Fellow Environment artists? Is Zbrush/Mudbox really worth the effort?
polycounter lvl 10
Hi All
Just wanted to ask other game environment artists about your thoughts on using zbrush ,mudbox ,etc as a primary tool for creating textures for environments. I`m asking because i`ve been working in the field for about 20 years and i`m still not convinced about the workflow of zbrush,etc in creating textures. Yes I realise that the end results can be realistic, but is it worth the effort in a game production environment? I`m not convinced.
So often I feel that whilst working on something in zbrush i`m wasting my valuable time doing the ground work and minute details , whereas i can knock up a texture in photoshop in a fraction of the time. Yes I realize photo sources vary ,lighting can be difficult to fix and normal maps wont always be as accurate, but if the end result looks convincing why bother spending hours producing something from scratch in zbrush?
For example I watched a tutorial earlier on how to make flagstones. He modelled the stones in maya on top of a source texture of flagstones, brought them into zbrush, then went through each stone and painted height details. Then added a grout layer, detailed that. Then went through texturing them all. And all I kept thinking was why not just bring your source texture of flagstones into photoshop, spend some time masking out the stones and then create different layers for the different maps. The tutorial must have lasted over an hour, it couldve taken 20mins in photoshop and crazybump.
Maybe (as my username suggests) i havent grasped zbrush fully , which is hampering the speed of my workflow. But working freelance I am always working to tight deadlines and feel its impossible to get things done with zbrush in the time I can do using good old fashion texture maps and photoshop. I really want to use it because i know the result will be correct, but when a gamer is running around a level would they really notice or care whether a normal or specular map might not be 100% accurate.
So basically, i`m asking am i the only one who thinks this way? or should join the bandwagon and put more time and effort into learning zbrush.
Thanks
Just wanted to ask other game environment artists about your thoughts on using zbrush ,mudbox ,etc as a primary tool for creating textures for environments. I`m asking because i`ve been working in the field for about 20 years and i`m still not convinced about the workflow of zbrush,etc in creating textures. Yes I realise that the end results can be realistic, but is it worth the effort in a game production environment? I`m not convinced.
So often I feel that whilst working on something in zbrush i`m wasting my valuable time doing the ground work and minute details , whereas i can knock up a texture in photoshop in a fraction of the time. Yes I realize photo sources vary ,lighting can be difficult to fix and normal maps wont always be as accurate, but if the end result looks convincing why bother spending hours producing something from scratch in zbrush?
For example I watched a tutorial earlier on how to make flagstones. He modelled the stones in maya on top of a source texture of flagstones, brought them into zbrush, then went through each stone and painted height details. Then added a grout layer, detailed that. Then went through texturing them all. And all I kept thinking was why not just bring your source texture of flagstones into photoshop, spend some time masking out the stones and then create different layers for the different maps. The tutorial must have lasted over an hour, it couldve taken 20mins in photoshop and crazybump.
Maybe (as my username suggests) i havent grasped zbrush fully , which is hampering the speed of my workflow. But working freelance I am always working to tight deadlines and feel its impossible to get things done with zbrush in the time I can do using good old fashion texture maps and photoshop. I really want to use it because i know the result will be correct, but when a gamer is running around a level would they really notice or care whether a normal or specular map might not be 100% accurate.
So basically, i`m asking am i the only one who thinks this way? or should join the bandwagon and put more time and effort into learning zbrush.
Thanks
Replies
Everyone has their own taste of what they prefer more / what is easier for them, but in my case, I can say without a doubt that the tools inherent in ZBrush and Mudbox have allowed me to do amazing things, and easier at that.
You seem to be focused on the aspect of "Accuracy." While important, it's good to know that accuracy really isn't of much concern in the first place, especially being a 3D game artist. Being completely representational to real life doesn't work out in most case scenarios when it comes to game art, especially with such technical things as normal maps and the like.
It's funny, because I prefer Z "BRUSH" to all my sculpting needs, and Mudbox to all my painting needs....you think by their names, it'd be vice-versa xD
Anyways, to the main point; Is it worth it? Absolutely.
