Dynamic Character Project: Male Human
Greetings all: time for the dreaded first post. ;P
Just a quick background: I hail primarily from the GameDev.net forums, after having recently taken interest in solidifying my rather sketchy modeling, rigging, and animation skills. I consider myself well rounded in many game-dev-related talents (including various aspects of coding, conceptual art, music and audio, graphic and texture art, and a small handful of other things), but my 3d work has always suffered at the hand of practicing other skills.
Next semester, I'll be undertaking a personal study that involves imitating the dynamic character creation and rendering seen in many next-gen games, where a player begins by choosing from a small handful of basic character templates/races, then customizes their appearance/hair/body type/&c.
I feel safe enough on the coding end of things, but I know I'm going to be facing quite a challenge on the 3d side of the coin. And, naturally, because of the nature of the project, if I make a mistake early on in one of the base meshes, it could spell disaster later on in the semester, so I want to be able to get as much critique and advice as I can at each step in the process. Word has it that PC has plenty of critique to go around.
Without any further prattle:
---
Subject: "Nude" and "bald" male human body and head.
Polycount: 2356 triangles
Timestamps: This update: 12/16/06; Started: 12/09/06.
Purpose: The initial mesh for the male gender of the human race. Will be rigged to and animated by a skeleton, morphed to morph target meshes (to change weight/musculature on the body and facial features/proportions on the head), unwrapped and given diffuse/normal/specular maps, and will have hair/headgear/clothing/armor mounted onto it.
Especially needs critique in these areas: Topology, readiness for animation, appropriateness for normal mapping, neutrality of form (for customization later).
Additionally, feel free to input on: Possible polygon preservation techniques, the character's style (or current lack thereof), my probable misinterpretation of bodily proportions.
Suspicious areas: Eyes (seem a little small or far apart), mouth (topology as well as shape of the lips), ears (general ugliness, lack of definition, shifty topology), ankles (brutish form, need to find a good reference).
Notes:
Working in Max 6 (almost 3 versions outdated!): if you'd like to take a look at the mesh, catch me on MSN (justinsmith@sevenelements.net). If you like it so far, let me know: it helps when I know I've taken a couple steps in the right direction. If you see problems, tell it to me straight: better to find out now then later.
Pardon the smoothing groups around the ears and lack thereof in certain other places. In-game lighting will be per-pixel, but the smoothing in the viewports throws me off.
Media:
Just a quick background: I hail primarily from the GameDev.net forums, after having recently taken interest in solidifying my rather sketchy modeling, rigging, and animation skills. I consider myself well rounded in many game-dev-related talents (including various aspects of coding, conceptual art, music and audio, graphic and texture art, and a small handful of other things), but my 3d work has always suffered at the hand of practicing other skills.
Next semester, I'll be undertaking a personal study that involves imitating the dynamic character creation and rendering seen in many next-gen games, where a player begins by choosing from a small handful of basic character templates/races, then customizes their appearance/hair/body type/&c.
I feel safe enough on the coding end of things, but I know I'm going to be facing quite a challenge on the 3d side of the coin. And, naturally, because of the nature of the project, if I make a mistake early on in one of the base meshes, it could spell disaster later on in the semester, so I want to be able to get as much critique and advice as I can at each step in the process. Word has it that PC has plenty of critique to go around.
Without any further prattle:
---
Subject: "Nude" and "bald" male human body and head.
Polycount: 2356 triangles
Timestamps: This update: 12/16/06; Started: 12/09/06.
Purpose: The initial mesh for the male gender of the human race. Will be rigged to and animated by a skeleton, morphed to morph target meshes (to change weight/musculature on the body and facial features/proportions on the head), unwrapped and given diffuse/normal/specular maps, and will have hair/headgear/clothing/armor mounted onto it.
Especially needs critique in these areas: Topology, readiness for animation, appropriateness for normal mapping, neutrality of form (for customization later).
Additionally, feel free to input on: Possible polygon preservation techniques, the character's style (or current lack thereof), my probable misinterpretation of bodily proportions.
