I go one step further, and type some code that produces a teapot automatically, and the teapot in turn generates and runs its own code. Then I go and make a coffee, and when I come back, the teapot has produced whatever I required.
What i want to say is that a good3d character (not a character artist ¬¬), a good 3d model should be done using polygons, subdivs. I think it's simple.. i really don't know how to explain it better , subdiv modelling is subdiv modelling, it's not related with something "more artistic" which is very subjetive.
Zbrush models are mainly static works and in most cases:
-you part from a bad base mesh to speed up work, useless for rigging, animation, etc. (a fast and dirty one)
-you don't care about topology, and at the end you could need to use retopology tools, something terrible for our workflow (face loops, muscle loops, etc.)
-you don't have a good base mesh with good uv mapping and deformation. Does anyone test their meshes for a good deformation? doing inorganic shapes only... i suppose the response is no.
-you can have a "mucus" model with nonsense wrinkels, folds and silly details not needed.
-etc.
Subdivs are not only for mechanic objects, that's something i want to say, and you can find in organic modelling as technical issues as with inorganic shapes, maybe more.
Talking about modelling, there is only artistic side if you really know how to model.
WTF, it looks like Zbrush is the unique tool for characters nowadays and the work with characters is so easy... and that's not true. As much, in a real production, you can use Zbrush only to give a fine detail pass to generate a bump or a displacement for a still render.
With inorganic shapes, one done, all done, is always the same mechanic thing, you learn the technique and that's all, you don't need to know anymore. And if you have a library of mechanical or inorganic subdiv meshes, all is copy and paste. With normal mapping floats is more of the same, making subdiv models like a weapon is easier using floats, and without floats too, is all technique and experience IMHO.
You should understand what i say if you only use Zbrush or Mudbox to add the fine detail you can't add with polygons. What I'm talking about is about Pure Subdiv modelling... hi-res modelling of characters for animation (a character also has inorganic shapes attached). It's a harder work but easy aswell.
3D is not only Game Art, and models for normal maps. Zbrush can help a lot but not much when you need to build a hi-res model for animation. With too many models you don't need Zbrush, and all is Subdiv work.
Another thing is that the people who paint realistic forms on a paper or canvas have better predisposition to do greater 3D. is fully proved!. We are not special for making 3d, anyone can do it, and it's bacause the majority of the common 3d tasks are mechanic, can be done by TOOL guys, 3d monkeys, scripts, etc... it's the reality.
blaizer not trying to be smatass or anything, but i cant make sense of your posts, i think its because you try to say the same thing but you change the line of thought completely on each time, could you just elaborate what you mean but in a shorter text/that makes sense ? ( no not beeing ass, interested in what you have to opinate actually )
Blaizer:so what your saying is that if your not making game art, but high poly characters for movies etc you should use sub-d in stead of zbrush>lowpoly>normal map... but this is a game art forum haha. it's quite obvious that deformation will suck if you create a hp with a crappy base mesh in zbrush and then try to use it for a scinematic, but usually nobody does that.. we just use the hp for the normal map for the lowpoly so who cares that the actual hp will deform poorly?
(i'm saying we, but actually i never do any high poly work hehe :P).
also... YOU SUCK! . (just to heat the thread up some more).
Blaizer, I somehow understand where you are coming from with this. BUT you forget one important point : retopology aka resurfacing.
Things are not linear anymore. You can use Zbrush, Mudbox, 3DCoat, heck even the free, loveable Organica to create striking silhouettes for characters (hence, NOT just surface details.). Even harsh primitives mashups work very well for this first step (I do that all the time now, then take that in Mud to wedge stuff together).
Then you can go many ways. Either detail the crap out of that step, or roughly retopo and reproject to go on sculpting. Then, you make your proper support mesh, whatever that is (subdividable cage for film displacement, or lowpoly 'outshape' for game art).
The beauty of todays tools is that nothing is locked anymore. Such a cool thing to embrace!
Yep, I'm with Pior on that. Modeling is easier today than it ever has been as far as the technical aspects go, which can have a nice bonus effect on exercising one's creativity. I have a wealth of tools and techniques at my disposal, that sometimes I just use on a pure whim to change it up. It is an awesome time to be a 3d artist today and I think it will only get better, as tools evolve to be less about technical bs and more about just doing your job as naturally as possible.
Using sculpting as a kind of example on how things are changing... some companies actively hire modelers purely for sculpting. That's all they do. Lower rez animation cages are then built by a more tech-savvy artist, UV'd and skinned to a rig, since there is less 'artistic' ability required to reproduce the low rez model from a high rez. I'm not saying this is a right or wrong method, but simply that it IS a method that people are now able to use, where before it generally wasn't.
With inorganic shapes, one done, all done, is always the same mechanic thing, you learn the technique and that's all, you don't need to know anymore. And if you have a library of mechanical or inorganic subdiv meshes, all is copy and paste. With normal mapping floats is more of the same, making subdiv models like a weapon is easier using floats, and without floats too, is all technique and experience IMHO.
I don't understand half of your posts but this quote is simply misinformed if not ignorant. To say inorganic modeling is all copy and paste is a load of crap.
I must think all comes due to my different way to do things. let's put a round of images so you could understand me better. 1 image = 1000 words hehehe
This is what i mean with a bad base mesh. Retopo tools are something bad for me, only lazy ppl uses it imho, you need to redo the base mesh when you could have starting doing something well, like a MAN does LOL. It's time you loose.
And this is how i make subdiv models (with a concept done previously).
The work is done quite fast... but it has its work, its time. It's not so "easy" to do as the crap base mesh i posted, which was done in 2 min and maped in 1 min (bad uv, bad seams, bad all). If you are going to do a 3d doodle, a model for normal maps... i undestand it (we work in the fast way), but for real production for animation.. start sculpting a model from a bad base mesh is time consuming! and time=money.