You may find the UI of ZBrush daunting, but so does everyone who's had a look at it lol.
You mentioned crazybump as something you use in your workflow. Have you had a go at nDo2? It is better in every conceivable way. You can spend less time in ZBrush now, and actually do all the detail work with nDo2, right in Photoshop. If you ask me, THAT sounds like the solution you may be looking for.
But here's the kicker... the bigger environment art pieces can get away with and should also have more detail. Zbrush gives you the fine detail, not just the minute. While its not the best painting applications (mudbox wins that one) is is great for generating those all important normal maps.
If you are serious about environment art, much less any game art (outside of dynamics) then you should give the sculpting workflow a shot. In my opinion it is much better then just generating normal maps from photos and doing it all in photoshop.
Zbrush is hard to get into, but once you do get into it and feel comfortable...it will reward you in many ways. Mudbox, well its just fun to paint in that app. You can also paint in zbrush but should use Xnormal to project the paintjob onto a low poly.
Doing this will allow you to produce both quality and quantity depending on a the time and asset type.
Maybe you simply need to take the time to look at good, recent games using such kind of assets to grasp a little better what is the actual point of the workflow ? Bulletstorm and Darksiders 2 would be good places to start.
http://images.wikia.com/bulletstorm/images/e/e2/Venus_maneater.jpg
http://www.google.com/imgres?start=10&num=10&um=1&hl=en&safe=off&biw=1024&bih=543&tbs=imgo:1,isz:l&tbm=isch&tbnid=e6yC-S0SfrOr9M:&imgrefurl=http://gamentrain.com/darksiders-2-rewards-players-with-bonuses-from-its-predecessor/&docid=-JJws0q3O0D41M&imgurl=http://gamentrain.com/blog/wp-content/uploads/2012/08/Darksiders_II_Darksiders_Unlocks1.jpg&w=1920&h=1080&ei=PMUpULeDFNOHyQHsnoCoDA&zoom=1&iact=hc&vpx=469&vpy=38&dur=282&hovh=168&hovw=300&tx=205&ty=76&sig=114046126784431069449&page=2&tbnh=125&tbnw=224&ndsp=12&ved=1t:429,r:6,s:10,i:25
Also for the sake of the discussion, it would be nice to see more precise examples of situations where you considered that sculpting was not necessary. Because indeed, some time it isn't.
However... sometimes ... it just is :
http://orbart.free.fr/index.php?Gallery=105
In general if the best people out there use a specific workflow, it might just be wise to learn it too
I can understand in some situations that zbrush/mudbox are the best solution ie carved walls, doors,pillars, etc and object detailing but general floor and wall textures I`m still not convinced. As an example, here is a video from zbrush that I thought would be quicker done in photoshop/crazy bump.
http://www.zbrushcentral.com/showthread.php?34651-Pixologic-Release-Learning-Series-Tileable-Textures
But like a couple of you said, you have to pick the tool for the job, sometimes its quicker in photoshop other times zbrush or maya.
I agree that Zbrush needs some practice to get around the interface, I did try the mudbox trial and felt alot more comfortable with it and it was great for painting. But the sculpting couldnt compete with zbrush.
Maybe i`ll just have to spend a week or two of downtime from work and learn zbrush, away from production to nail the interface and workflow as I think that is what is putting me off using it in production. Anyone have some good environment based tutorials for zbrush they could point me too?
Thanks
http://wiki.polycount.com/EnvironmentSculpting#Sculpting_Tilable_Maps
It might help to think of it as being just like painting a texture in Photoshop. Start rough, refine, until you get the result you want. The only difference is you're generating a 3d texture. You can use photosource in Zbrush, as custom alphas. The process is basically the same, 2d or 3d, and you're in control.
Actually,I`ve just got the mudbox 2013 trial and there are some quite useful new tools in there since i last checked it out. And from a Maya user its quite straight forward to use. So i`m going to spend a few days trialling them both.
