Agree that sculpting and retopologing is waste of time? Why not model it straigt?

I dont see the differense in works beetween people who do sculpting and retopologing and those who skips such steps and starts modelling it straight away. I think its just wasting of time. Very often i saw that what they have sculpted was not even used or just partially used so during the modelling stage they completely remade the look of the character which had no similarities with what was sculpted before. Or sometimes they sculpt so easy meshes to do like for example cartoonish bird. Why would someone spend 5 - 6 hours for sculpting and retopologing when you can model such a simple character in 1 hour time.

So what do you think?

1 Like

Because creating the base mesh as a retopo step rather than from scratch is a far easier process (a bit like connecting dots in that the form to follow is already there).

I usually use the shrinkwrap modifier for retopoing, you can also use a subsurf modifier on top of that to ensure an all-quad mesh when you’re done and the modifiers are applied. I guess it all depends on preferences though, I personally find it easier by way of the sculpt -> retopo process myself since I can easily chart out the topology and the tweaking required becomes minimal.

Dyntopo sculpting in Blender is really a great modeling tool, but I hate retopo. The other day I had the opportunity to try ZRemesher, it’s fantastic. If you do not have money to buy it, for now instant-meshes is the closest thing we have in Open-Source (but the result you get with it is pretty poor compared to that obtained with ZRemesher).

Sculpting and retopo is a top down design approach. You start with basic shapes and slowly refine them. For complex organic shapes it’s pretty much your only option.

Extrusion modeling is bottom up, you just start. If you have the drawing skills you can sketch ortho references and use that, but otherwise you’re likely to end up with some pretty funky proportions and shapes in your model, because you’re always kind of guessing at sizes.

Silverland, you’re actually very wrong. It’s not a waste of time. Many things would be better of sculpted first and then retopologized after. You can add finicky details on sculpt which you can’t do on low poly model and that detail could also be baked to the “low poly” version.

For certain kinds of sculpts, it’s easier to quickly model a base mesh and then use multi-res, but in general this is true.

I think it’s very much a matter of preferences and familiarity with tools. I do mostly naturalistic human figures, which are complex organic shapes, and I’ve never started with sculpting, nor with ortho plans, nor do I guess at proportions & sizes. My pre-digital training (painting & sculpture) informs my 3D modeling, of course, which others may not be able to claim, but it’s not a matter of single available options by any means. I’m much more comfortable with modeling->multires->sculpting as a workflow, so for me it works better than others.

Ok. See that most of you or all of you using the method that i dont like. So far i was using only referense images and was happy with that but maybe i will reconsider now this things and will try something that i hate to do.

For me dyntopo scultping is total freedom in 3-d artistry. I can crank out a mutant t-rex/mermaid/unicorn head in minutes, you just can make whatever the hell your mind desires. Whereas traditional modeling is a bit more technical and boring. I do both all the time, so it really is just whatever works for you. There is no right answer.

If you don’t like using sculpting to retopology and still get what you want at the end of a project, then don’t use it. We’re talking about a creative visual process. Essentially the end product is what you’re after. While certain techniques are well established, they’re still just a means to an end.

That doesn’t mean you can’t learn from something you don’t like though. :slight_smile:

I totally agree with Photox, dyntopo is total freedom, and I use both sculpting and modeling all the time too, but just one or the other, and I often use also a third option, that is sculpting the mesh as is after a fast building of the topology, with neither dyntopo nor multires, and I find that very handy and pleasant.

And I totally agree with Silverland too, I also hate to retopo.


1 Like

It’s funny how these things go and how something becomes a standardized methodology overnight. And suddenly it will seem like the only right way and like everyone is doing it. I think this is obviously due to CGI still being so new and evolving. Also that the tools are constantly evolving so fast it can feel so much like shifting sand. So it’s easy to become locked into a certain methodology because it can help give a sense of certainty. And we are all mostly guilty of it. It’s inevitable and human nature. There is also the fact that so much of this work is so complex too. There are so many things to know and keep track of.
But I think creative working methods are always up to the individuals and that working processes should never be set in stone for anything. Especially in creative work. What works for somebody else is not necessarily best for another.

