high res sculpting - why bother?
polycount lvl 666
i am thinking now that since a lot of studios seem to be using scans( even full body scans now) for realistic characters, why bother learning to zbrush/texture heads and bodies.
Ok, there is stylised stuff you could be doing, but in terms of realistic work, the only fun you are going to have is in maybe setting up the materials/lighting and retopo( what fun)
I kind of get how animators must be feeling since the introduction of mocap
or am I being too negative here
Ok, there is stylised stuff you could be doing, but in terms of realistic work, the only fun you are going to have is in maybe setting up the materials/lighting and retopo( what fun)
I kind of get how animators must be feeling since the introduction of mocap
or am I being too negative here
Replies
2.) Because you love it...
3.) Because scans won't replace 3D artists...just as mocap hasn't replaced the need for quality animators...
Also, you have to have good understanding of human anatomy if you want to do high quality stylized characters or even monsters. Any orc or elf or whatever is based on human anatomy.
Which is more cost efficient? finding the exact jacket you want on your character and scanning it... or modeling it from scratch? I'm willing to bet the control needed in most situations will sway in favor of an artist creating it from scratch.
I guess you could make original heads by morphing the base meshs together, so combining the normal map info from separate scans. still not my idea of fun though.
You are reduced to merely 'processing' rather than creating anything.
make art is button is nearly upon us
On the plus side, I dont think it will ever be so technical that people without art skills will be able to do it, but in day to day work, i think workflow will use more and more scanned/retopoed parts, with artists just creating the stylised parts needed to complete the design.
I think now more than ever, 3d artists are actually going to have to become better all round artists to not get eaten up by the advance of technology, or at least trapped in a retopology role.....
what more efficient, spending days sculpting the folds on cloth or scanning the info instead.
I have already worked on games where they have used scanned cloth/heads and they looked really nice too. that does worry me because actually making the art is the thing that I enjoy the most.
Smash the Spinning Jenny! Burn the Rolling Rosalind! Destroy the Going-up-and-down-a-bit-and-then-moving-along Gertrude.
5 extra points for guessing the quote
First and foremost I enjoy trying to break my personal barriers and improve, I won't improve or learn anything from using scans, not to mention not getting any creative input out of the process.
As already mentioned, what will the scan lovers do when there is no scans for character X? As game worlds become bigger and more complex, so the demand for the volume of characters rise, then it would be pretty impractical, time and cost, to scan every single one.
But it does seem that it's a tide that will come in no matter what.
I would suggest so see scans more like a perfect way of getting good 3d reference images for volumes and shapes.
Also it's great to create library material for close and mid distance surface details or location references for environment recreations.
Texturing in a photo realistic manner is also more like the job of an craftsman then of an artist.
But there are so many ways of doing stuff: art style, platform dependent preconditions or just imagine you want to create an alien or anything else where you for sure don't get a reference you can scan.
two things.
first, the excuse of "you still need artist to..." is not an excuse worth mentioning.
second, not sure if your analogy makes sense because camera came before photoshop. typically you would want to compare a newer technology that replaces or devalues an older technique. for example, mocap REDUCING the "value" for talented hand animators. there is a difference in reducing the need of someone vs reducing the value of someone.
anyways, let me give you couple of typical replies from someone who might claim to "embrace technology"
1. because you still need to know anatomy to clean up scan data and we all know how fun/creative/fulfilling process that can be...
2. because we can hide the fact that we used scan data and take credit of the entire work because it makes us look amazing...
on a serious note though, this is the nature of technological evolution and you just have to deal with it like everyone else. just hope and pray that you can retire before the day comes when ALL of your process is automated.
There's still a fair amount of knowledge needed to get a good end result though. We utilised a fully scanned workflow for the last game I worked on, and the difference between the work a good artist compared to a medicore artist was pretty staggering. The good artists could bring a solid grasp of anatomy to implement the changes we requested as a beautification pass in the sculpt, mastery of topology and baking to get a good, crisp normal map bake and strong texturing and material definition skills to get realistic materials and textures.
Saying that I don't feel the end results were that much better as the in-game lighting and materials still left the character looking CG anyway (game was on 360). I also felt that scanning really sucked the joy out of the artistic process and am now working on much more stylised projects and much happier for it!
I suppose I should be happy that at some point I would n't have to sculpt a pair of jeans from scratch, but it's that very challenge which is the fun/painful part
As rollin wrote though, the scans are good for reference too and I do look at scanned data sometimes to see where I am going wrong
I class my work as semi realistic anyway, but I think i am going to head back the other way and take it more on the cartoony side so I don't get totally frozen out of the market
MM, by the time I retire there will be robots to do all my work for me anyway.
that's why i bother doing high res sculpts.