Technically, you can get away with just Mudbox, which is much much much easier to learn. It can sculpt, get you the basic details and such...enough for normal map generation. Zbrush gives you much more detail, much higher poly and a lot of neat tools/features including but not limited to zpheres and a form of auto retopo. In fact, you can learn the basic sculpting approaches in mudbox and if you feel you hit the end of the road with mudbox's sculpting capability you can always move on to zbrush.
http://www.pixologic.com/zclassroom/homeroom/lesson/environment-with-tate-mosesian/
http://www.pixologic.com/zclassroom/homeroom/lesson/environment-with-tate-mosesian/
EDIT: @jograd10 .... didn't see that link -____-
it can make quick work out of larger forms. from there, for the more detailed stuff, surface noise and alphas are your best friend. it can also work alongside PS, so you can play to their strengths
In my opinion, 3D coat is only good for two things..maybe 3...and sculpting is not one of them. I believe 3D Coats strengths are in retopology and UV unwrapping. The third is their projection painting, which I actually this is quite bad..or rather so many other apps do it better.
I could see people getting 3d coat if they cant afford zbrush or any other number of apps out there.
It can be useful to learn (I really want to learn more of it) for your personal workflow though and if your studio lets you, include it in your professional workflow too.
Perhaps if you`re working at a company and you spend all your time making stone floors and walls over the course of a year but I have to make entire scenes in a few months with alot of variation in textures.
So this is my dilemma, I just dont think I have that time to build models, detail them , add texture and bake out maps to get one texture instead of getting decent photo resource and use photoshop, crazybump and maya depending on the texture and its done quick. And ok the final result may not be as pixel perfect as a zbrush texture but the average gamer wouldnt notice aslong as the overall effect looks good.
For now i think i`ll keep going at mudbox as its so quick and easy for doing painting and the modelling tools do the job. And like someone said, when I reach a point where I feel i need more tools i`ll look again at zbrush.
Thanks
You may ask next where to to use it. Well some people will say "for everything" and some "nahh photos all day!".
What I can say from my little personal experience:
1. It's very nice for creating rocks, big stone structures (columns, big blocks etc), destroyed elements you got my point I think.
2. I don't like making tightly packed grout less walls. Or tingly packed grout walls for that matter. I just don't see it worth spending time on these sometimes. But after this tutorial you guys linked I might give it another try.
3. I personally NEVER will do any hard sufrace in zbrush or any other sculpting software. I just like to have pin-point millimeter accuracy.
Diffrent studios/people have diffrent requirements. Neither photos nor zbrush will be going away any time soon. For more realistic things photos will still be definitly very good option, especially combined with ndo. You can also make photos look more stylized if you wish.
For very stylized things zbrush will be an very good option.
In either case (stylized and realistic), zbrush is very goo for things I mentioned in point 1.
Very nice tutorial guys. And for your concerns about it. Actually sculpting those bricks like this guy did, is pretty fast. They are not super heavy on small details. You probably spend more time at thinking how do them than actually doing them.
As I said you don't have do sclupt every wall. Still many studios rely on purerly photosourced texture. I faster, cheaper and well know. Age Of Conan, Crysis 1/2, The Witcher, The Secret World, TERA all those games use photosourced textures. Hell I have some photos from where they were derived in my library.
Even sclupted texture here heavly depend on photos for colors.
btw. Is there a way to download this tutorial ? Watching it over my uber-fast internet is really pain. With 20s loadings every 20s ;/.
For me it is, use daily basis, rarely texture in photoshop anymore.
Have you tried mudbox's texture painting yet? It blows 3D coat out of the water and then some. I made that mistake, and cant imagine using 3d Coat for anything other than retopo, but now thats easy in the newest zbrush update as well (or in blender).
Aside from the latest features...
https://www.youtube.com/watch?v=iLqt39YxSvE
The projection painting is probably the best I have seen so far, low poly, high poly, it doesnt matter. The brushes are smooth like photoshop and the interface is probably near flawless... at least in my opinion.
There's always been something wonky about 3D coats texture painting process.
I know that mudbox is becoming a common texturing tool now in game studios, along side zbrush for the sculpting.
you're not alone there
Zbrush managed to become that because of it's acceptance with various big studios and because its an integral part of many pipelines and workflows. Having Zbrush seats is pretty much required for any studio that makes more than casual / mobile games. Having 3DCoat (insert other "helper" package) is not.