For me, is the same to model hi-poly or Low-poly, 3dwork is 3dwork, you just put in practice all what you know. When i'm modelling a hi-res model, i start from a low low low poly model as you may see, nothing about poly to poly or edge extruding.
Here's another example..
To say inorganic modeling is all copy and paste is a load of crap.
I said: "And if you have a library of mechanical or inorganic subdiv meshes, all is copy and paste". I didn't say inorganic modelling is all copy and paste dude ¬¬ read well. i'm using copy paste all the time, i make things very fast. Do you have a bad workflow? hehehehe
bad forum to talk about subdiv modelling, not so many hi-res modellers here hehehe :poly121:
PD: actual game models, are quite similar to subdiv models.
perna, i talk about doing bad base meshes for sculpting, first image i posted vs second. With my base mesh i don't need to use retopology tools like others, is faster and more efficient imho.
Of course a model intended for animation needs to have a proper structure. But trying to hit that structure right from the start is not necessarily the only way to go. As a matter of fact, it *used* to be - before sculpting package appeared.
The point of Mud and Z is not just crazy zspheres or basemesh monsters, nor just high detail sculpting. It is also a fantastic way to explore shapes and proportions. That freedom cannot be reached with the traditional use of soft selection tools.
Also you seem to assume that a model is always made from a precise, locked concept. This is not always the case. This makes the free form approach of sculpting packages even more valuable.
Moreso : it is actually great to have a dirty sculpt as a base, even if the final product is to be a carton-like animated movie character. You can sculpt and adjust the shapes nicely and freely in mud, then create the clean subd model afterwards (using, guess what? retopology tools. Try Topogun). It works extremely well and saves you from hours spent trying to find a flow that doesn't pinch, stuff like that.
I wont even comment on your opinion on mechanical modelling. Seriously, look up some examples and think about it some more.
White boxy dragons get boring after a while But I like the way they look, I admit it!
I go one step further, and type some code that produces a teapot automatically, and the teapot in turn generates and runs its own code. Then I go and make a coffee, and when I come back, the teapot has produced whatever I required.
What could possibly be wrong with a mash-up computer language written by 100 different programming interns? MAXScript (btw that is proper spelling/caps) is so awesome I cannot stop cursing at it sometimes.
Of course a model intended for animation needs to have a proper structure. But trying to hit that structure right from the start is not necessarily the only way to go
Pior your are right, but the method you talk about is more time consuming at the end (subdiv model for animation). It's like to follow the longest path. If you know what you have in hands, why don't do things well from the start? lazyness?
I've worked several times doing retopos from sculpts, and takes more time, it's a nuisance. It was another experience working with some new tools, and that's all. For me it's not the best way to do things.
I usually explore shapes and forms in my mind, and i use the paper to "record" the idea, it may be very very old school but it's fast enough and the best for me. I make a draw/concept in minutes... why should i need to do a 3d doodle having an idea so solid? is to loooose time, seriously.
Don't think i assume that a model is always made from a precise, locked concept. Sometimes i model from an idea in my mind.. and i simplily build the model in 3d without any concept. For doodling, i prefer pencil and rubber, i'm a traditional guy. And again, due to lack of time, i avoid to do concepts too many times, but i always know what i'm doing.
It seems like almost all ppl here takes the long way, and i think is because you don't have a solid idea in mind. Doodling is fun, but believe me, you really lose time, too much. And Retopo is bad for a good workflow imho.
I'm not so kick-as perna, i run into bad proportions issues all the days because i still don't see some forms/volumes. As artist, i'm very green, as much as my fucking plant lol (I feel ashamed of what I did the day before). My eye is not so good, but i can say that technically (modelling), i haven't got so many problems. I only know how to use the tools i use and i'm constantly growing as artist.
Don't know how explain it better, i thought it was something very simple but it seems i'm the only one who can understand it hehehe. I will end saying this again: to make an human, is far difficult than making mechanical or inorganic model, that's all (at least for me). You got the uncanny valley theory and that's enough proof of how challenging (difficult) is to create a credible human.
Ok getting things back on track (since I was helping to derail it).
I'll post my work flow for environment art. Hopefully I'll fuck something up and contribute.
Design hands us a written description sometimes with crude drawings or photos.
The env team is broken up into sub-teams of 2 people (technically 4 if you count production and dev). Each group takes an environment per milestone (2 milestones per game, two games per year)
Person A, 3D. They make most of the 3D content and do some light texture work. They also wrangle the cameras, lighting and do simple mechanical animations. We have 3D characters on pre rendered background think old school RE, but from the FPS pers.
Person B, is in charge of most things 2D, but does some light modeling and unwrapping. Also draws up concepts, designs puzzles, touches up background renders, and delivers all assets to production to be put in the game.
Its really up to each team to get the job done how they see fit, but a normal env goes something like this:
Concepts, more often its a quick 3D mock up, passed it off to the 2D artist who fill in details while the 3D artist starts to flesh out the scene nav, lighting and block in some details.
Once its pretty much blocked out and a bunch of details have been dreamed up, we fill the scene out, sometimes other artists or interns help out each adding their own stink to the scene.
Depending on the 2D artist I'll create unwrap templates that are artist friendly. Before I get the textures back I'll rearange the UVs to a 2nd channel and then use RTT to transfer 1 to 2.
Game gets built...
Stuff happens...
Release party...
Down time to work on tools, archive libraries...
Launch Party...
Wash rinse repeat...
Ok it took me a while to find this post under all the faggotry, but i'll give it a shot.
Since i dont really have much experience working on the style of games that you do(lots of prerendered stuff, right?) I cant really offer much from a critique standpoint.
Instead, lets try flipping this around. Could you make a little writeup about what the pitfalls and problems are with your workflow?
first of all: i hate you all, you're the most pansy-ass bunch of elitist pricks i've ever laid eyes upon.
now that i've offended you all, its time for workflow.
*reference gathering.
