I have been thinking recently of not doing any more zbrush stuff. It's a simple thing really in that most of my work of late, bar 1 project has been subd,no zbrush in sight( except for morph targets)
It takes me about 5 times longer to sculpt something than to model it.
TBH my zbrushing is ok, but It makes me annoyed at times don't even get me started on retopo grr yeah model something, sculpt it then spend days rebuilding it?
Modelling, then texturing without zbrush in the pipeline is sooo much simpler, but i am having trouble dragging myself away from Zbrush as i have invested so much time in it.
have you guys had these kind of doubts?
I am not going back to 1000 poly models with 256 textures, but was thinking of going for the more pixar style, not sure really.
I think that suits mi style better
Replies
I don't enjoy sculpting as a process. when I was a 2d artist i was mainly in to watercolours and did not enjoy mediums that were 'thick' like acrylic or oils. its the same with sub d vs zbrush in that i see subd as watercolour and zbrush as oils and acrylic.
It's also more fun to sculpt when the object in question is completely organic (i.e think mud) in which case, sculpting actually feels like the way to go.
This is for organics though. I still wouldn't use zbrush for hard surface stuff, other than a quick concept sculpt.
I think some people are more predisposed to be good at sculpting than say texturing and maybe I am just a texture artist who can model a bit.
I really need to resolve this in my head or I will end up just getting stuck in a loop and
not create anything.
CheeseOnToast - I have tried working like that, but find it just too time consuming overall and I am not really a 'sketcher'.
Organic = Sculpt.
A lot of stellar artists work like you. It's a normal process. Sculpt some wrinkle details and folds but get your base shapes done in modelling.
Theres still quite a bit of work to be found as a handpainted artist.
I personally love the pixar style myself (making a sub-D modelled side project game).
Im actually using zbrush as a sketch tool then remodelling it in sub-D when proportions are where i want them.
You can break it down, for e.g, 3 days on the main shapes in, then spend 3 days in zbrush and see how much progress you can do on the details with a time contraint. That way it forces you to be efficient and keep track of time which helps a lot becoming faster or better in areas you might feel you are less confortable . In the end, even if the character isn't polished as it could be, you will have gotten some xp out of it .
EDIT: I also feel the process is tedious but in the end just make sure you have fun and keep learning all the time, thats the most important for personal pieces imo. : )
- Online videos of people working in Zbrush are usually 4x timelapse or more (probably because they look cooler that way, but also because watching someone work at 1x can be a bit of a waste of time indeed). When played back at normal speed, even videos coming from "experts" can be excruciatingly slow to watch, and honestly kind of inefficient.
- People's art background are different. I personally don't feel like Zbrush/Mudbox are successful at emulating clay *at all*, as they feel more like working with toothpaste especially when there is no structurally sound basemesh used as a based to be detailed. For some reason I feel like manipulating polygons is much closer to real sculpting, probably because flat polygons have a definite structure and direction to them - just like a slab of clay.
- And lastly, I think that expectations for time and efficiency vary greatly from one individual to the next. Some are totally fine sitting through a painful process (like doing retopo tasks in Zbrush instead of using a dedicated program for that), while others just can't. It's actually quite fascinating how personalities vary in that regard !
That being said, when collaborating with other artists I have no issue with them tackling a modeling tasks one way or another, as long as the results are good and timely. However, if a particular process makes a given modeling task either inaccurate or too slow, then things should be addressed.
I hope that makes sense somehow !
I think inefficiencies tend to arise when artists stick to a specific program just for the sake of avoiding to export out a model. Now of course I would love it if things were 100% unified ... but these days are long gone now (or maybe, far in the future !)
However I do grant the fact that working with polygonal modeling does require more tool research and being proactive with one's workflow optimisation - something that in and out itself can take years. That sure is one of the severe drawbacks of that approach.
Well, I used to model them only with "image planes" poly by poly, when I started in December 2012, but soon after I found the world of sculpting in Sculptris... that was the biggest relief ever. And soon I switched to ZBrush, and there's no going back.
