Topology research

#721

And believe me, if you spent days with careful modeling of a character with excellent topology, and you need to change something…it’s a nightmare.

   Then why not import your excellent topology into Mudbox/Zbrush and make the changes needed. That's how I work. I import my "perfect" mesh with "wrong" forms, subdivide some, brush away till I'm happy, go back to level zero, export and voila a perfect mesh with a perfect shape.

A clay sculpture of the creature is created, then it is scanned, and the modeler created the model with proper topology. Here you are the sculptor, and the modeler as well.

   That has nothing to do with work flow but all with clay modelers not being able to work with computers in those days. The model scanned by a technician (not a modeler) was the solution.
   
   Another assumption I more or less read is that the sculpt was a concept. i think it was more an end product to visualize the 2D designs to get a final approval and to be used by 2D animators to study the character.
   
   There is a lot of people around here (like me) that can model and sculpt but there is just a handful around here that can design a character. These people (the designers) don't even need a sculpting tool ...a simple pencil and some paper is enough. Once the designers are done with the designs there really is no need to start concept modeling nor concept sculpting. It's the task of modelers (in case the designer is not also a modeler) to translate the designs as good as possible into 3D characters. The route to take is up to the modeler.
   
   Like described I like to plot my poly-loops as early as possible and only roughly approximate the shape so I can later fine tune the finished mesh in Mudbox.
   
   This way of working is even possible when there is no real consensus yet about the design.

Not one polygon needed to be replaced nor did it hinder while sculpting. I totally agree it’s about form in the end …but there are more ways that lead to Rome.

#722

I understand what your saying Jester.

I have the same problem. Tyring to juggle good Topo and form at the same time. Sculpting the figure first allows the artist to focus purely on form and the character of the figure. Instead of being tied down by Topo rules.

Then when the sculpt is ready the modeler can focus on creating the topo useing what ever technique he feels is suitable for the task. For example if I wanted to have a super smooth model I would use Nurbs to begin my re topo tast than convert to polys. It’s a common method used for automotive models, but, i have found it to work on figures as well if you want really smooth faces or what not.

Anyway, I am going to try your method you mentioned. I e-mailed you about this already Jester but wanted to just make a post as well.

Thanks,
Ryan

#723

Bunk, very nice models.

Every road lead to Rome.

What I said that if you concentrate first on the form, and later on the topology you may gain time.

It takes few hours to create a perfect, detailed head by sculpting. It takes less then a half day to build the topology. Within one and a half day I have the cage of the head and the high polygonal model as well. And believe me, if you have a wrongly placed wrinkle it’s much easier to sculpt it, than change the topology, and take care of the form.

But it’s my, and that’s your way, and that’s why the world is interesting :slight_smile:

#724

Bunk, very nice models.

thanks …but it’s one and the same model. I tried to illustrate that this method even works when no proper design was done first :wink: Starts left up ends right down.

What I said that if you concentrate first on the form, and later on the topology you may gain time.

It takes few hours to create a perfect, detailed head by sculpting.

Wow you’re fast. I can use that amount of time just to create the shape of an eyelid or nose wing before I’m satisfied.

It takes less then a half day to build the topology.

Same here

Within one and a half day I have the cage of the head and the high polygonal model as well.

Me too since it’s the same mesh, further more I don’t have to bring the mesh back in for the creation of displacement maps as you will have to do. The shape of your topologized mesh will change because of that so you will have to export the low level as your new base mesh. Then you 'll have to relax your UV once more to accommodate the new shape (a step I can’t escape from either ;).

And believe me, if you have a wrongly placed wrinkle it’s much easier to sculpt it, than change the topology, and take care of the form.

Hold on. Wrinkle details come from displacement maps and if not I only add it after I’m done sculpting. So I only create the basic loops and add details (if necessary) after sculpting.

#725

Sorry pal, you don’t get my point. So this debate leads nowhere.

you said:“Me too since it’s the same mesh, further more I don’t have to bring the mesh back in for the creation of displacement maps as you will have to do. The shape of your topologized mesh will change because of that so you will have to export the low level as your new base mesh. Then you 'll have to relax your UV once more to accommodate the new shape (a step I can’t escape from either ;).”

That shows me pretty well, that you have no clue of my workflow.

  1. I sculpt, mostly from ZSpheres, rarely some basic geometry is built in XSI and used for sculpting
  2. When the sculpt is approved, I spend a day or so to build up an evenly distributed, all quad topology. I don’t make difference between animation and still. An edge looped model will work for both, if I follow the outlines.
  3. When the control mesh is ready, I make the UV set for it. Few hours at most, within a day usually, but in extreme cases it can take two days
  4. Reproject the mesh back to the new topology (ten minutes at most).
  5. This mesh will follow perfectly the details I sculpted. Displacement in ZBrush within no time.

For you your workflow is better, for me mine. So there’s no need to debate on it.

#726

Sorry for being a retard.

I understand every step here except this Jester…

  1. Reproject the mesh back to the new topology (ten minutes at most).

Reproject the new mesh?

I think I know how to do that but, maybe I am wrong. Would you mind going in a bit more depth about that?