*more reference gathering.
*check polycount, check webcomics, check various military modeling sites, check youtube.
*watch 2 hour documentary on chimps.
*check polycount again
*even more reference gathering.
*go to sleep, wake up next morning, check polycount.
*start modeling, look at reference
*look at reference some more, check polycount, check youtube.
*go to sleep again.
*start some actual modeling. get good progress.
*look at reference, check polycount, check youtube, curse, saying to myself i shouldn't be wasting time on the net.
*get depressed over lack of progress.
*quit modeling, saying to myself i'll finish it at a later time.
first of all: i hate you all, you're the most pansy-ass bunch of elitist pricks i've ever laid eyes upon.
now that i've offended you all, its time for workflow.
*reference gathering.
*more reference gathering.
*check polycount, check webcomics, check various military modeling sites, check youtube.
*watch 2 hour documentary on chimps.
*check polycount again
*even more reference gathering.
*go to sleep, wake up next morning, check polycount.
*start modeling, look at reference
*look at reference some more, check polycount, check youtube.
*go to sleep again.
*start some actual modeling. get good progress.
*look at reference, check polycount, check youtube, curse, saying to myself i shouldn't be wasting time on the net.
*get depressed over lack of progress.
*quit modeling, saying to myself i'll finish it at a later time.
What is all this check polycount business, sounds like a terrible waste of time.
Alright, since vig went through the trouble of either finding or making this silly flow cart, i'm going to post a few of my thoughts on prop workflows.
Now, first let me acknowledge that every project is different, and engine as well, so some principals may not always apply. But i would like to think so of these ideas are fairly universal.
Modeling in sets.
Whenever possible, i find that the most efficient way to make any prop is to make them in a set, now sometimes you need to do a one-off prop, or a few singular assets that are very generic and could be placed everywhere, but i think those cases are less common than people would like to believe, or should be at least.
Lets look at this from a practicle time standpoint:
Lets say on monday I give an artist a prop to make, he goes through, makes the high, low, uvs, bakes, textures, gets that all done monday.
Great! now lets say i give him a similar job tuesday, wednesday, etc. By the end of the week i should have 5 props, right?
Ok, now lets say i give an artist a set of 5 props to make, he does all the highpoly models at one, low, uv, bake, texture, all of these in phases. Chances are he's done by wednesday, or thursday at the latest. Thus saving 1-2 days on every small set of props that need to be developed.
Not only that, but his results are going to be more consistent, and cohesive, he wont spend a day making a bolt, and then a day spending the same amount of time on some ornate light fixture, instead its easier to prioritize the amount of work that goes into each prop. Also, if you can manage to combine props onto a single uv set, you will end up saving UV space in the end, chances are if you've got 5 separate props, all with 512s, you could have easily fit all of those assets on a single 1024, saving you a good chunk of memory.
Now, onto the issue of grouping props onto one uv set. First off, i'm going to start by saying that if you're working efficiently, spending a
few more minutes to better plan what will be needed, and work in groups, any cleanup time will likely be negetated, and you'll still be
making assets faster than if you were doing them one by one.
Ok, so when should you group assets?
In my opinion you should do this whenever possible, the speed gains you will get from being able to do all of your bakes at once, and texture
them all at once should not be understated. This shit saves a lot of REAL WORLD production time.
Now, i understand that it will not always make sense to combine a group of props onto one sheet, but i think if you're smart about it, it
becomes a pretty simple decision.
Here are some examples of situations where i would combine multiple props onto one sheet.
Say you have a set of very similar props, for example lets say you have a computer, a mouse, a monitor, a keyboard and a printer. Really,
this is one cohesive set of props, that SHOULD be placed in the same vicinity, its one prop set that is split up into multiple pieces
mostly just to give better variation in placement.
When you have one type of object, but with multiple variations, lets say a few types of crates, or a fence that has a few different peices. This is going to be a good time to combine objects, because it is safe to assume they will be used in a similar area.
There will be plenty of instances where you will not want to group objects, throwing something small that will be reused a lot onto a large, unique prop that will be used very little is a good example of where you REALLY DO NOT want to combine objects.
Now, onto the issue of set dressing.
So, lets say "Dave" is an idiot, and he takes your nice computer set, and only place mice around his level, 47 mice infact.
Does this mean that your grouping system is totally wrong? No, it simply means that "Dave" is a fucking moron.
Some things we can do to stop these situations before they start:
Spend some time, work with the LD/Set dressers. Make sure they understand that you want to have "themes" to your areas in the game, and that these themes should likely be set dressed with appropriate objects, that actually make sense to be in the area in the first place. The more cohessive and common sense your prop placement is, the better your game will feel.
Spending the extra time to be organized, and have a more unified goal/vision on how your scenes will be set up will be a major boost in the long run.
Now, there will inevetibly be cases where not all props in a set get used. Is the only solultion then to split each of these up, redo the uvs and RTT the textures onto a new sheet? Absolutely not, this is the worst case scenerio, and something that i would only do if say, only 1 or 2 small pieces of a 5 set prop sheet were be used.
If only one piece out of a set is being used, chances are this object is OUT OF PLACE, and can be easily be replaced by a prop that fits in this environment better, this should be an easy fix in most engines, much easier than reworking entire sets of assets.
Did "Dave" place a wire spool, mouse, crate, and bag of doritos all in one spot, in one level, where none of these props or similar props in the set are being used? The easily solution here is to fix "Dave"s work, not rework all of the assets.
So, work more efficiently, take some of that time you save by doing so and spend more of it planing what you will actually need, and working with the right people to make sure there isn't massive misuse of assets, and the world will be a much better place.
I completely agree with the exception that when you hit the unknowns.
I think its better to put the unknowns on their separate sheet instead of tagging them into the nearest wasted space and hope that no one places them as a separate object. You can easily tag unknowns together later after the dust settles and you know for sure if props where placed "correctly".