Not saying that 3D coat is bad, far from it, but it's still a "utility" in the games pipeline, not a corner stone.
Yes would be really nice to have that feature in mudbox, Its useful to project and sculpt in the same area of the mesh but you have to be careful not to move the camera and you loose the position of the projection. Its something I like about Zbrush, that you can sculpt and paint simultaneously.
Actually I went through the tutorial myself again and it was very quick to make the bricks. I had convincing looking bricks and probably spent a few minutes on each just adjusting the shape. Still bugs me about the workflow, all the faffing around turning on polypaint, going into projection master, turning off other things, it gets you dizzy trying to remember which function was in which menu.
Maybe i`m missing something but I dread using Zbrush`s interface compared to Mudbox , Photoshop or Maya. Learning zbrush must be like working in a shoe shop, having to go through loads of boxes, until you find the right shoes, and then you put everything back in place, then you have to open all the boxes again to find some you saw earlier but cant remember which box it was in :poly122:
Strongest features are:
The concept of "material" which is texture set for diffuse, bump and spec. And ability to paint all channels simultaneously.
Camera projection and box projection for stencil and materials with option to retain texture size despite of camera zoom.
Comprehensive brushes and strokes set, including splines, stamp and drag for precise hard surface texturing.
Mudbox lacks these tools. And there are two very nasty problems that make mudbox useless for texture painting IMHO. Color posterisation when painting with low opacity and very unstable work with lowpoly meshes with uneven topolgy, i.e. not evenly tesselated quads.
Zbrush is indeed great sculpting tool. But it can't paint textures so it can't be used alone. You need 3dcoat or other painting tool anyway.
painting in 3d sculpting tools solve the UV texture seams issue too.
Zbrush will no doubt always look better because you're re-building reality in fine detail to bake from, but time investment to get that quality is likely much higher too...
The trap there is that artists need to really be good with their idea/reference material to get a realistic final result with that new capability on offer.
There may be Zbrush users who could use PS/nDo2 but over-kill on Zbrush, or those who use PS/ndo2 and could do better with Zbrush but are happy to sacrifice some quality for whatever reason (budget, time, both, artistic style, really necessary? etc)
I'd really suggest learning Zbrush. I am in the same camp as the OP but I think it's as clear as day where Zbrush is useful and required.
But I do agree with the OP that a lot of what we see is just technically amazing painstakingly made wrongness, unfortunately
High end games usually have enough eyes on the art to make sure it makes sense and tells a good story, but some images I've seen have materials that just make no sense for where they are used.
Ie, modern sci-fi weapons with chips and scratches on everything like it's been through a sand blaster hehe... or old temples with sharp cracks and sharp chips when it'd all be smoothly weathered away.
But that just comes down to not great art working rather than a fault of any given workflow or another... but it's almost automatically avoided with photo-source methods, wheras Zbrush methods allow you to make mistakes more easily I guess.
I'm looking forward to the day when we can just use a fancy camera that shoots a pair of offset images, does a depth map, diffuse, spec and shininess map (using some IR flashes or something hehe), and then we can photosource again, but with Zbrush style base assets to work off
Dave
http://www.cgtextures.com/texview.php?id=77377&PHPSESSID=qrsb9lujot3s6nr0640n6n6eh5
Basically i`m building half of the arch and duplicating/mirroring it along its length.
Now i`m currently pulling my hair out with maya`s flaky transfer maps. But i`ve built it with a base cube manipulated to form the main sections with then columns. Then using color, bump maps,etc to add the detail through a number of maya shaders to the high poly meshes and transferring them over to the low poly.
I`ve currently spent about 8 hours and am pretty much done, probably a 1/3rd of that time was lost due to maya crashing grrr. And using xnormal for the ambient occlusion. Finally photoshop to add wear and tear to the transferred maps.
So i`m interested to hear other methods of creating this? in zbrush, mudbox,etc
PhotoSculpt is heading in this direction.
http://www.polycount.com/forum/showthread.php?t=70506
But it's still a bit of a difficult process, tiling for example.