Recently I’ve really come back big time to old fashioned classic box modeling. Especially for starting a character or any organic form that needs to be animated or rigged. This can always be upressed and used as a sculpting base for more refinement. Then later used as a starting base for the re topo.

The most important reason I realize that I’ve come to prefer this method is simply the precision and clarity. Modeling and designing a character for deformation and animation is just not at all like working on a sculpture for me. The process just feels very different. Animated characters also need clear line and form even if you are working on something more naturalistic or realist.

After the initial idea or sketches, It’s always got to be a solid design process. I find I often like to box model in separate sections and be able to pivot and move them around. Even attach them to a full FK armature. So that the form is always fluid and moving and I can always check how it would change shape in motion. Check clearly if the legs are the right length etc. Which is alway tricky however many characters anyone may have modeled.
I sort of see it as the CG equivalent of always keeping the clay fluid while sculpting traditionally. Making sure to always be working the whole form. Just because it’s cutting poly’s and pulling vertices doesn’t make it less organic in feeling for me.
There is also the fact that so often working with traditional tools there is the need to try and picture and analyze organic forms as solid clear shapes. Well in computer box modeling I always thought you are sort of working back the other way from that.

An excellent distinction & probably why I have always worked this way. For me, animation is a principal goal, so I work with that goal foremost in mind. I also think my “style” of 3D art (such as it may be) lends itself more to well-defined forms & volumes rather than miniscule detail, so modeling keeps me more honest in that regard.

You have a lot more freedom when sculpting, and sculpting is a lot easier than polymodelling from scratch. also polymodelling and maintaining a proper topology is a lot harder than retopo. sculpting is the way to go. i was in the same mindset as you before, but ive realized that sculpting is the future.

Doing a Dyntopo sculpt followed by retopology with shrinkwrap + subsurf will also get you an animation-ready mesh that is completely quad based. This can be followed by traditional modeling techniques on the new mesh (which is now easier because the form has already been established).

Though I do agree that box modeling from scratch in Blender would be far more practical, faster, and easier in 2.7x than it was when dyntopo was first introduced (noting the major enhancements to the toolset since then).

It depends first on your personal preference, but also has a lot to do with the project, so the workflow may change from one to another.

For example, for doing something like this you’re probably not going to need sculpt and retopo, simple box modeling will be enough to create such a model.

But, if you’re doing something like this traditional box modeling will be the slowest and most painful approach. For that kind of models is better to go with the “sculpt + retopo + bake textures” method.

I personally find that for complex, organic models is way better to start from a sculpt, it feels more natural; but for hard surface or inanimate objects, box modeling is still the best method. I haven’t seen anyone doing a sculpt of a car for example…
Also, with dyntopo and the recently improved vertex paint tools (see GSOC 2016 projects), you don’t even have to do retopo for a model that won’t be animated. If is going to be a still you just have to sculpt, paint and render.

So the retopology process is only needed when you need a mesh usable for animation and/or games.

Hi Silverland.

I think the message of this thread is be interested in all ways of doing things and always be open to trying new things.
But work which ever way feels right and most comfortable to you for the project in hand. There is no one way for everything.

I have to say that with re topo in Blender and reading Ace Dragon’s post above. Shrinkwrap and subsurf is a really brilliant way to go. I’ve used it several times and love working that way. It means you just need to block out the main poly flow as a low poly mesh. Very easy to work out the general flow and edge loops that way. The subsurf adds the rest. Then for the final detail bits like edges of lips and eyes. You just collapse down the mesh, re apply a shrink wrap and then work with what the subsurf has already given you pretty much for free.

What I like most about re topo in Blender is the fact that you are always still in full poly modeling mode with full use of all the tools and work arounds. If it’s a character for animation then re topo is always so much more than just drawing polys onto a sculpture.