I think there needs to be another step or procedure in place other then "I told the LD to place all these light posts together but not those" A naming convention, a flag in the editor a way to group or sort the props something other then just "Remember 6mo ago I told you the rat and garbage piles go with the wardrobe boxes!" It's not only Dave the moron but maybe a few other people that weren't told. Having a process makes that process doable and favorable I think in most cases, provided its there, and not on the "remember-system"
this conventional way of unwrapping things into the 0-1 uv space and painting a dedicated texture that fits just that object is completely obsolete if you ask me !
you should make textures that are SOOOOO useful that they can go on practically anything and everything else can be dealt with geometry/vertexcolors/lighting.
so don't put all sorts of object sheets next to each other ... instead use ONE AND THE SAME sheet on everything
Yeah absolutely, being better organized and having a documented process would help a ton in dealing with those sort of issues. Throughout the production of darkest of days i oftened wanted to make.... Like a spreadsheet i guess, of various props, good ways to use them, where they were located, etc. We had pretty terrible naming conventions so often times not only would a wierd prop get used, the LD had no idea there was something better to use in the first place.
I just started working in Maya like three months ago. I didn't read any tutorials, a friend just showed me how when you press spacebar you can do most of the stuff you need to from there.
So I start an average male human character in order to get used to working with Maya, I don't even know what the fuck the attribute editor does, and after half managing to finsih the model, scultp the high poly and unwrap some uv's using roadkill, I now can't understand how the hell you're supposed to put textures on shit in maya (truth be told nothing's easier than Max's material editor). I can't be bothered to look for tutorials and I'm just going about it in trial and error.
This the end result.
Hopefully, after a couple more days of going about it via all the wrong ways, something good will happen.
If this isn't a fucked up way to go about things, I don't know what the fuck is.
/edit oh by the way, I don't know what any of the menus in zbrush do, I just press space there as well. I heard there's a way to unwrap shit from zbrush and then you can polypaint or something on the high rez and then take that and put it on the low rez?
Fuck that shit - I might consider pluging the wacom in to paint the textures in photoshop. Then again, I might not, and just use the mouse.
/edit 2: oh and I also save every wip piece (exports, textures, maya files, everything) on my desktop. I then find out what the last thing I worked on was by exploring my desktop folder viewing things by "date modified".
I think what he's saying is that its hard to find a point to your post. All i've managed to come out with after reading is that you have no idea what you're doing. If you can write up some specific issues your having, or specific things you're doing that you think are stupid, it would be easier to respond.
So, you press the space bar? Really thats all i know after reading your post. And that you're lazy and cant find tutorials on your own.
Well, both my posts were tongue-in-cheek, and while I do work with a fair bit of trial and error, I do actually use tutorials :P
Seriously now, I was actually thinking of starting a thread about the whole saving and naming files - this is a real issue in my workflow. So, my real question would be, how do you guys save your files?
...but I do use a fair amount of spacebar truth be told.
Blaizar & Perna: Which retopo tools are you talking about? ZBrush?
I just lately ended up making new topology for an entire character mess, without having a zphere cage and that surely was slow as hell. I think it took me 15 hours to just click to have basicly the same mesh I already had done. (I got corrupted subdiv history and couldn't re-obtain the low-poly meshes.) It was also sort of a test for me how fast it would be working something so big with zbrush retopoly.
But I think it's not comparable to another approach, which I would take now that I have made the mistake. To build a fast cage with zspheres and then just modify the more problematic regions with retopology tools. I don't have a lot of experience yet of doing things this way, but atleast the tests that I have done made me think it could be really a time saving manner.
Perna: Would you mind explaning about your retopo-workflow?
Recent thing, working on two objects independantly, low and high res.
Missed out a part on the low res so made it in the high res file as a high res part then a low res part, saved file. Opened low res file and merged this low poly part but neglected to realise that the position of the low res mesh was offset from the high res. So there's probably a work around somewhere...
Often times there are times where you need to model something / make something that makes sense/ is realistic/ or realistically works.
But sometimes the tasks given to you there no real ref's / guides to follow. What do you do when your in this kinda situation?
heres a example of what I mean (Actually 2)
1: a while ago I was contracted to work on a military style cargo plane, (a 52 something number cant remember at the moment) but then I Was told "it needed to work functionally"
Now the issue is there were only maybe a handful of images of this online and no spec maps etc to look into so alot of it was purely guessing, in fact blueprints were not online since its a currently used military craft, all that was online were some photos. What do you do in a situation like that where you have a ROUGH idea of something but not the specifics but need the specifics to make it work?
2: This is more recent , ive been asking this question around online, but Im currently trying to make a medieval style blacksmith building . Now you can find some info about things like this on images on google and such , but the specifics on how it works/ what tools were used/ exact layout to temper steel I have found little or no info on. So to make this "functionally work" there's no real info? Since in this example blacksmiths is a dying art, where do you go to find the info necessary to help you complete your task?
normally my work flow for these situations is look for every image under the sun possible on the internet and any info based off Wikipedia and any friends via messenger who would know anything.
But what do you do when all those sources fail you? What is the best alternative to make something thats close to it/ works somewhat functionally but you have no real ref of it?
And what do you do when your on a short time frame and you cant blame it on a intern X_X
Fudge it. Fudge that shit as hard as you can dude, i mean, this is for a game right? Make it seem like it would work. Thats the best you can do with shit ref, and nobody explaining how they want it to work in the first place. You're not making cad model's, you're making game art. Nobody will die if you model a panel slightly too thick or something.
Fudge it. Fudge that shit as hard as you can dude, i mean, this is for a game right? Make it seem like it would work. Thats the best you can do with shit ref, and nobody explaining how they want it to work in the first place. You're not making cad model's, you're making game art. Nobody will die if you model a panel slightly too thick or something.
thats where that artistic BS major comes into full play dosent it?
