique do you use to make the muscles deform? How do you ensure that you get the muscle shape you want when contracted and expanded? Are you using a simple shrinkwrap for the skin? If so, what do you do to add additional detail that might not be in the muscle sim? Are there small tools that would make this technique more efficient?


it’s still ongoing “research” …

I use ShrinkWrap and Cast to “skin” the muscle system - ShrinkWrap for muscles, Cast for some rigid elements (bones).
Muscles are very simple mesh objects controlled by the character armature, with rotational symmetry around Y-axis where possible, to minimize bone roll problems (it just goes unnoticed, if such a muscle rolls around Y-axis). To simulate muscle contraction, I use bones with StretchTo constraints, and, at the moment in just one case shape keys: the “six pack” swell of rectus abdomini is modelled via shape key. Maybe, I’ll switch to something shape-key-based for the deltoides as well, because the current mechanical solution still annoys me with bone-roll issues (I ‘hate’ that silly bone roll!!!).

In the muscle sim for the last incarnation of the character, I used soft bodies on the muscle objects to achieve jiggle, to some degree. It worked, but is extremely cumbersome to animate, because each muscle has a softbody cache of it’s own, which needs to be managed.

In the future, I hope to implement jiggle for the whole system by a single softbody with vertex groups representing anatomical elements, which will be targetted by bones via StretchTo.

However, one big drawback of shrinkwrap-based muscle sim is, that the shrinkwrap modifier tends to project vertices in directions you don’t want. Here are some sketches, which illustrate the problem, and the solution I’ve found:

This is, what we’d like to happen: a skin vertex is pushed away by the muscle, and translated in outward direction:


However, if the muscle is positioned in a disadvantageous way, related to the skin, the vertex may be “sucked in” instead, and will be projected to the opposite side of the muscle object. This happens just, because the vertices on the opposite side are closer to the skin vertex:

Another problem arising when using “Nearest Surface Point” shrinkwrap mode is, that vertices may also be translated sideways, because the muscle object’s lateral bounds are closer to the skin vertex, than it’s outwards bounds:
The solution is simple, but somewhat brute force: just eliminate back sides, by joining up a common target object, which has no “inner” bounds, using booleans. This, of course, has a performance penalty, but on the other hand even reduces the count of ShrinkWrap modifiers on the character mesh itself. Before CarveLib integration, Boolean operations have been to buggy to produce results good enough for this solution (as booleans tended to create additional, redundant, disconnected vertices or even mesh parts, which confused ShrinkWrap), but as of now, this seems to be a good way:
This also allows to give muscle groups common influence over much larger parts of the skin mesh - you don’t need to fit influence areas for, say, biceps brachii, tightly around the muscle, but, as you now use a common target for biceps and triceps brachii, allow biceps deformation around the whole upper arm area, which greatly improves deformation and skin slide effect, because now biceps deformation effects do not require the armature to “prepare” the respective, biceps-only skin area by pulling it to a proper position.

Here are front & back images of the muscle system:

These two show front & back view of the merged upper arm, forearm, torso, thigh and shin target objects:

Concerning additional detail layer “over” the muscle system: I plan to use (driven) displacement for that purpose.

Additionally, there are several means to reduce/increase muscle system influence: one is (driven) Smooth, which I already use, and another one is provided by the VertexWeight-Modifiers, which are also a great addition in the toolset available for anatomy simulation.

Here is an animated test of the standing splits (also showing trouble with damn hair dynamics …):

Another progress image with the anatomy simulation:

I’ve added some subtle skin crease effects (by use of driven displace), more muscular detail, a grain of (currently hand-animatable) gravity effects in the abdominal area, improved volume squeeze in knee & elbow joints (also by use of driven displacement).

However, work is proceeding really slooooowly … some nasty, painful inflammation in my shoulder and back musculature keeps me from sitting in front of a Wacom & computer screen too much, and if I do anyway, it should earn me some money …