Wierd 3D modeling method?

melviso

Was just going through youtube to find modeling tut for 3dmax and I found this.This must be very hectic but does help when it comes to topology and getting the character's likeness headon:

[ame="http://www.youtube.com/watch?v=dxnSXBW1qwY"]3d anime: modeling head in 10minute - YouTube[/ame]

I am wondering if any of u have tried this method?Does it work in tems of speed?

pior

Nothing weird here!
(except the model itself hehe)

Bal

That's how I used to model when I started out heh.
Nowadays I don't really see the point in doing this kind of modeling for most things. It's so much more organic and responsive to just sculpt everything, and then do a clean retopo once the shapes are satisfying.

Ace-Angel

Bal wrote: »

That's how I used to model when I started out heh.
Nowadays I don't really see the point in doing this kind of modeling for most things. It's so much more organic and responsive to just sculpt everything, and then do a clean retopo once the shapes are satisfying.

Many guys in the Cinematic industry don't do that, they hate it. I think many peeps at Weta even refuse to sculpt anything unless they have a proper low-poly SD model on which they can add details only for later on displacement and rendering process.

cryrid

I am wondering if any of u have tried this method?Does it work in tems of speed?

I used to do that about 5 years ago. But I didn't have access to a sculpting tool back then, nor was I aware of the concept of retopologizing.

There was an older tutorial for modeling anime heads using that method which seemed popular enough at the time. I don't think it's faster today though, especially compared to sculpting. It can somewhat get the topology out of the way, but then you're left with a lot of verts to move into place without having any planes established. That might not be so bad for certain anime faces which seem flatter in nature to me anyway (especially around the eyes). It can also rely more on having orthographic references of the head you're making.

gsokol

The music from that video makes me want to jab a pencil in my ears.

Theres nothing wrong with that method if it works..but I don't usually see people work that way.

melviso

I have tried Mudbox & ZBrush and I am beginning to think I kinda prefer traditional modeling to ZBRushing or Mudboxing.I don't know why.I think its way faster and I kinda feel I have more control over my mesh especially with inserting edge loops where u need to and also when modeling stuff that u need to extrude from a mesh like the collar of a shirt,cables,ropes e.t.c.I know zspheres takes care of that but retopologing all those wires is gonna be work.

My personal project is being delayed by the whole high poly first thing,retopo which can be time consuming if model has complex details and painting the different maps in Mudbox.Then I found this site:
http://haikai.net/
From what I saw in the diffuse map models.This dude can paint diffuse maps so well,that they look really good.So he models the character and paints only diffuse maps.They look kinda like they have normal maps on them.So I am gonna use this style to be faster and the model will look way better after compositing in After Effects for the animation shots.

cryrid

My personal project is being delayed by the whole high poly first thing,retopo which can be time consuming if model has complex details

You're going to have to model them in either way. You can either worry about the form and topology at the same time, or hammer out the form first and then snap the topology to it.

What tools are you using for the retopology? If you're just building it face-by-face in zbrush then that's going to be a pretty slow process, but if you're using other tools then things like wires shouldn't take long at all.

melviso

Yep,I use the retopology tools in ZBrush,I have also used Topogun.Some of the problems with using these tools and retopology itself are double edges,manifold geo,double vertices.Sometimes,when u import the retopo mesh into Maya,the normals for some parts are reversed.I kept wondering why my normals from xnormal were so screwed up.Mental ray renders my normals inverted.Doing retopo for fingers is also a pain,lol.Getting vertices and edges between the fingers.There are a lot of problems and for sake of speed,hammering out the form and topology at the same time might sound difficult but with practice,should become a plaything.

Doing high poly which at the end of the day is gonna look like the supposed low poly is gonna take longer...I think.so rather work on the lowpoly fom scartch instead.Only downside is,normal map cannot be painted so u would need a high poly mesh with details.This workflow would be faster for me since there won't be normal maps.

maze

Ace-Angel wrote: »

Many guys in the Cinematic industry don't do that, they hate it. I think many peeps at Weta even refuse to sculpt anything unless they have a proper low-poly SD model on which they can add details only for later on displacement and rendering process.

Yes, or actually doing a concept in zbrush with zpheres or dynamesh, then creating a proper topology, getting it approved. And only then redoing all the detail in zbrush and extract disp, etc.

Mesh topology is really important in film/cinematics, needs to play nice with displacement. Also not all models need to be subdiv ready, there is a lot of low, and mid res modeling for a lot of shots as well depending on the camera closeness.