Indeed, that's where the "Artist" in you needs to come out.
When all is said and done, it doesn't matter if your blacksmith shop is 100% accurate to one found in England in 1542. If It doesn't LOOK right, then it's not right.
So in this case, it's the opposite. Make it LOOK right. Who gives a shit if the the forge should have been 3' to the right. If it LOOKS like it would work, and you have the basic elements of what 90% of people expect in a blacksmith shop (anvil, forge, tools, bucket of water) then that's all that really matters. The 2% of people who are going to call you on your shit being out of spec aren't worth it.
wow after all that about my points being nonsense, I'm glad the actual author might be able to get something out of an alternative viewpoint.
I hate this attitude that you can't do form and function. People are just unwilling to explore it because it's too much effort. Not directed at you Boogotti, im talking about a general vibe i get in game art. Like, people can shun technical correctness, but they should at least experiment with it before they completely dismiss it as some degenerative approach to design which heeds invariably poorer results.
I guess I'll just keep making the point that, even if you're making a fantasy model of an orc, people will still talk to you about the anatomy of the thing and making sure the muscle flow is correct. Perhaps i like a bit more "sci" than "fi" in stuff, but I think part of the whole appeal is that someone's thought about a design and there's no otiose details, everything has a function, and if they manage to pull a nice form together without having to sacrifice simplicity and function, it is a truly great design.
Perfection is achieved, not when there is nothing more to add, but when there is nothing left to take away.
Sorry to go off on that tangent, but I'm quite devoted to supporting this school of thought because I see it as underrepresented.
Again, sick work Boogotti, and I'm glad you're keeping an open mind, even if you don't alter the design at all based on my crits, the fact that you read and considered them can only be a good thing.
Honestly, unless you're an engineer, you drive a tank, or you have way too much time on your hands and jerk off to tank manuals, how would someone know to even think about most of the points you brought up. Now your points make sense after the fact, but for someone who knows nothing about freaking tank engineering, really what do you expect?
You want people to go out and do all the R&D, design, test and build a tank before they can consider making a model? Really, this sort of mentality that you seem to have is just ridiculous and is quite unreasonable to expect.
Do you want to be completely anal retentive when you make a model? That is fine, most people have better things to do with their time.
This school of thought is underrepresented because it a gross *waste of time*, especially in any sort of professional environment(unless you're a tank engineer).
"Incoming projectiles ricochet when they hit armour" is a concept that has been true since armour's first use in warfare. Even old plate mail is designed in such a way as to direct projectiles AWAY from the user.
Surely it's not very difficult for anyone to grasp the concept of a moving thing bouncing off a static thing, and that if you have a "nook or a cranny" when the moving thing hits it it will actually be guided into its target.
Like, this isn't fucking phD shit here.
Again not directed at you specifically Boogotti
edit: I'd also like to point out there's no reason to imply i'm some kind of nerd or wanker even if this stuff WASN'T something you could derive from a couple of minutes thought or a quick skim over wikipedia's article on armour. And I still put it to you that nobody would bat an eyelid if this same argument was over anatomy.
Replies
Zbrush models are mainly static works and in most cases:
-you part from a bad base mesh to speed up work, useless for rigging, animation, etc. (a fast and dirty one)
-you don't care about topology, and at the end you could need to use retopology tools, something terrible for our workflow (face loops, muscle loops, etc.)
-you don't have a good base mesh with good uv mapping and deformation. Does anyone test their meshes for a good deformation? doing inorganic shapes only... i suppose the response is no.
-you can have a "mucus" model with nonsense wrinkels, folds and silly details not needed.
-etc.
Subdivs are not only for mechanic objects, that's something i want to say, and you can find in organic modelling as technical issues as with inorganic shapes, maybe more.
Talking about modelling, there is only artistic side if you really know how to model.
WTF, it looks like Zbrush is the unique tool for characters nowadays and the work with characters is so easy... and that's not true. As much, in a real production, you can use Zbrush only to give a fine detail pass to generate a bump or a displacement for a still render.
Check out this link, i'm sure WoW addicts will find it cool hehehe: http://www.cgchannel.com/news/viewfeature.jsp?newsid=8144&pageid=0
With inorganic shapes, one done, all done, is always the same mechanic thing, you learn the technique and that's all, you don't need to know anymore. And if you have a library of mechanical or inorganic subdiv meshes, all is copy and paste. With normal mapping floats is more of the same, making subdiv models like a weapon is easier using floats, and without floats too, is all technique and experience IMHO.
You should understand what i say if you only use Zbrush or Mudbox to add the fine detail you can't add with polygons. What I'm talking about is about Pure Subdiv modelling... hi-res modelling of characters for animation (a character also has inorganic shapes attached). It's a harder work but easy aswell.
3D is not only Game Art, and models for normal maps. Zbrush can help a lot but not much when you need to build a hi-res model for animation. With too many models you don't need Zbrush, and all is Subdiv work.
Another thing is that the people who paint realistic forms on a paper or canvas have better predisposition to do greater 3D. is fully proved!. We are not special for making 3d, anyone can do it, and it's bacause the majority of the common 3d tasks are mechanic, can be done by TOOL guys, 3d monkeys, scripts, etc... it's the reality.
(i'm saying we, but actually i never do any high poly work hehe :P).
also... YOU SUCK!
Things are not linear anymore. You can use Zbrush, Mudbox, 3DCoat, heck even the free, loveable Organica to create striking silhouettes for characters (hence, NOT just surface details.). Even harsh primitives mashups work very well for this first step (I do that all the time now, then take that in Mud to wedge stuff together).
Then you can go many ways. Either detail the crap out of that step, or roughly retopo and reproject to go on sculpting. Then, you make your proper support mesh, whatever that is (subdividable cage for film displacement, or lowpoly 'outshape' for game art).
The beauty of todays tools is that nothing is locked anymore. Such a cool thing to embrace!