Than look at stylized games, your not going to see darksiders or wolf among us, portal or tf2 using scans are we.
Scans aren't redefining how we do art, there just a other tool in the toolbelt. Did the invention of the camera prevent people from painting from life? Did the invention of cnc machines and 3d printing stop people from sculpting in clay?
BTW your tin foil hat isn't in style anymore mate, maybe get your self a nice flat cap, or perhaps a homburg.
My point is that in the future that particular skill {ie sculpting) might not be required as much, even if it is fun.
personally I enjoy the texturing and materials side too,so I would still have the chance to be a artistic after a fashion
passerby, I think your wrong on that, lots of jobs on offer these days require photorealistic skills.That really does get pushed a lot these days.
Of course there are stylised games duhh and I did mention that further up the page.I have toyed with the idea of going back to more stylistic work, but I like more gritty stuff these days.
I think 3d scanning 'is' redefining how do art and for AAA games that's only going to get more prevelant.
I think 3d scanners will get smaller and more widespread, so it would be really easy to scan a whole body, rather than spend days sculpting it. so many characters have cloth bits these days, or imagine your average prop modeller,how much of their output could be replaced by using a hand held scanner.
also with the imrovement in auto retopo these days, how long before that is no longer a task in the pipeline.
maybe the most creative job to have in the future will be shader or lighting artist?
sure, if your goal is to replicate reality perfectly, whatever tool does that the best will become standard much like how sculpting replaced subd modeling for characters.
On Call of Modern Battlefield Gears 12, maybe.
On something like Okami or Bioshock Infinite where the art direction isn't "scannable"? Somebody is still going to have to cut geometry and paint those suckers.
There are also many, many other types of objects that go into a game that are not characters, only a very small % of a game is the character assets, many of which are too large (environments, buildings) to scan but still may require some sculpting.
I would like to elaborate on that by saying that you need to know zbrush, you need to know how to sculpt, you need and understanding of anatomy and how clothing sits, now more than ever. Because 3d scanning exists ONLY in the realm of zbrush, in tight anatomy and in complex materials. If a studio wants to use scan data to drive a high fidelity, realistic aesthetic to their new title, you'd better know how to work with incredibly detailed character models. It wouldn't hurt to understand the physics behind your 3d capture system, how it captures or does not capture different materials. How to capture cross polarised textures, how to use all the data you're getting effectively.
That said it is true that in the field of "realistic" games, 3d scans will significantly reduce the workload of character artists in particular. But I fail to see this as a bad thing! Certainly in the long term! It means smaller studios being able to pull of bigger more ambitious projects with less people. Do you really want monolithic 250+ person studios being the only way AAA games get made? Side note: I can't see how anyone would complain about increasingly useful automated retopo tools. Is there some well paid retopo artist out there thinking, "man I sure love my retopo job, It's quite scary that soon the art of retopo will be lost to mankind." Noone loves retopo that much... what was I talking about? I think I've got side tracked...
Anyway, back to cleaning up this scan data I've been working on since 8am.
Would anyone actually be interested in this, and what questions would you have? The caveat is I basically only do work for VFX and Film...
https://philippefaraut.com/
http://www.richardmacdonald.com/
http://www.schellstudio.com/
http://imaginerick.com/
Now of course finding a "Zbrush guy" would be easier and probably more affordable ... But I would imagine that a superbly skilled traditional sculptor could very well find a place in a game studio. Now of course that would be more of a pre-production kind of job but hey in some cases that's even more fun ...
i would guess the main issue with traditional sculpting-to-digital process is that it is too slow and too messy. i could sculpt a head with lot of details (realistic or stylized) in 30-60 minutes but the most experienced traditional sculptor probably couldn't sculpt something to that level of detail in few hours. i could be wrong but i havent seen traditional sculpting done as fast as digital. also no ctrl+z in traditional sculpting and easier iteration/revision possibilities in digital medium.
some scanning tech relies on multiple photographic viewpoints which makes it even more accessible.
