Update on Post #20 Click Here Or just keep scrolling down all the way.
Original Post
Just a question I have for artists here. Mostly this is about sculpting characters.
I have made multiple characters. Some better than others :P But the issues I have is; usually they look just about the same. Especially the face.
How could I follow the reference or the concept art better here ? Any tips ? I think most of the faces I sculpts I make look the same because I follow the face planes guide.
Replies
What search terms are we supposed to be putting into Google to explore the more esoteric facial anatomy information?
I've posted this elsewhere but this should help. They are grouped, all except the top singular one is caucasuan, whereas he is african.
https://drive.google.com/file/d/0B-qJDC2a45e8d3Z0SzVZNmRwNHc/view?usp=sharing
Bottom left is women, bottom right is men, top left is children.
I think one of the biggest culprits of sameface is not drawing/modelling different skulls.
Im guessing you have a thing for redheads? :P Look at Isla Fischer and Christina Hendricks. Both Redheads of about the same age and skin tone and both good looking. But they are worlds apart in what their faces look like. Jawline, Chin shape, Nose shape all completely different but they are both "correct" because they exist and we can see them. Dont be afraid to stray a little bit.
Fatten out the nose, widen the bridge, maybe the nose was broken once? All your noses are small and slender and follow the same basic shape, although racial background plays a part in the shape of your nose not every Caucasian person has a skinny nose like Emma Stone. Look at Zoey Deschenel's nose.
Your Jawlines follow the same shapes as well. Jennifer Aniston has a big square jaw. Scarlett Johansson's has a more rounded swoop. Then there's every possible option in between. Its all ok to do!
And that is just the major features. You can tweak all the more subtle features too, like maybe one eyelid hangs lower than the other? You literally have endless options in Facial Construction. I spent 3 years taking Passport Photos of 1000s of people every week. There are so many weird and wonderful faces out there.
What I would suggest in future is to pick a real life reference (not necessarily a celebrity). Find a good photo. Dissect the Photo marking out all your facial planes and then sculpt that way. Or even better, mix and match Photos. Maybe take an eye from one and a nose from another and build an entirely new face. Just be aware of Racial Differences in Faces.
And it's even more obvious with a lot of anime characters.
As Ysalex said, studying real people will help, and then you can inject thoses features back into stylized characters.
And if you want to make different stylized characters, try to pick some from completely different cartoonists.
well your argument already falls apart, within one style characters may look alike, but they don't have to.
just compare UP to Big Hero 6, both stylized, all different characters. and in Up no Character looked like the others, besides some random unimportant characters. As they come from the same basemesh/rig for production reasons.
Style does not mean everything looks the same, if they do it is a chosen style decision.
And seriously Merida does not look like most other disney princesses.
Hehe,
"Fat Pads of the face".
[ame]https://www.youtube.com/watch?v=JutKaXbQ6dQ[/ame]
@PyrZern: If your looking to "hit" character likeness in your stylized sculpts, then I would echo what others have said. Start practicing realistic head anatomy. All the time and effort you put into it will translate with any stylized characters you do.
Everytime facial anatomy is brought up I have to bring these books up by Philippe Faraut. Get these as they are worth the money. The amount of information that you can learn in these books can transform you as an artist.:thumbup:
http://artists.pixelovely.com/practice-tools/face-expression-practice/
@mikezoo: I was curious about Philippe Faraut's books after watching some videos of his. Thanks for confirming they are awesome as well!
2D artists imagine their character features as shapes to create variety. Here's some examples of tutorials I keep.
I think it's good to view variety in the most exaggerated terms cause I personally tend to under sell facial features. So when i'm aware i'm doing that i try to go farther past what i deem acceptable.
I also think it's good to try and separate whats "Attractive" and whats "Interesting" in your head.
I always thought disney's thing was both intentional and a result of keeping the same artists and directors around. Also people being inspired by past works and artists there.
Styles tend to be consistent with the same artists or maybe groups, but there's tons of variety out there with stylized drawings.
As with anime, yeah that's kinda true especially with certain facial features (head shape mainly), depends on where that artist takes their inspiration from. But when a lot them start out as complete amateurs there is a lot of unique stuff there still.
Part of my fun is picking apart differences in artists work that i think is interesting. It's a lot more than just faces, anatomy and expression is as big a part in style as anything else.
If anyone got any more awesome stuff to share, don't let me stop you :P
So I been working on these the last few days... I plan to add fine details and sculpt some personality or facial expression into these as well (they will stay sculpts, not going lowpoly). Should I add details first, or should I add in expressions first ??
Supposedly.... 9th, 10th, 11th, and of course 12th, accordingly.
Note; sculpting in fats still hard for me... They keep turning into blobs.
Note; would love to go more into realistic style, than the minimal-stylized-whatever I am doing... but then they keep turning into blobs as well...
Yes, I use the same basemesh I made exactly just for this bust practice... Same for males and females, actually.
( depending on your taste in artists )
A lot of famous artists are accused of secretly painting themselves in all their works
( the Mona Lisa among many other DaVinci paintings are accused of being disguised portraits! )
I doubt the majority of examples have anything to do with some secret disguise. And whether or not Leonardo really did paint himself into his work can't be proven ( unless they really do get permission to dig up his bones link ) But I personally was unknowingly guilty of as much in my teens and early twenties ( ID heavy years to be sure ). And recognizing and admitting as much irked the hell out of me ( Hi! my name is Roger and my paintings are ego accidents ). But since then I have been a strong proponent of exorcising ones ego from one's art and preaching that one's ego can easily be yer worst enemy. I would like to believe such a philosophy really helps one to observe more honestly. ( Under such a philosophy: Rappers r not musicians cuz the ego is not an instrument! . tho I suppose purposeful egocentric work might be valid for someone exceptionally honest/talented like Bukowski, Van Gough, Cobain )
Not that you are doing the same...