Using sculpting as a kind of example on how things are changing... some companies actively hire modelers purely for sculpting. That's all they do. Lower rez animation cages are then built by a more tech-savvy artist, UV'd and skinned to a rig, since there is less 'artistic' ability required to reproduce the low rez model from a high rez. I'm not saying this is a right or wrong method, but simply that it IS a method that people are now able to use, where before it generally wasn't.
I don't understand half of your posts but this quote is simply misinformed if not ignorant. To say inorganic modeling is all copy and paste is a load of crap.
I must think all comes due to my different way to do things. let's put a round of images so you could understand me better. 1 image = 1000 words hehehe
This is what i mean with a bad base mesh. Retopo tools are something bad for me, only lazy ppl uses it imho, you need to redo the base mesh when you could have starting doing something well, like a MAN does LOL. It's time you loose.
And this is how i make subdiv models (with a concept done previously).
The work is done quite fast... but it has its work, its time. It's not so "easy" to do as the crap base mesh i posted, which was done in 2 min and maped in 1 min (bad uv, bad seams, bad all). If you are going to do a 3d doodle, a model for normal maps... i undestand it (we work in the fast way), but for real production for animation.. start sculpting a model from a bad base mesh is time consuming! and time=money.
For me, is the same to model hi-poly or Low-poly, 3dwork is 3dwork, you just put in practice all what you know. When i'm modelling a hi-res model, i start from a low low low poly model as you may see, nothing about poly to poly or edge extruding.
Here's another example..
I said: "And if you have a library of mechanical or inorganic subdiv meshes, all is copy and paste". I didn't say inorganic modelling is all copy and paste dude ¬¬ read well. i'm using copy paste all the time, i make things very fast. Do you have a bad workflow? hehehehe
bad forum to talk about subdiv modelling, not so many hi-res modellers here hehehe :poly121:
PD: actual game models, are quite similar to subdiv models.
A lo mejor en español se me entiende jejejeje
Of course a model intended for animation needs to have a proper structure. But trying to hit that structure right from the start is not necessarily the only way to go. As a matter of fact, it *used* to be - before sculpting package appeared.
The point of Mud and Z is not just crazy zspheres or basemesh monsters, nor just high detail sculpting. It is also a fantastic way to explore shapes and proportions. That freedom cannot be reached with the traditional use of soft selection tools.
Also you seem to assume that a model is always made from a precise, locked concept. This is not always the case. This makes the free form approach of sculpting packages even more valuable.
Moreso : it is actually great to have a dirty sculpt as a base, even if the final product is to be a carton-like animated movie character. You can sculpt and adjust the shapes nicely and freely in mud, then create the clean subd model afterwards (using, guess what? retopology tools. Try Topogun). It works extremely well and saves you from hours spent trying to find a flow that doesn't pinch, stuff like that.
I wont even comment on your opinion on mechanical modelling. Seriously, look up some examples and think about it some more.
White boxy dragons get boring after a while
LIAR! Maya doesn't have the Utah teapot!
I've worked several times doing retopos from sculpts, and takes more time, it's a nuisance. It was another experience working with some new tools, and that's all. For me it's not the best way to do things.
I usually explore shapes and forms in my mind, and i use the paper to "record" the idea, it may be very very old school but it's fast enough and the best for me. I make a draw/concept in minutes... why should i need to do a 3d doodle having an idea so solid? is to loooose time, seriously.
Don't think i assume that a model is always made from a precise, locked concept. Sometimes i model from an idea in my mind.. and i simplily build the model in 3d without any concept. For doodling, i prefer pencil and rubber, i'm a traditional guy. And again, due to lack of time, i avoid to do concepts too many times, but i always know what i'm doing.
It seems like almost all ppl here takes the long way, and i think is because you don't have a solid idea in mind. Doodling is fun, but believe me, you really lose time, too much. And Retopo is bad for a good workflow imho.
I'm not so kick-as perna, i run into bad proportions issues all the days because i still don't see some forms/volumes. As artist, i'm very green, as much as my fucking plant lol (I feel ashamed of what I did the day before). My eye is not so good, but i can say that technically (modelling), i haven't got so many problems. I only know how to use the tools i use and i'm constantly growing as artist.
Don't know how explain it better, i thought it was something very simple but it seems i'm the only one who can understand it hehehe. I will end saying this again: to make an human, is far difficult than making mechanical or inorganic model, that's all (at least for me). You got the uncanny valley theory and that's enough proof of how challenging (difficult) is to create a credible human.
I think it's odd that you find retopo'ing to be the long way.
Blaizer it seems as though this thread was made for you.
now this is classy:
#include "colors.inc"
camera {
location <7.5, 20, -20>
look_at <0, 5, 0>
}
plane { // the floor
y, 0 // along the x-z plane (y is the normal vector)
pigment { checker color Black color White } // checkered pattern
}
sphere {
<0, 10, 0>, 4
pigment { color White }
finish {
reflection 0.7
phong 1
}
}
light_source { <0, 20, 10> color Yellow }
light_source { <10, 20, -10> color Red }
light_source { <-10, 20, -10> color Blue }
Ok it took me a while to find this post under all the faggotry, but i'll give it a shot.
Since i dont really have much experience working on the style of games that you do(lots of prerendered stuff, right?) I cant really offer much from a critique standpoint.
Instead, lets try flipping this around. Could you make a little writeup about what the pitfalls and problems are with your workflow?
now that i've offended you all, its time for workflow.
*reference gathering.
*more reference gathering.
*check polycount, check webcomics, check various military modeling sites, check youtube.
*watch 2 hour documentary on chimps.
*check polycount again
*even more reference gathering.
*go to sleep, wake up next morning, check polycount.
*start modeling, look at reference
*look at reference some more, check polycount, check youtube.
*go to sleep again.
*start some actual modeling. get good progress.
