Alice (MGQ)
Current:
Hi polycount! Loved the detailed WIP or making of threads like the Hazardous course ones on here, and decided to join. Never made a character and am rather inexperienced with 3D in general, so I want to start a thread of my own were I slowly build a character from scratch and learn in the process. I'm going to try and write out as much of the process as I can, and hopefully that will help me learn and make it easier to critique both my workflow and execution.
As the character to make I picked Monster Lord Alice from (NSFW) Monster Girl Quest.
There are a few reasons that I decided to pick her to make, aside from having a nice character design and being surprisingly well written considering the source:
I'm going for a high poly character for next gen or cinematic, and aiming for a blend of stylization and realism with her.
Please critique my work as being judged by professional standards; while it is silly to compare my work to professionals, I'd rather you do and tell me my work is terrible, I learn everything that is wrong with it, and try and push both it and my abilities past my current limits, than let even minor flaws slide by in complacency.
Hopefully I'll be able to learn a lot and come up with nice results, with the help and critique of polycount, and I'll try and not fall flat on my face in the process
Hi polycount! Loved the detailed WIP or making of threads like the Hazardous course ones on here, and decided to join. Never made a character and am rather inexperienced with 3D in general, so I want to start a thread of my own were I slowly build a character from scratch and learn in the process. I'm going to try and write out as much of the process as I can, and hopefully that will help me learn and make it easier to critique both my workflow and execution.
As the character to make I picked Monster Lord Alice from (NSFW) Monster Girl Quest.
There are a few reasons that I decided to pick her to make, aside from having a nice character design and being surprisingly well written considering the source:
- Nothing to hide poor anatomy behind (though admittedly no practice making legs/feet).
- Not a lot of clothes/accessories to have to make.
- Large variety of surfaces/types of objects to try sculpting: scales, skin, horn, metal, silk, flowers/plants, gemstones, and hair.
- Likewise, there is a large array of textures and materials needed, ranging from matte to reflective, differing sharpness of specular highlights, reflection/refraction, SSS, etc.
- At the same time, I don't have to do too much of any one type of surface/object, and so it shouldn't get monotonous.
- Sculpting and texturing tile-able patterns like scales is something I really want to learn to do, and yet am intimidated by, so it'll be a good challenge.
- Stylized concept makes it a fun challenge to imagine how certain parts would work in more detailed 3D, such as skin to scale transition.
- Eyes aren't too large, so I can try and attempt them as 3D shapes instead of just painted on.
I'm going for a high poly character for next gen or cinematic, and aiming for a blend of stylization and realism with her.
Please critique my work as being judged by professional standards; while it is silly to compare my work to professionals, I'd rather you do and tell me my work is terrible, I learn everything that is wrong with it, and try and push both it and my abilities past my current limits, than let even minor flaws slide by in complacency.
Hopefully I'll be able to learn a lot and come up with nice results, with the help and critique of polycount, and I'll try and not fall flat on my face in the process
Replies
I heard that a good workflow starts with planning, research and gathering references before drawing the concept or starting in 3D. Fortunately, I don't have to worry about the creating concept, but I still think I should do both planning and research, as well as gathering reference.
Now, obviously the concept is one that takes a relaxed approach to biology, but I still would like to make it as plausible as I can. A good rule of concept art try and ground the design in reality as much as possible, even if it is outlandish, allowing for easier suspension of disbelief and a better concept. So I too am going to try and ground Alice in reality as much as possible without changing the already existing concept.
In doing so I'll start with a dominant part of the concept: the tail. To figure out how long to make her tail, I'm going base her tail on the extinct Titanoboa, which measured 12 to 15 meters and had a max diameter of 1 meter.T he snake tail merges with her body at its thickest part, and I'll give her a hip to tail ratio of 1:12 to 1:15.
For scales, I plan on basing it off of python scales, as a type of python is the largest living snake and the little bit of back scales shown on the concept resembles a python. Measuring her she fits the 8 head tall heroic proportions (except her head is stylized), and so that is what I plan on using for the body. That puts her hips at 1.5 heads, and her tail will then be between 18 and 22.5 heads.
A problem I ran into is that human spines end in a coccyx, which would not work with her lower snake half, especially since she appears to have the crotch and butt of a human. So trying to imagine what her spine would look like, I came up with two ideas: Either her spine temporarily moves to the center of her hip cross-section, or her spine temporarily forks into two and runs through the area where the back of her legs would be if she was human, before joining together to form her snake half's spine.
I'm leaning towards the forked spine route, as it is both possible, and in fact happens in snakes quite often compared to other animals (pictured is a two-headed snake), and allows for a much smoother curvature in the spine than a single spinal column would necessitate. It also frees me up to handle the human to snake transition with a much more simple blending of the two parts.