Suspicious areas: Eyes (seem a little small or far apart), mouth (topology as well as shape of the lips), ears (general ugliness, lack of definition, shifty topology), ankles (brutish form, need to find a good reference).
Notes:
Working in Max 6 (almost 3 versions outdated!): if you'd like to take a look at the mesh, catch me on MSN (justinsmith@sevenelements.net). If you like it so far, let me know: it helps when I know I've taken a couple steps in the right direction. If you see problems, tell it to me straight: better to find out now then later.
Pardon the smoothing groups around the ears and lack thereof in certain other places. In-game lighting will be per-pixel, but the smoothing in the viewports throws me off.
Media:
Replies
ill post some more usefull crits tomorrow sorry
well, to start your character doesnt seem to have the triangles it sas it has at a first glance i would give it 1500 tris
[/ QUOTE ]
Hmmm... In Max, I count by converting a duplicate of my mesh to an Editable Patch (triangulates the entire mesh), then I use the 7 hotkey to check the polygon count. 'Tis how I've always done it, but is there another more widely used way of doing things that could possibly give a lower count?
[ QUOTE ]
I would also recomend that you start slow
[/ QUOTE ]
Yup: learning takes time! Took me many years to get where I am in the other things I do, so I know it's gonna be a slow [and probably painful] process to get my 3d up to speed.
Reason I chose to go with a next-gen goal is partly because I have an entire semester to focus on it, and I think I would be under-challenging myself spending an entire semester on 6th/5th generation assets, and partly because the academics here are very focused on next-gen techniques: perhaps a bit to the downfall of students who become more familiar with the technical aspect of generating assets than the artistic aspect of making them come to life. :{
Oh, and just to clarify: 3d's definitely not my strongest point, but I have had my helping of it, including character, object, and environment work for UT2004 mods. Next-gen is certainly a jump up from my past work, but I'm not brand new. In other words, I'm not blindly jumping in over my head: my doom is premeditated. ;}
Aaaanyway, I'll eagerly await your more detailed critique. :>
Polycount: Unchanged.
Timestamps: This update: 12/16/06; Last update: 12/16/06.
Especially needs critique in these areas: Topology, readiness for animation, form.
Suspicious areas: Mouth, ears.
Notes:
Got some useful critique about the position and roundness of the eyes from a colleague. He also mentioned that the eyebrows, jawline, and cheekbone had lost some of their definition since the last time he'd seen it, probably due to my going through and redoing most of the major topology.
Went ahead and made some changes and adjustments. I get the feeling that the face is going to be the most difficult and time-consuming aspect of the mesh, but I could be proven wrong. ;P
Media:
To-do:
Correct the mouth:
<ul type="square">[*]Draw back the corners to give it a rounded shape. Right now, he looks like a fish.
[*]Round out the shape of the teeth to fit the horseshoe-shaped bite of a human mouth.
[/list]
Work on the ears:
<ul type="square">[*]Probably start from scratch. They're ugly.
[/list]
2 much time spent writing and not enough screwin around me thinks
[/ QUOTE ]
:P
I write fast.
Polycount: Unchanged.
Timestamps: This update: 12/17/06: 1.57am; Last update: 12/16/06, 11.54pm.
Especially needs critique in these areas: Topology, readiness for animation, form.
Suspicious areas: Mouth, ears, eyebrows.
Notes:
Rebuilt the ears: they're cleaner, but still very much open for critique. Seems like most of the detail in the ears will be handled in the normal map, but I feel like there's something I could be doing to give the geometry a nicer touch...
Also fixed up the mouth: he now enjoys the rounded jaw that his fish-faced former self lacked.
Trimmed down the balloon-for-a-skull and brought the face in a bit. I need to work around the jawline to see how much definition I can get out of it with some reworked topology.
Still screwing around with the eyes- they're hard to position without eyebrows textured on, which leads me to believe that I should probably apply a quick and temporary eyebrow texture. <_<
Anyway, well past my hour of sanity. I'll take another look at this in the morning with fresh eyes and your critique. I got a bit of [fairly positive] feedback on the topology and form of the body, off the forums, but I'm still hoping for critique on it, here. I haven't done anything to the body since the OP.
