Question for Prop Modelers
polycounter lvl 12
Hello everyone. I'm making an effort to optimize the amount of time it takes to create assets. I'm not to familiar with how long something should take. I spoke to a student at Gnomon who mentioned that a prop will take 2-3 days for an assignment. I have made 2 props so far. The second prop took me 6 days to make and that's after aggressively half-assing a substantial portion of the work.
2 days was unheard of for me. I literally did not believe her when she told me this yet consensus from other members on the forum made it abundantly clear that this was the norm. Students have other assignments they need to stay on top of. I live alone, play zero games and watch zero movies and only communicate via Skype and Facebook. On top of that I have all the time in the world and it still takes me way to long. It won't take a genius to come to the conclusion that I'm doing something wrong.
I would like to know why this is. So prop guys, how do you divide up your time? Do you do modeling on day 1 and texturing and lighting day 2? Do you have a schedule for how long each aspect of the pipeline should take? The reason I ask is because I plan on timing myself and comparing. I can isolate which aspect of the pipeline I am doing wrong.
Example: It used to take me an entire day to bake out my special maps but a user mentioned that this process only took him a couple of hours. I sent him my work and we compared workflows. The difference between me and him was the amount of time fixing cage errors in xNormal. The difference between our cages is that he triangulated the cage mesh and that somehow fixes all those "Voodoo errors" errors xNormal would throw back at me. I hope I can turn 6 days into 3.
2 days was unheard of for me. I literally did not believe her when she told me this yet consensus from other members on the forum made it abundantly clear that this was the norm. Students have other assignments they need to stay on top of. I live alone, play zero games and watch zero movies and only communicate via Skype and Facebook. On top of that I have all the time in the world and it still takes me way to long. It won't take a genius to come to the conclusion that I'm doing something wrong.
I would like to know why this is. So prop guys, how do you divide up your time? Do you do modeling on day 1 and texturing and lighting day 2? Do you have a schedule for how long each aspect of the pipeline should take? The reason I ask is because I plan on timing myself and comparing. I can isolate which aspect of the pipeline I am doing wrong.
Example: It used to take me an entire day to bake out my special maps but a user mentioned that this process only took him a couple of hours. I sent him my work and we compared workflows. The difference between me and him was the amount of time fixing cage errors in xNormal. The difference between our cages is that he triangulated the cage mesh and that somehow fixes all those "Voodoo errors" errors xNormal would throw back at me. I hope I can turn 6 days into 3.
Replies
It's good to identify where you are slow and how to improve, but at the end of the day, you'll get faster by practicing more, the more experience you have, the easier it is to solve common problems and troubleshoot issues.
You shouldn't be spending a day tweaking your cage, that seems like you have some fundamental problems when it comes to understanding how normal maps work. Check the stickied threads in the tech talk section.
But if I've been doing cages wrong this whole time, then that begs the question: what else am i doing horribly wrong? Next asset I'll log my time and upload my sheet.
Rather than tell you how much time I spend, I'll break down my basic process for hard surface stuff:
1. Blockout, get the proportions sorted out with basic geometry before you commit to detailing. The more detail you do here, the more time you can save in the highpoly
2. Highpoly, with the highpoyl already set, this is mostly a matter of adding support loops or modeling finer details, floaters, etc
3. Lowpoly, start from either hte highpoly cage, blockout, or remodel sections that are very complex in high but will have basic shape in low
4. Simple uvs, no final packing yet
5. Testbakes, bake sure you don't have any projection errors or any other issues, if so, cut in some geometry to fix in the lowpoly geometry, make any highpoly changes to improve bakes at this point as well
6. Final uv packing
7. Final bakes, including material mask bake (assign a colored material to each material type on the highpoly)
8. Base materials, get your material read down with simple mats, usually just basic colors for difffuse/spec/gloss, maybe some overlays for surface types that need it
9. Final detailing, wear, unique stuff, etc for the texture
I'm sure you get this a lot, but I'm a big fan of your work and contributions EarthQuake
Dude, of course it's going slowly. Once you've made two hundred the pace will pick up.
I know it's really tough sometimes. Myself i have been doing general environment art for two years now, and the past year i feel like i almost not progressed even though i have.
What's more likely going on is that you are still developing a feel for certain workflows, and don't yet have the type of decision making skills that allow you to quickly realise 'hey, this'll go wrong this way, let's do it that way': You're really deliberating.
What you can do about this is to do a screenrecording of yourself and then play in back in fast forward later to see the weak parts. But also to see what you do right: Maybe you are really good at UVs or something.
Then, after identifying it, it helps to figure out why the mistakes you make cause errors. If you have a bad cage, why is it a bad cage, can you visualise how the rays go with baking?