The transition between scales and skin is definitely one of the areas I want to nail, and so I try to gather real life examples of skin merging with scale. I found two mammals that have this, the armadillo and the pangolin. Never heard of the pangolin before this, but it is awesome! Perks of getting references: discovering cool stuff
Moving up, for the skin itself I'm going to stick to as human look as possible, but still having the blue shade of the concept. I thought of being lazy and referencing Avatar, but decided to try and find examples in nature. Blue skin on mammals is very rare, and it occurs in only a few species like the West African mandrill and the vervet monkey. Fortunately, they can exhibit blue hues more vivid than I need, so the blue I'm going for is possible on a mammal.
Alice is by the concept a symbiotic creature that combines a large variety of lifeforms, though I'm trying to err on the side of warm-blooded mammal/human whenever possible; the flowers in her hair aren't accessories but actually a part of her body. To me they appear to be similar to hibiscus, with wider petals than the one's of the scarlet hibiscus in the pictured refs. I think I'm going to add in some of the shape of lilies and remove the stamen from the hibiscus. Also, the thorns on the stems of the flowers are more dull and round than my ref.
The horns on her head seems to start like sheep horns in the back but quickly go very flat, even flatter than the elk antlers. I think that I would like to combine the smooth surface of the elk with the light ring-like striations as seen on the sheep.
Her accessories are that part I'm most uncertain about, but also since they are independent of her model I feel I'll have more room to alter them later on. Her jewelery and the metal parts of her clothes look gold or gold-plated and have insets of reddish-orange gems. For the gems it appears that I will use rubies/garnet colors with a yellow shade; also, the stones seems to have either a cat's eye/star ruby effect to them, or have another gem embedded inside of them. The remaining metalwork of a purplish color is something I'll just decide on later. Her gloves I'll make silk, as silk would seem technology/culturally appropriate, but her skirt... The shade/specular suggest same material as gloves, but they also appear to have a solid shape. Maybe metal plates covered by black silk and with a purple silk trim? Going to think about that.
Hopefully the more I plan the better the end result and the easier it is to not hide flaws as artistic license. Also, sorry for being verbose, especially from a beginner whose process probably has little to offer others.
If you spotted any errors in the initial planning or have suggestions for improvement I would love it if you let me know. Any ideas on how to make it more biologically sound, more interesting, foreseeable problems that will come up later, etc. would be really appreciated.
Throttlekitty: Nobody likes to do things that feel like work or a chore
Currently building a base mesh in Maya before taking it into Mudbox and have a couple questions:
What is a better idea of sculpting the tail: to make a base for the entire tail and sculpt that, or to sculpt on a segment of the tail and then duplicate it out to build the tail? The belly scales seem like they would lend themselves well to segmentation, but the back scales would be more of a problem. I could try and make it so that the scales tile across segments, but would that lead to too little variability across the tail? Unlike lizards, snake scales seem very evenly patterned, and would need "hero" scales in only a few spots, so would this be a good idea?
If I'm going to sculpt the tail as a segment, does anyone know a good way to build a tail from it and combine it with the torso in Mudbox? Since I'd want to get fairly high details on the segment before duplicating it to build the tail, I'm concerned that Maya, were it would be relatively simple, can't handle the poly count necessary. Not sure how I would use the "Sculpt with maps" feature for this either, since I'm not sure how the high level segment would be baked for each part of a complete low poly tail.
Working on building a base mesh with good edge flow and even quads, and it's going alright but I'm running in to one annoyance. When I make vertexes of valence 3, I sometimes get an odd artifact, like a tiny sharp spot right at the vertex (valence 5 is fine). It only appears at higher subdivisions, but is really visible and refuses to go away with smoothing.
The valence 3 vertex in the center of the left plane will cause this noticeable artifact, especially on concave surfaces (less on convex, none on perfectly flat). But the vertex on the right plane doesn't produce the artifact when subdivided. Is it because the one on the right has more even quads compared to the kites formed on the left? Is there anyway to get rid of this artifact? I've been trying to make sure my topology uses more valence 5 instead or has the look of the plane on the right, but I would love to know how to get rid of that artifact.
Going to fix up proportions of fingers, hands, body etc. while sculpting. 4.6k tris in the torso and another 2.1k in the tail for a total of 6.7k. Didn't bother showing the tail right now as it is really long and not very interesting, but I did decide to have it widen slightly after connecting to the torso before narrowing to the tip since it looked better like that.
Aiming for rough forms, not really trying for details yet. The clothes are just placeholders and I haven't started the arms/hands yet. Want to work a bit more on the back of the trunk and shoulders before posting a back view.
Any critiques and thoughts on anatomy/proportions really appreciated.