Media:
To-do:
-Sharpen the jawline. Pull in the skull along the upper end of the back side of the neck, to make the jaw more prominent, and to give the ears a stronger position: right now they're free-floating on the sides of his melon.
lol everybody run, he speaks like a nerdrobot
objective: to take the piss, for fex sake
TARGET AQUIRED KILL KILL KILL ALL MALE GENDERS OF THE HUMAN RACE
[/ QUOTE ]
:P
for the front view. His hands and feet are enormous in comparison to his head - shoulders not far enough apart
Generally its lacking a good silhouette.
scroll down.
As for the model, I totally agree about not using solid references. He just looks like massive hands and very thick lower legs (ankles especially).
And perhaps not put a link to your website, when there is no website to be found. Just was a bit dissapointing to go to your site and see nothing but an 'under construction' notice. Just a thought.
But, keep plugging away at it. The base is there, now just tweak the hell out of it to get it looking how it should.
I'll see if I can get my hands on some stable references. Without getting too far off topic, does anyone recommend (or recommend against) 3d.sk? Seems like a valuable resource for a reasonable price.
@Johny:
Thanks a ton- a very informative read. :}
Re: Writing
I think Smirnoffka summed it up well. It honestly only takes me a few minutes to write up a response, and I figure the more I spill my brains here before I start jacking around with my mesh, the easier it is for someone to point out a potential problem or a problem with my train of thought before I put it into action.
I also occassionally glance back at the thread in the morning, to make sure I didn't forget anything from 4am the previous night. ;} Easy way to get right back into the groove.
Re: Website
I'll have something better than an under-construction page up before the weekend.
Re: Model
Spent most of tonight working on another project, so I'll save the update for tomorrow. I'll put all this critique to good use: I'm actually kinda glad to hear that the model's shape is the problem, and not the topology.
not aquired : defining a visual target of what the morph system is supposed to look like.
:P
What I mean is, you obviously planned your objectives quite well, but why didn't you put together a series of human males pictures side by side in photoshop to help you visualise what you are aiming at? Looking at references for ears or hands will only get you this far - you need to look at the whole picture.
Speaking of which, why do you want to fight with something as complex as a morphable human body system since you consider that you still have progress to when it comes modelling experience? I'd say go easy and humble on this, you'll have more fun.
Just my grumpy early morning thoughts! Good luck with the project.
why do you want to fight with something as complex as a morphable human body system since you consider that you still have progress to when it comes modelling experience?
[/ QUOTE ]
Ah, that's just the trick. This project is by no means a 3D art project: it covers a good number of skills, including graphic design, software design/development, user interface, and perhaps a little audio. I know I've got the non-3d areas covered, and I figure that this project has a healthy enough reliance upon my 3d skills that it'll be good practice for me, but that it doesn't rely so heavily on the 3d end of things that the project will tank if I don't nail the characters.
And yeah- I'll get my hands on some clean references, ASAP.
Polycount: Unchanged.
Timestamps: This update: 12/18/06, 9.46; Last update: 12/17/06, 1.57am.
Especially needs critique in these areas: Form, neutrality of form/personality for easier customization later.
Suspicious areas: Mouth, ears, nose.
Notes:
Mmm... shelled out for a 3d.sk subscription, and I must admit, a good set of references makes a world of difference, eh? Looked up a moderately neutral male face, and wow. Instantly noticed many of the problems with the mesh: hopefully you'll be able to see some improvement in this update.
Closed the mouth and brought the eyes to a more natural position: easier to work with them in their natural stances, and I can reopen them when I get around to texturing, anyway.
I'm gonna attack the body, next (fixing proportions and reworking the shoulders/back), but I'm especially looking for more feedback on the head: it's definitely the bigger challenge, here. Would the mesh benefit from more geometry around the ears/nose, or should I leave those things to normal mapping?
Media:
To-do:
<ul type="square">[*][Possibly?] Add more detail to key areas of the face.