I don't think it's necessarily a bad thing to be a bit slower in the learning process. Experts tend not to be people who do everything right the first time, but rather the people who know how to deal with all the ways it can go wrong. Because in the field, things will go wrong and fixing it is the largest part of your day job.
Over time, that adds up.
To add. With 3DS Max and the double smooth method, try to build the low poly while you're doing the high poly. If there's a cylinder, build it with how many sides you want in the low poly.
That alone has saved a ton of time for me. In some cases I just turn off both of my turbosmooths and the low poly is finished.
Trademark that. Bake-Sure: The be all, end all texture map baking tool.
That's why I usually apply a shrink wrap on my low poly (targeting the subd high poly).. or is it a bad idea?
it's a good idea if it nets you a result you're after.
Why test bakes though? Once you finish the final UV important errors would be clearly visible.
I usually do
UV
cage,
test bake,
if there are any errors, if there are I'd add some blockers.
final bake.
Side Note:
What about leaves or other very flat surfaces? (I do see that you said that this method is for hard surfaces but I don't know the process of thin surfaces)
I know there is no simple answer, but on last interview I was asked how long it would take me to make a prop or a weapon.
Next-gen assets, highly detailed, like most of things in a production of AAA titles right now.
Unfortunately it seems I am too slow and now I wonder, generally, how fast top artists are.
Any straight answer?
EG: I want to Speed up my high poly creation.
Just find something, and make highpolies of it. 5-6 Just make a bunch.
Now you can't tell me that by the 3rd clone, you haven't gotten faster. 1 cause you know the object so well you could make it without reference, and 2. because you know what not to do.
Do the same thing for your low poly/retopo. Again notice by the 3rd or 4th clone you are miles faster than when you started.
Keep doing this for each step.
Then at the end, try to find what is is that still slows you down, try to google problem solve, post your art on PC, ask for tips. You gotta have a kind of lazy annoyance for things. Lazy people find shortcuts to help them reach their end goal faster, or more efficiently. Think, "I don't wanna take 10 minutes to do this, can't I find a way to do it in 5?"
You can then start setting yourself hard goals. Have 5 concepts ready. Say each prop is going to be 2 days of my time. Post it on polycount
if you don't finish a prop in the allotted time, stop working on it and move on. See how much faster you are by prop 5.
You have to do these kinds of exercises to get better, and then your speed will naturally get better.
I think you've nailed it on the head here. You've only been active on polycount for a month. For any assets you're working on now and in the future, post updates constantly on here. People should be quick to point out if you are doing something wrong. The easiest way to get faster is to get an industry gig cuz you will learn all the little things that make the process go faster. I actually don't think you should worry too much about speed. Quality is way more important for an artist trying to break into the industry. I was not the fastest before I got my first gig and I certainly wasn't very fast before getting my first next gen gig. But I tried to hit a certain bar of quality to give myself a chance at getting a job. Once you're in, you learn stuff REALLY FAST and you get 10x faster before you know it. I promise you, if you were to get a job in the industry, you would probably be doing those assets you fret about in 2 days before you knew it. So, while efficiency IS important, don't rush just to try and finish something that someone tells you they do in 2 days. Make it looks REALLY good and then figure out small things to help it go faster the next prop you work on.
http://3dmotive.com/series/the-briefcase-workflow.html
I find that I work fastest when I can conceptualize the entire process from start to finish at once. That takes practice but at some point you will be able to plan several steps ahead and optimize your effort for the end goal.
I have been using this n52 gamer pad te for the last 7 years I've worked in the game industry for example. I actually own 3 of them so I have one at each job site (I work at 2 locations) and 1 at home. I don't know a lot of people that use this sort of thing, but for me its really helped a lot with speed. Bascially I can map any key to it and it frees up my right hand for the mouse or stylus. Everything is accessible from my left hand alone in terms of shortcuts and to me its a lot more comfortable than having my hand at the keyboard. You have to go strictly off memorization cause the keys don't have letters, but once you get used to it its super fast. Anyway, this isn't for everyone, but its just one example how you can take time to improve your speed in the long run. It took me about a week to setup my gamerpad and get used to it, and I still tweak things every now and then.
But to become fast it just takes time. You get used to ui, workflow, and simply your brain starts working two steps ahead where you are unable to do that at first. If you have ever seen a really talented concept artist work, its like every stroke is perfect or they get from point a to point b extremely fast due to efficiency. Something like that just takes time and some are faster learners than others. I consider myself a tad on the slow side, but I just work more than others and it makes up for it. Keep at it dude, you will speed up over time!