Planning on roughing out the arms and hands next.
Are the proportions and forms correct so far? Having never drawn or sculpted humans (or anything basically) I'd love to get a second pair of eyes to check if I'm on target.
The subdivision level is a little high for the level of detail - like in the previous posts for 'roughing out forms' the clavicle and tits are clearly defined already and there's definition on her six pack - but the larger forms like her ribcage, hips, and arms just isn't there yet.
I'd recommend raising her arms to a 45 degree angle to make sculpting them a bit easier on yourself. Stepping down a few levels to the point where the model gets blocky and working as much detail in on that level as you possibly can. This should force you into getting those large forms and hitting the silhouette. Then when you've gotten as much detail out of that base layer then subdivide and do the same thing on the next level up.
Does the basemesh have a head?
Also of note if you're working on female forms is the work of Paul Richards especially his notes from Babe Lab. link is possibly NSFW - http://babelab.blogspot.com/
It's an interesting concept and should be a lot of fun to try translating into 3d. Keep at it, this'll be cool to see develop over time.
I try to alternate between subdivision levels, pushing big shapes on the low and then building up on the high with strokes since the brushes feel better on the higher levels, but you are right, I need to focus more on the overall shapes. I made some of the details like the clavicle and abs to help act as landmarks in getting the rough shapes, but I spent too much time focusing on the details.
There is no head yet, I want to get more comfortable sculpting before I make it since it is the most critical part. I do have proportions set up for it though, so I'm basing the body off of that.
Thanks again for the critique
She's curvier now, but breasts still need to be made bigger; her illiac crest may need to be smoothed over and she might need to be made more voluptuous to be truer to the concept. Are her proportions and volumes looking on target and does her body-type seem to match the concept?
Will try and cycle over parts in iterations to not get bogged down and miss mistakes; I'll probably do the back next. Any tips to help 'refresh' your perspective while working?
I should post up more of the concept to make it easier to make critiques, but unfortunately this is basically the only other useful image I have of her:
And fixed up her a little, hopefully the stretch is gone now:
Made a base head, and like you guys suggested tried to get major shapes blocked in before building subdivisions.
The end of the jaw and ear area will need retopo, but hopefully the general topology of the face/head will work; most importantly, I'd like to know if I'm on track with the main volumes.
I'll get there eventually
http://www.hazardousarts.com/Miku_Tutorial.pdf
Haz doesn't just pull nice looking anime proportions out of his butt or directly copy the size and shape of anime stuff, dude knows his anatomy and proportions and where and when to exaggerate them to get a similar look. If you don't know the structure underneath and how faces are formed in general you'll just get something that looks kinda creepy, which is what you have now.
I highly recommend studying anatomy and form before finishing this up if you want a super nice result. You've got plenty of material and surface detail reference but you should be collecting reference for attractive women, anatomy, proportions, etc just as much if not moreso than surface details. The reference might not match what you want to accomplish exactly but it's up to you to apply what you learn to what you want to do.
Search pinterest, google images, or go to your local bookstore and buy some books on anatomy. There's absolutely tons of stuff out there to help you.
Good luck!
ceebee: I agree 100% about anatomy, and that real proportions are the foundation for stylization that can't be ignored. I've had the amazing Thieme books on anatomy in front of me, and I have no excuse for not using them more. And yes, it looks creepy right now.
I think I finally was able to wrap my head around a critical part of the stylization. The way the jaw looked in different illustrations confused the hell out of me, and I was trying to recreate it in 3D and failing miserably.
But it was so simple: it wasn't actually a jawline, but the jaw + the cheeks that formed the shape. It just looked like the jaw since the 2d drawings don't show depth well.
She'll get eyelid and eyelash geometry, but for now as I work on large forms she has giant, empty eye sockets. Lotsa work to be done, but now I have hope of being able to do it eventually
I posted a quick breakdown of facial proportions in another thread that you might find handy:
http://www.polycount.com/forum/showthread.php?t=121501
There are also tons of resources on youtube that deal with this topic, take advantage of them.
Starting to get more of the forms I'd like. Also, changed up the topology a bit.
Worked on the ears, changed up head and face forms. Tried adding eyelids, but it failed horribly, will do it a bit later so for now she still has empty eye-sockets.
Just like ceebee said, i think u should take a step back and study anatomey/ face proportions. your model just does not look right even if it is a stylized model you are aiming for. I just did a quick guidline for you. hope it helps.
cheers and good luck!
I think most of the larger forms are there now (though if I'm wrong I'd love for someone to correct me), so time to subdivide and start working on the mouth, nose, brow, etc.
Don't think the head topology will change much for the high poly, so will attach it to the body next.
Also edited some images to be clickable thumbnails to make the thread cleaner.