#727

j3st3r I’m not discussing your work flow. Since there is nothing wrong with it.
I’m discussing the general advice: “Sculpt first, then build the topology.”
The claim entirely depends on what is modeled imho. If it’s some obscure character then yes it’s the fastest route.
If it’s just another head I don’t see why starting from a base mesh with the loops in place should limit the artistic freedom. If you can sculpt a blob or whatever into a character you can certainly sculpt that character from the mesh at hand.
Advantage is that the low-res mesh is auto shaped into the form of the High-res mesh. Meaning a lot of retopogizing that doesn’t need to be done. I think that -in those cases- adjusting the mesh is faster then the retopo route from scratch.

#728

The key sentence is there: if it’s just a head.

Anyway, I prefer doing even heads from the scratch. I am not bound to the topology, I can work more freely. And I can retopologize it in short time. That’s good for practicing, experimenting. What I said, that it’s easier (to me) to sculpt hte character first, then retopo it. I’m not against using simple geometries (like basemeshes) during sculpting process. But I’m sure that sculpting is faster than edge loop modeling, and easier to correct mistakes.

#729

Just another thought: Jester, have you ever thought about shrinkwrapping a base model on a sculpt instead of retopologizing? Of course you’ll be left with a lowres version of your sculpt, but you can bake the normal map back on the lowres version.

#730

ZBrush has a feature (as well as topogun has) where you transfer your sculpted details onto your new geometry

toontje, believe me I did. Rebuilding the topology is faster. What you forget, that ZB and MB change all subdiv levels of geometry when you sculpt. A stroke on the highest level affects the lowers also, and vica versa. So your base mesh has no worth. Second, a muscular man needs a totally different topology than a smooth babe. Shrinkwrapping is not the solution here.

There were a video by Gnomon (God bless them) where the artist, Zac Petroc used that shrinkwrapping method. That method inspired ZBrush retopology tools, and Topogun as well. Retopology is one step beyond shrinkwrapping. Formerly I used your methods, I tried all of them. To me, with my attitude this one is the more effective. I don’t understand why you are so strongly against it, although it’s almost obvious that you hadn’t try it? I suggest to all doubter to give that method a try

I am working in 3D insdustry since 1998, and I was always an explorer. Whenever I read something that was interesting, I tried it. That’s not different. I gave a try to sculpting, now I cannot imagine to work other way. To me (and it’s strictly to me) that method is a more creative way of working.

#731

This thread has certainly picked up steam in the last few days :eek:

#732

Yeah… the last time there was so much discussion about modeling tech, we’ve been talking about how subdivs will replace NURBS :wink:

#733

Jester thank you for your advices. I will do the mental switch and make a couple of methods according to your workflow. I hope you don’t take offense if I appear headstrong, I was just looking for that solid POV/ sales pitch. :slight_smile:

#734

Jester, I am not questioning your workflow and I never said I was against it. I was just wondering about this step bcause I wasn’t sure exactly what it was.

Anyway
thanks for your comments on this.

#735

My wife wants to know why I need to learn how to paint to do 3D modeling.

So my question is:

Do I need to know how to paint in order to be a good digital artist?

#736

Art requires the same skills and a lot of similar knowledge regardless of the tools. Anatomy, composition, color theory, a trained eye and so on.

You can develop these while practicing on a computer with 3D graphics, but it is a lot faster and easier with a pencil or even with paint (or Photoshop and a tablet, although that will require some extra effort to learn the tools).
Formal education can help a lot here, but you can learn a lot on the internet as well. But you can only develop your sight through practice.

So you do not need to know how to use oil paint and canvas, but it is absolutely required to develop artistic skills somehow. Otherwise your stuff won’t look good.

There’s a lot of threads here on this subject, by the way…

#737

No two people have the same definition of what good is. And very few people do art the same way.

#738

My initial reaction to the question is, “Hell no!” because painting is such a physical form of art, and the initial skill set that is required to be of any use to cg art isn’t very related.
Composition and color theory are definitely areas of overlap though. If you consider zbrush, then it makes a lot of sense that an understanding of painting will be helpful. You might relate to this analogy; 3d modeling is like practicing law, and painting is like acting. They both require that the individual adopts a role, and the characteristics of each have cross-over skills. But because 3d modeling, like the legal field, can be soooo technical, arduous, and esoteric, it doesn’t make sense to trade direct education in the field for an indirect, somewhat related, education in painting. So the analogy is that a student of law would be wasting precious time if he were to take acting lessons, even though there may be overlap.
All that said; the two fields of art go hand in hand concerning the appreciation of composition and art in general. And when you find yourself in the desirable position that would require color and shape compositions in scenes or production level conceptual design, you may learn that your creative process has been fostered by your ability to paint. When considering concept design, nothing is more efficient than sketching.

Also, if you feel that you need to immerse yourself in related areas like painting or sculpting in order to fully understand concepts of modeling, then it makes sense to do so. A very specific knowledge of your toolset is the most important thing to someone who is learning the craft, that means understanding the application’s limitations and possibilties, and having an imprinted map in your brain of what the programs can do and how to do it; you will be sacrificing some of this understanding when you spend time painting, but you may gain things like “appreciation”, “understanding”, “well-roundedness”, “open creativity”.

I know that my experiences with non-digital mediums of art have contributed in one way or another, but at the end of the day I think that it comes down to learning how to paint with the program (so to speak)

#739

No problem, guys! Religious debates are always hot :slight_smile:

The painting skill helps a lot. Not necessary the technique, but the theory behind it.

#740

In order to keep the image loading to a minimum on this thread I have moved my various images to a single updated location further in this thread.