Do you need feedback or help for a CG project you're currently working on? This is the place to ask.
I decided to recycle two anatomical studies: the skull and a torso and make a character from my favorite cartoon when I was a kid.
Everything was done in Blender except the texturing in Substance Painter.
What do you think? I'm not fully convinced by the one with Grayskull castle as the background since it's a cartoon photo and the character is realistic.



Hey, it's looking really nice. I think the background fits well and the little rim light at the right shoulder completes the job. Also I think that red light makes a nice contrast with the main (ambient?) light, it emphasizes a lot the shapes of the skull. Honestly I prefer a lot more your first renders, I find the second ones more boring as there's less contrast and less color variation in my opinion.Β 

Good work!
Sebastian Beltramini replied
  ·  2 replies
Decided to use the skull study on an actual scene.

Finished the initial blockout:


I'm trying to reproduce this beautiful image:


Quick texture concepting using the initial blockout:


I'll need to work more on the mail chain. The current volume looks more like cloth than metal.
I think I'll also add some particles over it to make it look like something coming out of soil.
Ignore the lack of proper UV's for the base.
First actual render. Still work to do: integrate it better with the ground, add some rocks and dirt and the chainmail doesn't really look like one, so I'm thinking of adding a few actual rings over it.

Zacharias Reinhardt replied
  ·  5 replies
Guys, how is it looking now?

I completely changed the lighting and added a few rings scattered using particles.

I'm thinking about adding some rocks and volumetrics for the environment.

Zacharias Reinhardt replied
  ·  6 replies
Something I'm working on. Posted it on FB but trying to help this forum come alive.
I still need to adjust a few things I realized after posting it, such as the end of the zygomatic bone and the placement of the ear hole.

Initial block-out:

After this, I carved the mandible so I could separate it from the cranium using facesets.Β 
Though it was a lot more of work, this would allow me to pose it should I want to use the skull for a character or another sculpt that has the mouth more open, or even remove it altogether.


I kept working on secondary forms and slightly on tertiary until I got to this point.
I also added the teeth as it was difficult for me to gauge proportions and shape without them.


Hi hi, so a few questions are bubbling - its my first game character that im designing for use in unity- so really this is for people that have a bit of experience in the workflow related to that -

So some questions

A little list:
1) Do I need to flatten the mesh, so that it only uses one texture? or can parts remain separate, are there any guideline related to this? - such as optimal texture sizes, triangle count etc...

2) Unity doesn't do hair particles - are there any alternative ways of adding hair other than hair cards? (I was interested to do fur trim on another costume - so any ideas on how to tackle - while keeping low tri count would be great.

3)Should I model the mouth and teeth (I can pin the face verts and there is plenty of room left on the face texture sheet (I separated head and body texture - hope thats ok)

Oh this might sound dumb - but should I model the eyes as shut, and then do shape keys? I'm not sure if that's what is the done thing...(i'm just assuming)

Anything you would change about the design? does it look weird, did I miss that doesn't work well? (still have to make sword belt, arrow holder, pendant and belt with pocket/bags and the hair cards/hair)

Anyway thoughts appreciated
cheers

Natalie


Some updates:
Changes and additions to costume, probably ave to sculpt the leggings
Need to make the hair cards, belts/equipment straps and maybe some rope (as all good adventures need ;P)
at least found a tutorial for hair here:
https://www.youtube.com/watch?v=DRDAtcm2cUE

Hi Nat, I will try to answer a few questions, hope I can help.

  1. This actually depends on your target in my opinion. You can separate some parts if yuo want to prioritize texture quality and final look. Normally for games try to use only one texture because they prioritize optimization over look, I believe you can achieve this making one big UV unwrapping (2k+) and then resizing to lower the file size and trying to lose as less detail as possible. In example, modelsΒ  in Dota 2 are made of separated parts (arms, head, body, legs, etc...), this allow people to make high detailed parts using very low resolution textures, in LoL they usually try to use one UV set for their models. I think your model can fit one UV set and also you'll get rid of that color cut at the neck, If you have lack of space try to scale down the inner parts (not visible) in the UV (at least is what works for me). Also, as I think your model can be mirrored as the left and right side looks identically, this wil give you more space in the UV if you didn't do it of course. Check https://sketchfab.com/ybourykina's work, you'll see gets a lot of details on 1-3 uv sets (you can inspect textures in Thresh and Ahri models).

  2. Actually can't help much here, just seen people using cloth simulation for hairs. As it is a tail maybe you can use this youtube.com/watch?v=aPeb444dxnk&list=... If you're going to make animations and then export to Unity it might be useful, I would take a look to the cloth simulation, maybe is easier.
  3. I think only if you want to practice face rig actually

    These are some of my opinions and of course I can be wrong, then correct me. Hope is useful for everyone!


Natalie V replied
  ·  1 reply
Hi Pedro - yay thanks for awesome thoughts!
is great to have someone to bounce ideas off! so thanks for the detailed response it gives me a great start!