*look at reference, check polycount, check youtube, curse, saying to myself i shouldn't be wasting time on the net.
*get depressed over lack of progress.
*quit modeling, saying to myself i'll finish it at a later time.
my trim method sucks and it's a pain in the ass to unwrap:
*Make big room
*select faces to trim, inset x units and extrude down x units
What is all this check polycount business, sounds like a terrible waste of time.
Alright, since vig went through the trouble of either finding or making this silly flow cart, i'm going to post a few of my thoughts on prop workflows.
Now, first let me acknowledge that every project is different, and engine as well, so some principals may not always apply. But i would like to think so of these ideas are fairly universal.
Modeling in sets.
Whenever possible, i find that the most efficient way to make any prop is to make them in a set, now sometimes you need to do a one-off prop, or a few singular assets that are very generic and could be placed everywhere, but i think those cases are less common than people would like to believe, or should be at least.
Lets look at this from a practicle time standpoint:
Lets say on monday I give an artist a prop to make, he goes through, makes the high, low, uvs, bakes, textures, gets that all done monday.
Great! now lets say i give him a similar job tuesday, wednesday, etc. By the end of the week i should have 5 props, right?
Ok, now lets say i give an artist a set of 5 props to make, he does all the highpoly models at one, low, uv, bake, texture, all of these in phases. Chances are he's done by wednesday, or thursday at the latest. Thus saving 1-2 days on every small set of props that need to be developed.
Not only that, but his results are going to be more consistent, and cohesive, he wont spend a day making a bolt, and then a day spending the same amount of time on some ornate light fixture, instead its easier to prioritize the amount of work that goes into each prop. Also, if you can manage to combine props onto a single uv set, you will end up saving UV space in the end, chances are if you've got 5 separate props, all with 512s, you could have easily fit all of those assets on a single 1024, saving you a good chunk of memory.
Now, onto the issue of grouping props onto one uv set. First off, i'm going to start by saying that if you're working efficiently, spending a
few more minutes to better plan what will be needed, and work in groups, any cleanup time will likely be negetated, and you'll still be
making assets faster than if you were doing them one by one.
Ok, so when should you group assets?
In my opinion you should do this whenever possible, the speed gains you will get from being able to do all of your bakes at once, and texture
them all at once should not be understated. This shit saves a lot of REAL WORLD production time.
Now, i understand that it will not always make sense to combine a group of props onto one sheet, but i think if you're smart about it, it
becomes a pretty simple decision.
Here are some examples of situations where i would combine multiple props onto one sheet.
Say you have a set of very similar props, for example lets say you have a computer, a mouse, a monitor, a keyboard and a printer. Really,
this is one cohesive set of props, that SHOULD be placed in the same vicinity, its one prop set that is split up into multiple pieces
mostly just to give better variation in placement.
When you have one type of object, but with multiple variations, lets say a few types of crates, or a fence that has a few different peices. This is going to be a good time to combine objects, because it is safe to assume they will be used in a similar area.
There will be plenty of instances where you will not want to group objects, throwing something small that will be reused a lot onto a large, unique prop that will be used very little is a good example of where you REALLY DO NOT want to combine objects.
Now, onto the issue of set dressing.
So, lets say "Dave" is an idiot, and he takes your nice computer set, and only place mice around his level, 47 mice infact.
Does this mean that your grouping system is totally wrong? No, it simply means that "Dave" is a fucking moron.
Some things we can do to stop these situations before they start:
Spend some time, work with the LD/Set dressers. Make sure they understand that you want to have "themes" to your areas in the game, and that these themes should likely be set dressed with appropriate objects, that actually make sense to be in the area in the first place. The more cohessive and common sense your prop placement is, the better your game will feel.
Spending the extra time to be organized, and have a more unified goal/vision on how your scenes will be set up will be a major boost in the long run.
Now, there will inevetibly be cases where not all props in a set get used. Is the only solultion then to split each of these up, redo the uvs and RTT the textures onto a new sheet? Absolutely not, this is the worst case scenerio, and something that i would only do if say, only 1 or 2 small pieces of a 5 set prop sheet were be used.
If only one piece out of a set is being used, chances are this object is OUT OF PLACE, and can be easily be replaced by a prop that fits in this environment better, this should be an easy fix in most engines, much easier than reworking entire sets of assets.
Did "Dave" place a wire spool, mouse, crate, and bag of doritos all in one spot, in one level, where none of these props or similar props in the set are being used? The easily solution here is to fix "Dave"s work, not rework all of the assets.
So, work more efficiently, take some of that time you save by doing so and spend more of it planing what you will actually need, and working with the right people to make sure there isn't massive misuse of assets, and the world will be a much better place.
I think its better to put the unknowns on their separate sheet instead of tagging them into the nearest wasted space and hope that no one places them as a separate object. You can easily tag unknowns together later after the dust settles and you know for sure if props where placed "correctly".
I think there needs to be another step or procedure in place other then "I told the LD to place all these light posts together but not those" A naming convention, a flag in the editor a way to group or sort the props something other then just "Remember 6mo ago I told you the rat and garbage piles go with the wardrobe boxes!" It's not only Dave the moron but maybe a few other people that weren't told. Having a process makes that process doable and favorable I think in most cases, provided its there, and not on the "remember-system"
you should make textures that are SOOOOO useful that they can go on practically anything and everything else can be dealt with geometry/vertexcolors/lighting.
so don't put all sorts of object sheets next to each other ... instead use ONE AND THE SAME sheet on everything
and than bacon !
So I start an average male human character in order to get used to working with Maya, I don't even know what the fuck the attribute editor does, and after half managing to finsih the model, scultp the high poly and unwrap some uv's using roadkill, I now can't understand how the hell you're supposed to put textures on shit in maya (truth be told nothing's easier than Max's material editor). I can't be bothered to look for tutorials and I'm just going about it in trial and error.
This the end result.