[*]Rework body proportions with proper references.
[*]Rework shoulder/back topology for better animation.
[/list]
I think the lowpoly would definitely benefit from a little more detail. You probably don't need more on the ears, but I believe the nose is a bit too formless to expect a normal map alone to sell it.
Polycount: 2584.
Timestamps: This update: 12/19/06, 6.18pm; Last update: 12/18/06, 9.46pm.
Especially needs critique in these areas: Anatomical accuracy, neutrality of form/personality for easier customization later, quality of topology for animation/deformation.
Suspicious areas: Ears.
Notes:
Put those recently-acquired references to use on the body, this time around, and although the mistakes on the body weren't quite as awful as the mistakes on the head, they were still pretty big. Fortunately, this revision features:
<ul type="square">[*]More anatomically correct arms->hands and legs->feet
[*]A little bit of revision on the torso
[*]Some rescaling on the head to match the body
[*]A few extra loops on the arms/legs.
[/list]
In addition, I added more geometry to the nose, including proper nostrils, a little more bridge definition, and a good deal more depth/distinction from the cheek.
At this point in time, I'm leaving this open to critique while I spend a bit of time getting something up over at fex-online.com. Once that's done, it's texture time.
Media:
Pre-nose-work:
The face is too big for the head for an average male - the lower lids should be halfway down the hed structure. The nose is probably little too short, and the mouth is too big in all dimensions.
He seems a little sausage lipped - make sure those lips turn inside the mouth and don't simply get thinner at the corners .
It's hard to tell from the final shot, but I think the ears are too big, but certainly they are too low.
The chest is too skinny in profile for a generic person, and you might want to beef up the thighs a little bit.
He feet are not good - but you don't need toes at that polycount. Really, mitten feet for that polycount.
This model has had a marked improvement every time it has been posted.
[/ QUOTE ]
Thank you very much. I'll look into your critiques and hopefully we'll see another noticeable jump in quality with the next update. :}
I think I might keep some of the definition in the feet: although I probably ought to weld at least some of the smaller toes. Main reason I initially put so much detail into the feet is that the project is set in a somewhat accentuated fantasty style, and many characters will likely feature enlarged hands/feet, akin to the Fable series. Additionally, the project is based primarily in tropical/sub-tropical scenery, and characters may frequently be seen in sandals or barefoot, so just a little extra care is in order: even if that just means a separate "big toe".
I'll dig up some close-up refs for the feet (and the hands could probably use a little refinement, too). No sense in wasting polygons if I'm not gonna nail the form, eh?
Again: thank you very much for the encouragement and the critique.
Polycount: 2576. (Added a loop to torso, removed detail from toes)
Timestamps: This update: 12/21/06, 2.27am; Last update: 12/19/06, 6.18pm.
Especially needs critique in these areas: Anatomical accuracy, neutrality of form/personality for easier customization later, quality of topology for animation/deformation.
Suspicious areas: Ears, feet, hips, chest.
Notes:
Addressed Rick's critiques. Face has been shrunk and repositioned, ears have been shrunk and elevated, nose has been elongated, mouth has been shrunk and repositioned, lips have been turned inward at the edges. Chest has been buffed (just a little: I want to keep him lean and rely on the morphs to tone him out), thighs have been buffed (a little, once again), waist area just above the hips has been slightly buffed. Toes have been welded and dramatically simplified, feet have been edited to match references (if anyone knows of some superb foot references, that'd be awesome).
Again: thank you very much for the critique, Rick. :}
Anyway, still working on setting something basic up at the website: probably just going to stick with a blog for now, until I have some worthwhile assets to make a portfolio out of.
The mesh is down to tweaks and fixes like the ones Rick pointed out. I'm pretty eager to unwrap this and find out just how suitable Silo 2 is for normal map generation- as well as how quickly I can get the hang of hi-poly sculpting.
Media (click for wireframes):
GIF Turnaround (2.0MB) <--- Anyone know of some cheap/free image-sequence-to-video software? GIFs suck.
http://psyk0.cgsociety.org/gallery/