I love sculpting, but sometimes it feels a bit frustrating adjusting the anatomy and proportions (for too long) when you you'd really like to move on to clothing and hair. But I still want to sculpt the underlaying anatomy properly before anything else. I'm an anatomy snob, kinda. Always finding errors and areas for improvement from my sculpts...
For me the retopology/modeling, UV unwrapping and texturing is more faster than sculpting. It's just that anatomy and folds are slowing me down. Even though it feels frustrating from time to time, I still enjoy the sculpting process.
I never felt comfortable using Zbrush. I hate the interface so much. This is why I began using Mudbox. Mudbox felt more fluid to me but even then I couldn't help but feel that I was wasting a lot of time on details that wouldn't even appear in the normal map.
It's easy to get carried away in those packages and forget the overall focus of your asset.
Pior i agree about sculpting programs not really emulating real clay. I mean who uses a brush to sculpt with
Joking aside I feel it would be nice if you could have some hybrid method where you could paint subd shapes and, have them automatically merge in to the surrounding topology.would be nice for cloth for example and would not need a retopo stage ie WYSIWYG
or have a big library of common cloth shapes that you could just add to your mesh with some clever editor.
I thought about trying it in max using redndered splines on top of a base mesh
( this would be for doing cloth folds), but I am not really tech enough and its probabyl not even going to work:/
Not really enjoying cloth simulators stil, as they are too hard to control, but for certain situations they are ok.
I have mentioned cloth a lot here as I seem to do mainly cloth in zbrush these days, but my point still applies to heads and other things.
I find the disconnnect between the final in game mesh and the original base mesh disconcerting and too convoluted.
I have seen people sculpt detailed cloth then do an amazingly detailed retopo, which could virtually be on its own without any additional normal maps.
I am thinking what is the point of that ie virtually doing it twice. what if you have to make corrections at the last minute?
Its like he is only sculpting this to get the base for the retopo
which looks great, but there is not too much need for a normal map on some of the cloth bits
This one is more obvious. It would make a great subd model in its own right
My overall point is that building something and subdividing it looks much nicer than a lumpy cloth/head model and in terms of speed is much quicker and nicer to look at IMHO. I would not like to build subd cloth from scratch. I have tried it and its just not viable for me time wise( and artistically)
I might try and approach it a bit different ie if I was doing a shirt for example, quickly block out the folds with basic forms and then retopo it straight away with the intention of further deveoping the retopo version( maybe just in the 3d prigram)
It might be more what I am looking for and i do think that maybe I am more suited for the disney style of work, where the fine detail is minimal and the focus is more on silhouettes and materials.
The closest I have got to nice looking final mesh from zbrush is when I decimate a high poly mesh down to about 2000 polys( see below). It picks out the main shapes/silhouettes brilliantly, but outputs tris. I would still have to do a retopo
I love the minimalist/faceted look of this, but with good retopo and and 1 subdivision it would look ok , in that the superfluous lumpy shapes have gone leaving clean recognisable shapes like a structure or framework.
I nee a new tool called 'topo painter' which paints directional topolgy and automatically and blends in between the main shapes with nice even quads lol
what i really love about the sculpting workflow is that it's less rigid and allows me to kind of "free flow" any ideas i might have.
i find it easier to expand on those ideas in zbrush than i would in max where i have to consider topology and subdivision consequences of any changes.
i am improving on the actual sculpting side, it's getting an end result that I am happy with at the end of it all.
skankerzero - yeah wish I could just do everything in max, would make things a bit easier. zbrush is just so great for doing facial morph targets though, that's one aspect I really enjoy
I did try Mudbox, but it was a bit slow on my machine
This kinda left me scratching my head. Sub-d is great but for anatomy/cloth zbrush is a very powerful tool. I learned heads/anatomy/cloth in sub-d and although its my subjective opinion, zbrush is just so far ahead of trying to do this stuff in sub-d its not even funny. I couldn't even imagine going back to that very slow, much less results focused workflow. - as for 'lumpy', not sure why you threw that in there, it only looks lumpy if its poorly done, and people do just as much poor work in sub-d as sculpting.