List of Fun stuff :D
Yes I will/ am mirroring the texture who can resist 50% savings eh :P I have only unwrapped the face so far I just wanted to test some colours.Β  (but I should probably extend that down to hide the neck cut -

Separated maps vs one draw-call....
I had planned on making the face/arms/hands as a map, then one for the costume, so parts can be swapped out - but yes maybe its better performance wise to be a bit more efficient for draw-calls to Join everything in a single Atlas. I wonder if there are any tools in blender for creating texture atlases... hmmm I did have Bake-tool - ill have to see if it still works with this new blender maybe it can do some magic on it :P
https://blendermarket.com/products/baketool

Some more thoughts on eyes...
I did a little look around some ripped play-station models, it seems like they keep the eyes open but more relaxed, so maybe ill have to do a little tweak to the current model. They also do a neat little trick with the eyes by adding an overlay to make more shadow from the eyelashes, (although im not 100% on how it works with rigging)


Revisions
they also avoid using hair, so maybe ill have to rethink my eyelashes since that will be a bit too many verts??? I also liked how they are doing the eyebrows as a diff and normals (why didn't I think of that!!) so I think between that and your ideas hopefully it will be a bit further down the track :)
So I guess its hair cards then ...eep time to figure that out then.




I did find and purchase Dynamic bones for the hair and costume (where it needs physics to stop it going into the mesh. its pretty easy to set up - much like blenders jiggle bone - But yes your idea is I think a nice way of going about it :)

I tried using it a while back before doing the sculpting course
an old model :PΒ 
https://www.youtube.com/watch?v=jNtG6wMRwno

Anyway please stay in touch its so useful to have somebody to talk to about this stuff :D Thanks so much!!

Cheers

Natalie





Hi,

I don't have professional game production experience but by doing personnels projects I struggled to make optimised models and found some workarounds. It really boils down to what will be the use of the model.

1. when I contribute models for 0 AD, an RTS antiquity based open source game, I need to comply to the design document and specs of their engine. Which implies low poly count, largest animals have 1000 polys, and small texture footprint (512px square is the norm for largest animals). So we flatten models, bake highpoly to low and then uv-unwrap with overlapped symmetric layout as to have the maximum resolution possible given the constraints. In your case, is it a third person type of game where you mostly see the back of the character, or a 3/4 top view like those rpg ala Baldur Gate and Divinity Original Sin, where you can't really zoom to much. Separating the head texture goes with the mouth and eyes modelling because it would be essential if there are close ups in game of the characters talking and making expressions. I don't know if the industry does a swap of textures when doing close-up/cinematic and then when playing the face has a lower texture. It would also help if it was the kind of game where the character can be customised, mostly with the skin colour.

2. For hair there is hair cards for a realistic look, but you can also do the curve technique and animate the hair with bones and then just bake and export the animation. Those two videos would help:Β 
- hair modeling : https://www.youtube.com/watch?v=t2XjdzzWCqI
- rigging and physics: https://www.youtube.com/watch?v=gTdRGZA4TXA

3. As Pedro said, it's really about face rigging or not. Does the model needs to express emotions and talk (well, more like does the player is going to see the model talk and act).

I would add that if it's to make a unity asset to sell or just distribute on the store, you can always make the two options. A high detail model with facial rig, and a low model for gameplay. There is also games that like to have the characters talk by just showing a bust while the screen keeps the same gameplay camera. That would make two models, one game ready low engine impact and a bust to do the dialogue scenes.

As for the model per se, at first I though the legs where a thinner than the rest, but on hindsight the shape and proportions match the overall character. I don't know if you already unwrap the clothes or if it's just a placeholder to get a feel but there is stretching on the thighs and the mail belt. The thigh the stretch is vertical, or it's intentional but feels like stretching. The specular of those nail's heads feels also more intense than it should. I don't know your workflow but I would sculpt/model the details of the metalic parts of the armor on the stomach and the upper legs then bake the normals to pop out the form a bit more. The chainmail belt is also flat instead of following the curve of the hips, it breaks the harmony of the model. And as another small silly detail, the chainmail seems to be under the leather nailed shirt. But the bust is bare and we don't see a fabric shirt or the chain mail. Does it mean that she has just a leather protection and half a chainmail shirt on the stomach? I know it's silly, but that what you see after taking to long on analysing the model.

And with all that, is the inspiration Astrid from How To Train your Dragon? She looks like the cold cousin on her mother side. :D
Natalie V replied
  ·  2 replies
Hello everyone!
l am currently working on this character and he is basically a plant Machine Hybrid that is throwing plant Bombs.

Altough this is just a quick clay render l feel like his Face is not drawing enough attention and I catch myself looking more at his arms and belt than his face. in the end his monocle will be glowing and he will be looking at a small plant on the ground, so maybe his face isn't really the problem, but I need to get rid of some of the detail on his body? What do you mink? Would be happy about some input :) I would also be very greatful about some tutorial tips about (procedural) shading in Blender, have a great day!

There is a lot more detail on his body vs his face, with a long distance between his body and face that doesn't contain any detail.Β  If you added details leading from his body up to his head would that help lead the eye to his head more?Β  Hmmmmm.Β  While modeling have you played around shortening his upper body so his head is closer to his lower body, even by just a bit, to see if makes his head any more noticeable if it's not so far removed from the more detailed parts?
Noah Bujnoch replied
  ·  3 replies
Perhaps you'd need to revisit the concept?
Maybe you could switch the widest part (currently in the lower area) with the narrower that's on top?
The face would be closer to the main focal point and then the eyes could go in the other directions of the figure.
Also, it could suggest a brain instead of jaw or mouth opposite to the original.
Something like this:
Noah Bujnoch replied
  ·  1 reply