1 - I'm going to replace my terrain with a static mesh, for better results
2 - Going to look in to changing my tree foliage
3 - Going to use the mesh placement tool to replace my grass once I've swapped my terrain out for statich mesh.
Also! Super Pro-tip: I started launching Photoshop, Mudbox and UDK from STEAM! This way I can log my hours I'm putting in each week. Woot!
Also! Super Pro-tip: I started launching Photoshop, Mudbox and UDK from STEAM! This way I can log my hours I'm putting in each week. Woot!
Don't let the man know how much time you use your programs....
I haven't checked in to this thread in a few days, nice progress going on here. Only crit I have is the water, but that's been covered (looks like the lilypads are just in air and not the water).
Did you look at the water materials in the UDK? They do a decent job of ripple effects, although they do look like they are meant for things like lakes or ponds.
I think he's talkin 'bout using the edit normal modifier in Max.
don't think unreal picks up edited mesh normals. Had a right pain with it once, don't think it works. Maybe it's fixed now though.
Turning works, of course.
Question: With all the duplicate meshes used, does that automatically count as "instanced geometry" or is there a button/technique that needs to be pressed/used?
from what I know, instanced geometry usually refers to the mesh`s impact on the memory footprint.
a good way to think about it is a balancing act- you have three budgets: tris on screen
draw count
memory
Whenever your building a scene you have to be very aware of this:
You cant have too many 1 off meshes. This is a poor use of the memory. You`ll quickly find that model budgets are smaller than you think and game art is sort of about re-use and recycling.
you cant have too many individual assets placed in the scene because that kills your draw count
and you cant have assets that are super high poly because that kills you tri`s on screen budget.
I think instancing refers to reusing a mesh over and over because that's the best approach to keeping your memory budget from going over, but you have to be careful because you run the chance of killing your draws.
I have heard of another form of instancing where you can duplicate a mesh as much as you want, and the engine doesn't create more draws, but it had the restriction of the asset in question needing to be extremely low poly (under 20 tris) so it was great for small plants and shrubs
but this is all radiant speak. I`d be curious to find out how unreal differs, and conversely how its the same
adam the work you have been doing here is priceless. I have taken a lot notes from this thread and it keeps getting better. Thanks for sharing all of your work.
iansmithartist - Oh ya? What've you learned so far?
workflow ideas, ways of breaking down and digesting a large scene, its the parts in-between that has always thrown me. I can make a prop or a texture sure, but knowing how to bring it all together and making it work, now that's another kettle of fish.
Its also humbling to see someone who is in the industry still asking questions and taking c&c. :thumbup:
Its also humbling to see someone who is in the industry still asking questions and taking c&c. :thumbup:
Oh he didn't get into the industry by working, I bet he did some....favors.... yeah, we'll go with that....
Just kidding, I agree with ian here. I'm learning a lot from this. More people should break their work down as much as it is here. It's a lot easier to follow and I feel like I'm getting more out of it overall.
iansmithartist - Awesome. I'm really happy to hear you're finding this thread useful. The idea with it was to 'break barriers', as a professional, and really open up as a "WIP" thread. Any professional who tells you they know everything about their craft is full of shit
DarthNater - While I agree, its quite an undertaking and probably viewed as something thats not worth their time. It's much easier to post images and wire shots and let people figure out the rest for themselves.
As well - and this goes for everyone - when you post WIP's, if you break things down you actually learn a bit yourself in additional to showing others a thing or two.
Adam's last point is absolutely true. As someone who's working at improving his skills to get into the industry, my WIP thread here and my blog that I started keep me motivated and sane has done wonders to help me grow professionally and personally. I've learned a lot from both threads like Adam's and my own thread (I'm looking at you Matroskin). The best bonus is seeing how far something has come. It's a great feeling of accomplishment, and knowing you have much more to learn just makes you want to keep going.
Hey adam... great thread. I was curious as to how you went about getting the atmospheric fog in the editor. I'm shooting for something similar in my "Old Bait Shop" Scene. Amazing progress so far!