Hopefully, after a couple more days of going about it via all the wrong ways, something good will happen.
If this isn't a fucked up way to go about things, I don't know what the fuck is.
/edit oh by the way, I don't know what any of the menus in zbrush do, I just press space there as well. I heard there's a way to unwrap shit from zbrush and then you can polypaint or something on the high rez and then take that and put it on the low rez?
Fuck that shit - I might consider pluging the wacom in to paint the textures in photoshop. Then again, I might not, and just use the mouse.
/edit 2: oh and I also save every wip piece (exports, textures, maya files, everything) on my desktop. I then find out what the last thing I worked on was by exploring my desktop folder viewing things by "date modified".
I thought the point of this thread was to critique our asset creation workflow, not our posting one?
I think what he's saying is that its hard to find a point to your post. All i've managed to come out with after reading is that you have no idea what you're doing. If you can write up some specific issues your having, or specific things you're doing that you think are stupid, it would be easier to respond.
So, you press the space bar? Really thats all i know after reading your post. And that you're lazy and cant find tutorials on your own.
Seriously now, I was actually thinking of starting a thread about the whole saving and naming files - this is a real issue in my workflow. So, my real question would be, how do you guys save your files?
...but I do use a fair amount of spacebar truth be told.
I just lately ended up making new topology for an entire character mess, without having a zphere cage and that surely was slow as hell. I think it took me 15 hours to just click to have basicly the same mesh I already had done. (I got corrupted subdiv history and couldn't re-obtain the low-poly meshes.) It was also sort of a test for me how fast it would be working something so big with zbrush retopoly.
But I think it's not comparable to another approach, which I would take now that I have made the mistake. To build a fast cage with zspheres and then just modify the more problematic regions with retopology tools. I don't have a lot of experience yet of doing things this way, but atleast the tests that I have done made me think it could be really a time saving manner.
Perna: Would you mind explaning about your retopo-workflow?
Missed out a part on the low res so made it in the high res file as a high res part then a low res part, saved file. Opened low res file and merged this low poly part but neglected to realise that the position of the low res mesh was offset from the high res. So there's probably a work around somewhere...
Often times there are times where you need to model something / make something that makes sense/ is realistic/ or realistically works.
But sometimes the tasks given to you there no real ref's / guides to follow. What do you do when your in this kinda situation?
heres a example of what I mean (Actually 2)
1: a while ago I was contracted to work on a military style cargo plane, (a 52 something number cant remember at the moment) but then I Was told "it needed to work functionally"
Now the issue is there were only maybe a handful of images of this online and no spec maps etc to look into so alot of it was purely guessing, in fact blueprints were not online since its a currently used military craft, all that was online were some photos. What do you do in a situation like that where you have a ROUGH idea of something but not the specifics but need the specifics to make it work?
2: This is more recent , ive been asking this question around online, but Im currently trying to make a medieval style blacksmith building . Now you can find some info about things like this on images on google and such , but the specifics on how it works/ what tools were used/ exact layout to temper steel I have found little or no info on. So to make this "functionally work" there's no real info? Since in this example blacksmiths is a dying art, where do you go to find the info necessary to help you complete your task?
normally my work flow for these situations is look for every image under the sun possible on the internet and any info based off Wikipedia and any friends via messenger who would know anything.
But what do you do when all those sources fail you? What is the best alternative to make something thats close to it/ works somewhat functionally but you have no real ref of it?
And what do you do when your on a short time frame and you cant blame it on a intern X_X
(please dont tell me the answer is bullshit it)
thats where that artistic BS major comes into full play dosent it?
When all is said and done, it doesn't matter if your blacksmith shop is 100% accurate to one found in England in 1542. If It doesn't LOOK right, then it's not right.
So in this case, it's the opposite. Make it LOOK right. Who gives a shit if the the forge should have been 3' to the right. If it LOOKS like it would work, and you have the basic elements of what 90% of people expect in a blacksmith shop (anvil, forge, tools, bucket of water) then that's all that really matters. The 2% of people who are going to call you on your shit being out of spec aren't worth it.
I hate this attitude that you can't do form and function. People are just unwilling to explore it because it's too much effort. Not directed at you Boogotti, im talking about a general vibe i get in game art. Like, people can shun technical correctness, but they should at least experiment with it before they completely dismiss it as some degenerative approach to design which heeds invariably poorer results.
I guess I'll just keep making the point that, even if you're making a fantasy model of an orc, people will still talk to you about the anatomy of the thing and making sure the muscle flow is correct. Perhaps i like a bit more "sci" than "fi" in stuff, but I think part of the whole appeal is that someone's thought about a design and there's no otiose details, everything has a function, and if they manage to pull a nice form together without having to sacrifice simplicity and function, it is a truly great design.
Sorry to go off on that tangent, but I'm quite devoted to supporting this school of thought because I see it as underrepresented.
Again, sick work Boogotti, and I'm glad you're keeping an open mind, even if you don't alter the design at all based on my crits, the fact that you read and considered them can only be a good thing.
You want people to go out and do all the R&D, design, test and build a tank before they can consider making a model? Really, this sort of mentality that you seem to have is just ridiculous and is quite unreasonable to expect.
Do you want to be completely anal retentive when you make a model? That is fine, most people have better things to do with their time.
This school of thought is underrepresented because it a gross *waste of time*, especially in any sort of professional environment(unless you're a tank engineer).
Surely it's not very difficult for anyone to grasp the concept of a moving thing bouncing off a static thing, and that if you have a "nook or a cranny" when the moving thing hits it it will actually be guided into its target.
Like, this isn't fucking phD shit here.
Again not directed at you specifically Boogotti
edit: I'd also like to point out there's no reason to imply i'm some kind of nerd or wanker even if this stuff WASN'T something you could derive from a couple of minutes thought or a quick skim over wikipedia's article on armour. And I still put it to you that nobody would bat an eyelid if this same argument was over anatomy.