As for the rest of it, if you feel more comfortable doing sub-d then do it. If the quality is there and you can do it without losing any speed then there is no reason not too. But I highly suspect that for at least realistic characters you're going to see a pretty steep dropoff in time and quality. Beyond that, if a client says to me, "its generally nice but can you modify x" - it takes me like 10 seconds to smooth out that area and redo it, whereas with sub-d its much more difficult to modify.
Personally I like both sub-d and zbrush - again I learned cg my first couple years in sub-d so I am definitely not afraid of it. I do quick sub-d for specific parts then pull them into zbrush and sculpt on them. Same for hard-surface - sub-d then if I need an extra push I bring it into zbrush for booleans and IMM and wear etc.
And as for retopo, its just another way to work. I do it that way because I think I get a better game-res mesh when I know where my detail is and what my finalized surface looks like, and I find it easier/more efficient/faster than trying to modify the sub-d or basemesh. Even when I'm making a basemesh from something I did in sub-d I tend to make the game-res from scratch unless its like true hard-surface, like a gun for instance. This goes double for heads/clothes.
Last - clothing sims. I think they're more powerful than you're giving them credit for, but that might be because you haven't spent much time with it. You don't have to do the whole mesh in sim, just do something super basic and use it as a base-mesh, that has been my go-to workflow for about two years now and its much more efficient for me at least.
with sub d I never intended to try and build realistic cloth with it, it just takes too long, but there is the rub, do i want to do more creative stuff with less detail, or focus on hyper detailed stuff.
I get the same money for both , but one takes me 5 times longer.
seems like a no brainer, but i am finding it hard giving up on stuff, just because it's hard:)
most of my clients want 'less' detail than I generally put in, so modification it's not really a problem with subd , it's the overall shapes they tend to worry about and they are easy to change.
i tend to use whatever suits the task at hand, but in the last five years nearly all my 'games' work has been non zbrush, so i am thinking maybe its time to stop practising this stuff so much in my spare time
texturing is my main thing really, so basic models with nice texture should maybe become more of my focus.
Yeah if you're going to compare the workflow for something with no real detail to a much more complicated model then sure it's kind of a no brainer that one is faster.
But if we're talking about more modern realistic characters than it's not even a competition, you simply can't get the same results from sub-d and certainly not in the same timeframe.
Plus pbr is much more reliant on normal maps for detail than since you can't just stick all that faked lighting into the diffuse anymore, making the higher res and mid range details you get out of a zbrush sculpt that much more important.
I think it is also handy to start off modeling using subd to get the feel of edgeflow which makes retopoing less of a chore. I dont do everything in Z but I like to force myself to try things out as it often makes you look at your process from another angle. It can be inspiring when you are making stuff for yourself.
some people maybe consider themselves as digital sculptors, but i don't, it's just another tool to help my workflow.
Kanga - yeah I think maybe I have to work a bit looser with z, i just blocking out and focusing on the main shapes, even doing a quick retopo on test areas of my mesh, see how that is working.
I love the cleaness of sub d, but also can't totally abandon sculpting where it helps me work a bit quicker.
maybe just answering my own questions here, but it helps to sound things out sometimes.
I might even make a library of commom cloth shapes which I can quickly add to my subd work.
As long you do not change major shapes, the lowpoly, UVs etc do not require any change, fire and forget.
Have you tried marvellous designer for clothing ? It requires retopo but you can get a good base for many things very quickly.
i am aiming to just get the main forms down in z, then build a subd version over that
I do use marvelous designer and although its quite nice, it has that look doesn't it?
I often use max cloth to get a base also, but its not as good as MD by a long shot
anyway I am going to experimient some more, and see what i can come up with