I think the question is more that if you don't need to use Zbrush then don't.
In isolation if you can get a better result with it, then do.
But I think many are forced down the path of using it simply because it's a good consistent tool for basing your texture/poly assets off in a large team. Suddenly it just makes sense to do everything in it, even basic stuff, because the workflow see's it the same.
Changing method for a floor/wall that looks basic then just means a different treatment for 1000's of floors/walls which might cost more time in the end.
I'm sure a game today could be made excellently without Zbrush, but I bet the teams would have a harder time working consistently with their layers and processes in Photoshop and whatever else.
Over thousands of assets and having every artist possibly working or re-working every other asset it suddenly makes sense to go the Z brush way I think, even for the bump map for a pencil or something.
Indeed from my perspective I'd see a strength in an artist that can achieve a great look with any tools they are given. That is what an artist should be capable of.
Even if they then use a tool that is overkill for some of their work, that might just be the way it goes.
But that doesn't make any other way wrong.
My 2p though, 8hrs seems a lot for an archway. Surely you could be knocking something like that out in Zbrush with dev tools for your specific engine in 1hr?
Despite my initial thoughts that Zbrush might be overkill for that I guess it's no worse than manually managing individual simpler but more numerous steps.
Hmmmm
Dave
I`m really getting fed up with maya/mudbox combo after this week, so many bugs and crashes in both. I basically had to end up turning off history in maya and saving before every transfer map. Then i was getting hangs in mudbox for undoing functions. Anyone else have these problems?
So i think i`m going to look again at zbrush which seems alot more flexible and stable.
How useful is this? How do you manage to get intricate details and into the higher resolution of texturing.
How does this work when you UV it, do you texture first then UV thereafter? What about painterly vs. "realistic" textures? I can understand "realistic" textures being brushed on but what about more stylistic textures with faux geometry and torchlight style for example, do you do these in your sculpter? Thanks.
Looks good but I can only dream about the poly count and texture VRAM that environment would use. I'm still having to heavily rely on repeating textures and low poly that console development requires and without any streaming tech so a whole level needs to fit into memory.
I think you'll find that an environment like that isn't that expensive in either polygon count or texture memory - at least by Ps3/xbox standards.
And the trick to making a whole level that has that kind of texture quality and poly count is to re-use the same assets and style throughout most of it.
I moved onto one particular game object and it was driving me nuts in maya and mudbox with the normal map distorting. This morning I went into zbrush and within 10mins starting from scratch had a perfect normal map and even 2-3x faster calculating than the other packages. I then imported the low poly object into mudbox and painted the textures , probably 30mins work total!
So now I think i`ll be spending more time with zbrush and learning the tools better.
Thanks
Registered just to resurrect this.
The thing about ZBrush and Pixologic is they have a clear plan. There are many big brands out there that make certain kinds of productions a lot easier because it is their niche and ZBrush is constantly moving towards their own niche as well.
I am no professional, no master, not by a long shot, but I have an affinity towards "freehand" and a love for hard surfaces mixed with organic. Zbrush, with each update, is adding "freehand hard surface" to their initial niche of "freehand organic" with some controlled organic and hard surface on side.
Tried mudbox, 3d Coat (to an extent) and I believe, if you don't lose sight of the fact that it is the artist's responsibility to be accurate and prepare with their references well, Zbrush is by far the easiest to put together complex tools and thus maps.
Like I said, I am not a professional in any kind of arts, digital or traditional, but probably just after my collection of microphones and voice processing pedals, Zbrush has been my best investment both time-wise and moneys-wise.
Cheers...
What is your goal there? Do you have the time to learn the quirky and unique Zbrush interface and "tradition?" How involved you would be with freehand organic (and now hard surface) sculpting as opposed to controlled (box?) modeling? Do you need to have low polycount on your models? Or can you technically and job-wise afford millions of polygons? Where do you stand on texturing by hand?
Cheers...
I already know how to use zbrush but I used trial version so i cant use it anymore and I like zbrush thousands times more than mudbox because sculpting in zbrush seems to be more responsive and "detailed",right now im using mudbox only because i can get it for free with the student license.