Just something to consider.
On the subject of hitting likeness. I try to find spatial landmarks of negative and positive spaces just like Jacque Choi's adipose example
( awesome example! checked out yer escorche on yer website as well... very helpful )
On the subject of getting stuck with "sameness"...
I would wonder who yer influences are. Perhaps more varied influences might jump start more variety in your work? ( varied inspirations )
Also...
I find that everything I do is god awful about a week after it is done. Best feeling in the world. ( that u r still improving )
lastly, I love yer avatar the best. I think she has loads of charm.
Using Neox's comparison of the style in UP vs Big Hero 6...
Where what is charming is varied by shape very differently?
Maybe it would be helpful to sculpt the gesture of whatever is charming in yer concepts...
into custom base meshes, so u don't have to fight what the generic edgeloops suggest.
( ps. what do u do with the under-boobage when using this base mesh on a male! :poly136: )
http://www.polycount.com/forum/showthread.php?t=124922
If you really want to work on getting away from what you're doing just start from a sphere and dynamesh.
Just be careful what you type into a search engine when looking for references, especially when you're at work.
A while back I did a Google image search for one of my favourite art studios, 'Massive Black'
...saw some pretty disturbing images.
first of all it is horrible, but secondly it limits you atm
just use dynamesh for now, start from a sphere or a super simple base head, don't shortcut. Just start from scratch,each time.
No premade noses, ears or eyes just start from zero every time
Make an Old Eastern European Jewish Man.
Young Asian Woman
Teenage Black Male
Scandinavian Teenage girl
and fuck basemeshes, dynamesh sphere always.
The image posted by JacqueChoi is an assignment for this course, which I highly recommend if you want to explore the more esoteric facial anatomy information:
http://www.scott-eaton.com/portraiture-facial-anatomy-online-course
yes!
im taking his zbrush course next month!!
I took all his courses because of you recommending them on the forum earlier. I'm very glad I did! :poly142:
Ref: Carlo Oberg
Started from a sphere. Left one is with dynamic subdiv.
- Avoid shift-snapping to front view in Zbrush at all costs.
- Polypaint/project some basic skintones as early as possible, and also take the time to build a basic eyeball as it will help you shape the eye sockets properly.
Good luck !
But the lips are looking awfully... different to your ref. And what you have in the nose area on the side profile seems like it's protruding too much/too large...
Look; you can only make so many assumptions.
I wanna say that upper lip overhang probably isn't right; and that the lower lip juts out much more, along with the nose being flatter. But without a ref for this guy, you don't know that for sure and also it's easy to accidentally make A) something racist-looking and duck-lips. Also, stop and think about those sharp lines and corners you always leave on your lips. Are they really there on your ref? What do they represent under the surface?
You always super simplify the eye area and some of it definitely comes from making simple cuts instead of identifying and building up the forms in the area (which are actually quite complex). Do some 2D studies of the eye of your ref and identify some key angles and overlaps.
Last but probably the most like... obvious, necks are generally protruding from the middle of the head. Maybe he just has really bad posture in your sculpt but that's a bad habit to get into.
Here's a quick ass redline full of assumptions and half-based on my OWN habits and face (and this is why people choose well-known celebrities to do likenesses of). Next time, pick a face you can find all angles of, probably. I feel like you underestimate your requirement for observation.
Sorry for popping my head in to poke holes in this, haha... It's a good step in the right direction, and I'm looking forward to seeing more.
Isn't there a dynamic perspective, or any other way to match the camera to a 3D app or engine? Also why avoid shift snapping to front view, because of lens distortion?
i noticed that zbrush perspective is not accurate by default base on Angle of View slider(or compare to other software's perspective), i think there are 2 prameters that alter zbrush's perspective accurcy, which are: LOCAL TRANSFORMATION and DYNAMIC PESPECTIVE.
by default LOCAL TRANSFORMATION is on and DYNAMIC PESPECTIVE is off, wich cuase unaccurate perspective base on Angle of View,
turning off LOCAL TRANSFORMATION, or turning on DYNAMIC PESPECTIVE will makes zbrush perspective same as other software's .
I am still not sure about thatو is anyone here to confirm that !?>
zbrush is 2.5d it has no perspective, it kinda distorts the pixels based on their position in the boundingbox, which means the perspective changes, the bigger an object gets
a lot of reference from the web also has the subject tilting their head ever so slightly forward, very visible on actor/model shoots. that can throw you off pretty bad when trying to align things.
for likenesses i agree with pior that getting some color onto the head quickly can help a lot but it does not make for the prettiest sculpts in my experience. you start relying on texture a bit early and shapes will not all be that visible to you unless you toggle frequently.
also basemeshes. improve - reuse. human heads are not that different to warrant reinventing the wheel all the time. who wants to do a base ear topology for each head?
if your model happens to be ever so lightly tilted up or down (even by a few degrees), shift-snapping to front view will lead you to *assume* that it is perfectly straight, while it isn't. Getting into the habit of not shift snapping to front view will force you to manually align the view to the actual front plane of the face.
You can try this out easily by loading an existing head model, rotating it manually to what seems like a good front view, and then shift snap to front. More often than not, these two views (shift snapped and not shift snapped) will be slightly different.
TLDR - imagine the consequences of shift-snapping to front view on a model similar to this :
http://img07.deviantart.net/60ff/i/2009/313/4/1/blackprofile_by_i_pw.jpg
I'd say that one's safest bet would be to train oneself in photography in order to be able to recognize lens type easily. This is very often overlooked, but lens type has a huge impact on the way a face looks - and not being familiar with this basically means that one big factor of the likeness equation is being ignored.