I am wary of saying that you can't scan things that don't exist , because in a roundabout way you already can (simple 3d objects from 2d drawings)
maybe in a few years time , you can turn a detailed concept with multiple views in to a 3d model - automatically with nice automatic retopo + colour info
Then the whole thing will have turned full circle and 2d guy will be king again:)
Even a tidy mesh is only a base for sculpting. I think experienced character artist underestimate the importance of their knowledge and skills when working with scan data. I also believe that as scanning becomes commonplace, people will start to see it as a starting point, not a quick fix. Look at the sculpts naughty dog are making. This is not possi le to generate with 3d scanners. Even if you had managed to scan such a large, organic, moving object at sub millimetre accuracy, the data would be insanely hard to use. Raw scan data is not human friendly and most high end computers have trouble with that kind of unorganized mess of verticies. However scanning an actor or mannequin, the costume, and photographing it would give you useable data to create a digital double in less time than a straight up sculpt but would require all the same technical skills.
And yess you can convert 2d drawings into 3d models... It's called sculpting. Don't let the siggraph "primitives from a photograph" tech fool you, what you are thinking of is decades away and for concept art, probably impossible.
This video re the fifa stuff is pretty nifty. they say it took them half a day to get the model looking 'perfect', then half a day for rigging. seems way quicker then sculpting to me
they seem to go straight to the finished retopo mesh, not sure if that is generated automatically or they retopo cloud points or whatever, but the results are pretty good.
I am guessing they have a base mesh which has some kind of tracking points then distorts to the multiple viewpoints.
http://www.youtube.com/watch?feature=player_detailpage&v=ZBIjwU6EbOA
re the 2d to 3d stuff I know most of the existing stuff is very basic, but as I mentioned in my post, a few years away - but I doubt it wil be decades away, maybe 5.
photomodelling has advanced a lot in the last few years, so I don't think it's at all clever to be so dismissive
BUT that is completely besides the point of this discussion i believe.
i really really dont want to clean up scan data and would never want to get a job where all i do is create models based on scan data.
i too am moving away from realism because of that.
btw, Naughty Dog didnt use any scan data AFAIK for lust of us characters. so scan data were not a starting point for that. good for them for making that choice.
however, more and more studios are preferring to just scan a real person instead of making one digitally. i could list lot of popular names but i might be breaking NDAs.
scans should only be used for digital double. it should not cut into the creative process and as a result make studios lazy and resort to 3d scanning.
how many people you know who do this for a living, ie. hand sculpt people for a living.
comments like that using the word "still" makes no logical point because as long as there is ONE single human being in the entire world who is still hand sculpting then it is STILL being done.
Well, people that do it for a living are in a different industry. I don't have any Hollywood connections, I'd guess the lack of available work for sculpters may be balanced by the lack of sculpters. I know it's hard to find people that can do low poly, diffuse only work without the aid of a sculpting program.
On the other hand, I don't personally know anybody that works with scan data.
Scans look like animated corpses to me anyway, there has to be a niche for good artists, right?
there really cant be an argument over the scarcity of traditional sculptors, but they are scarce partially because of various reasons like scarcity of proper training, lack of demand in that work and all of this are because of another medium filled up the void.
and if we are gonna talk about a different industry other than game industry then i dont even know what to say because at that point it is "neither here nor there" kind of a logic.
Also, i dont think iv seen any commercially available scans that are directly useable as a normalmap source. usually you will have a lot of surface noise that will need cleaning up and you will need to add surface detail like pores and fabric patterns etc manually.
Ruz so what's stopping you from taking stylized work? It's gotta be better than doing scan cleanup.
Not a chance. Even if such a thing would be somewhat possible, it would still look automated and "manufactured".
The whole "completely automating art" sort of schtick never catches, simply because you need human touch and individualism during the entire pipeline.
You need to optimize, stylize, and keep everything dynamic.
That's why handpainted textures tend to look better than photobashed textures, or why kitbashing tends to look less appealing than unique modelling.
meshpotato - my somewhat tngue in cheek point there was that the concept guy would be king and his style would pretty individual and central to the artistic look , then his 2d orthos( which would be pretty detailed) could be turned in to 3d with an automated process.
might happen one day then I will come back and laugh at you, but i am pretty crap at concepting so i would be boned
Eh? o_O
I think they have a standard head mesh they project onto the cleaned up scan, and then project textures from camera positions in maya or nuke or something else.
Ok I think I'm starting to see your point and maybe agree with you - under certain circumstances. Under 360/PS3 gen art budgets, it would be really easy to take a nice scan, work on if for a few hours, fix up the textures, topo, rig and chuck it in game. But When you look at the high end character art of the new current gen titles, you are looking at a level of quality and realism that is hard to break through to and very hard to get right.
This is where I see 3d scanning, taking a very hard to reach level of realism and bringing it much closer, making it more available. It won't go all the way. But it make this level of detail more affordable.