So this past week or so I've been busy (read: relaxing and taking it easy not doing 3D at home) as I wanted a little break (sue me!). But today I jumped right back on the horse and busted out the full-size fence piece I'll use to do the perimeter fence seen in the original concept.
The idea behind this process is this:
Even though I will have 1 full size and 1 high size piece for this fence, everything I'll need can be achieved with just the full size being made in highpoly.
The shot below has the model completely made. This is simply for pimping reasons. When I go to bake it, the model will be broken down to its "lowest common denominators" and baked to a low poly of the same sense.
I'll post pictures of this tomorrow.
Don't forget this is a portfolio piece in addition to whats happening here on Polycount. I don't mind spending the time on something being shown in its complete form if the end result is decent enough to show on the portfolio.
As for the highpoly itself:
All soft edges are typically softer (read: rounder) than I'd normally do for a realistic piece. I'd like the thicker sizes to lend themselves to the style I'm trying to go for.
I wanted a couple of subtle comical elements to be present in this piece.
Dr. Q purposely has the razor wire lowered to face level as it'll increase the chances of doing facial damage to his opponents. Since he himself has his face in a glass bubble he'd like others to be apprehensive over their face as much as he is.
The drawback being how easily this can be bypassed, but how can one think of that when there are faces to be ripped?
The gap between the fences are big enough for a person (read: Spy, Special Agent(s), etc.) As the image says, he really should fire his head of security over this one.
I won't do a lot of local detail or damage in the highpoly. If this was a realistic piece I'd subdivide it evenly, bring it in to Mudbox, and get some local details happening. But for now, for this particular style, I'm happy with achieving that inside of Max.
The lowpoly will be a fence set of a few pieces (1 full, 1 half, and variances of each) that are all derived from this one highpoly mesh.
Which we'll see when I post pictures tomorrow.
Enough chatter, here's the highpoly. Check it out unmarked and in JPEG format here.
Is the diamond grate fence part an alpha mapped texture or is it actually high poly geometry? If it's geometry could you please do a quick breakdown of how you achieved it? I really need to start doing some more advanced modeling and knowing how to make a nice grate/fence would be great to know.
Can't wait to see your next update! I need to go back through this thread and write down some notes. There's a lot of good tips and techniques in here that would be very useful!
Is the diamond grate fence part an alpha mapped texture or is it actually high poly geometry? If it's geometry could you please do a quick breakdown of how you achieved it? I really need to start doing some more advanced modeling and knowing how to make a nice grate/fence would be great to know.
I made some chain link fence pieces like that a while ago, the idea is to make part of the chain that you can just duplicate over and over to make the whole fence. Its basically one cylinder that bends forward then down in a V then bends back and goes up, I could whip up a quick image tutorial.
ZacD - Feel free I'll do up my own either tonight or tomorrow. I'm watching a video on Imbalu at the moment which is quite interesting, we'll see if I feeling like opening Max when its over.
Is the diamond grate fence part an alpha mapped texture or is it actually high poly geometry? If it's geometry could you please do a quick breakdown of how you achieved it? I really need to start doing some more advanced modeling and knowing how to make a nice grate/fence would be great to know.
Im sure there are tons of ways to make this, but here is how I would do it.
The scene is looking great Adam, it's is very cool that you are continuing to provide such detailed updates. I am still anxiously awaiting the appearance of the Owl turret .
The process I used to get my spline shape was different than yours but that's arbitrary. Your tutorial is pretty well 'it' for how I made my chain-link fence. Thank you for writing that up!
Wow.
Every time i visit this thread there is new found awesomeness to be had!
Even for a totalitarian regime, i think masses of face level razor wire is particularly harsh...i guess that just how the Evil Genius Rolls. ^^
I was making a highpoly chain link fence a while ago and found that when I had that much 'chain' I had an un-fucking-believable amount of polys on screen. How high poly are your chains? anything turbosmoothed? my bends looked awful without it.
I was making a highpoly chain link fence a while ago and found that when I had that much 'chain' I had an un-fucking-believable amount of polys on screen. How high poly are your chains? anything turbosmoothed? my bends looked awful without it.
Look at nrek's reply, you shouldn't need to smooth it beyond that.