Aye, I think you have to see this as a way of cutting down the time you spend on getting an individual character to a very high level of detail, and allowing you to spend that time on more art across the project.
This is pretty much how I feel about it. As an environment artist who works on head and body scans, I feel really dumb trying to sculpt anything on a head scan. And I see people who don't have anatomy skills wrecking a scan they thought they could fix by hand. (the trick by the way, is you must get the cleanest data you can get at the capture stage. Rule #1!)
I already see it right now, and scanning at this level is only starting to ramp up.
And we will all be out of our jerbs!!!, as china or india (or thailand, lots of stuff is moving to thailand, cambodia etc. because they are cheaper than china these days o_O) has enough space and people willing to do this to ramp up big scanning services for cheap.
And the quality isn't even inferior or bad, it's good and cheap.
But cheap isn't everything, communication, beeing close to your clients, working directly with them, is something you can't do (at the moment) with thousands of miles between the partners.
So whats to do now? I think concentrating on what can't be scanned - looking at your portfolio i don't get why you fear it, your stuff is stylized, not a single realistic piece in there, but maybe it is dated. But the stuff you do, no scanner will do for quite a while. The question is rather do you want to do the realistic work?
Because then you should better get used to working with scans and or simulation for cloth for example - no way around it, realism should be based on realism doing something from scrath is a waste of time if you look at it from the production point of view.
Especially with the new consoles around, the new productions to do. The level of detail they create, all this costs money, workflows will have to remove jobs that have been around for quite a while, because the amount of assets and details in those assets can't be produced for the same cost as last gen productions without making things faster. But mocap didn't remove classical animators, 3d didn't kill all 2d artists (many lost their jobs in the film or gaming industry for sure), but thats the thing with progress, some will not adapt - but you can't change that.
Heck not long ago, we created LODs by hand, now lots of productions use Simplygon, i don't cry a single tear after losing the job of creating LODs, i'm happy this monkey work is gone, and there is still a lot of monkey work that can and should be automated, it's a waste of money and talent to do certain tasks by hand.
How many people actually try, 100%, to make a living from doing such things? Not impossible but the stigma is that you will be a starving artist. I went to an art school for a time and constantly heard from teachers about getting into painting and sculpting and being a starving artist. You know, Pixar sculpts their characters and such before modeling.
Only when you can start scanning original ideas from paper to computer, then will scanning take over. For there is tons and tons of stuff that cannot be scanned because it does not exist.
Also how practical is it to scan animals that won't sit still? What about large objects?
both depend on your solution, a one shot camera solution, not a laser/projector based system should be able to scan an animal, unless it is moving very fast.
big objects can also work, drones come to mind, isn't there a kickstarter project claiming to do exactly this? I doubt it is doable in a decent quality now tho.
But that stuff can't be done NOW, doesn't mean it will never work, this stuff is beeing worked on and improved constantly. Who would have thought a couple of years back, that you could do decentish scans with a console camera tool (kinect), or a smartphone? Nobody, still it is doable now and will get better as time goes by.
You know, recently we bought ZBrush, dDo and nDo at my job. Last I checked It didn't start doing all the work for me. You know what it did do? It made my workflow much faster and raised the bar for what I was able to make within deadlines. Depending on those deadlines I can either spend the same amount of time on something but raise the quality significantly or I can just jam it out in no time for faster turnaround.
If you work in a studio, you are there to make someone money. If the options are to scan three people into a system in the time it takes you to sculpt one, what do you think is the more efficient option? Learn new toolsets/workflows and stay relevant.
^_^
Makes jobs soooo much easier.
Would have definitely helped us out, if we had used them on Thief.
Regarding scanning : I think it's great that the teams behind games like Fifa have access to cheap scanning techniques, as hand-sculpting all these athlete faces would be the most tedious and boring job ever. The scanning results are very creepy and lifeless, especially when animated, but I think that's fine for that purpose.
And then there are projects like The Last of us, which relies on very complex and subtle animations systems ... and interestingly, I think that if straight up scans/doubles had been used, the result would probably have been much less convincing. (as shown by LA noire and Two Souls, whose characters sometimes manage to look right, but more often than not look very stiff and weird).
So at the end of the day ... scanning tech might actually be limited by the much more impalpable notion of suspension of disbelief. And I think that an awesomely skilled (traditional or digital) organic sculptor, able to give his/her creation that extra little spark of life, will always be able to find a job at studios where the art leads/AD are aware of such things.