But there's actually something wrong with nrek's fence, its more of a weaving pattern with the chains going forward and back, I'll show you what I mean.
Chain fences have a high point and a low point, the weaving only happens on the corners.
The blue is the lowest part, the red is the highest.
Just went back through the entire thread and wrote down a few notes! There's a lot of great information in here. Your rock technique should be a great help once I finally start working on a TF2 level I've been wanting to do for 2 years.
Unreal pros: Any optimization tricks besides LOD's for models?
U can also use antiportals. It can be quite efficient considering that u have many hills and "rock walls" wich can contain antiportal inside them Though i haven't checked if they kept them in UnrealEn 3... Should be still there i hope.
Btw, looks great, but grass planes don't seem to be well integrated with ground surface. there is not enough smooth transition between ground and grass planes. Making alpha not straight on the bottom could help.
Nice fence! This is a cool thread indeed, I would have had no ideas on how to go about rocks like that. And all the concepts and planning, is nice to see your process.
While Sir Adam inspires me to do some kickass works of tri-dimensional art, Mr ZacD makes me want to do some nasty things to him.
On a serious note, this whole project is a great piece of study on workflow and a lot of other aspects of environment art for many of us! Thank you Adam for sharing this with us! I've been following this since the start, and have jotted down shitloads of matter in the process.
Hey guys - thanks! Unfortunately work & the new website have been keeping me busy so I haven't had much chance to work on it. I am hoping to get back in to it come next week.
Replies
1 - I'm going to replace my terrain with a static mesh, for better results
2 - Going to look in to changing my tree foliage
3 - Going to use the mesh placement tool to replace my grass once I've swapped my terrain out for statich mesh.
Also! Super Pro-tip: I started launching Photoshop, Mudbox and UDK from STEAM! This way I can log my hours I'm putting in each week. Woot!
Don't let the man know how much time you use your programs....
I haven't checked in to this thread in a few days, nice progress going on here. Only crit I have is the water, but that's been covered (looks like the lilypads are just in air and not the water).
Did you look at the water materials in the UDK? They do a decent job of ripple effects, although they do look like they are meant for things like lakes or ponds.
Hmm you sure? Cause I don't see my 200 hours in Microsoft Word :P
haha that's an awesome idea! So steam also logs the hours of non-steam applications you've been running??
Turning works, of course.
a good way to think about it is a balancing act- you have three budgets:
tris on screen
draw count
memory
Whenever your building a scene you have to be very aware of this:
You cant have too many 1 off meshes. This is a poor use of the memory. You`ll quickly find that model budgets are smaller than you think and game art is sort of about re-use and recycling.
you cant have too many individual assets placed in the scene because that kills your draw count
and you cant have assets that are super high poly because that kills you tri`s on screen budget.
I think instancing refers to reusing a mesh over and over because that's the best approach to keeping your memory budget from going over, but you have to be careful because you run the chance of killing your draws.
I have heard of another form of instancing where you can duplicate a mesh as much as you want, and the engine doesn't create more draws, but it had the restriction of the asset in question needing to be extremely low poly (under 20 tris) so it was great for small plants and shrubs
but this is all radiant speak. I`d be curious to find out how unreal differs, and conversely how its the same
-nice lookin scene adam :P
konstruct - thanks!
Picard - woo!
Any progress on those characters? i think i spotted a basemesh in previous screenies.
Edit: Does using the beta UI have anything to do with it maybe?
just use manic time. free. keeps track of my time in my apps.
workflow ideas, ways of breaking down and digesting a large scene, its the parts in-between that has always thrown me. I can make a prop or a texture sure, but knowing how to bring it all together and making it work, now that's another kettle of fish.
Its also humbling to see someone who is in the industry still asking questions and taking c&c. :thumbup:
Oh he didn't get into the industry by working, I bet he did some....favors.... yeah, we'll go with that....
Just kidding, I agree with ian here. I'm learning a lot from this. More people should break their work down as much as it is here. It's a lot easier to follow and I feel like I'm getting more out of it overall.
DarthNater - While I agree, its quite an undertaking and probably viewed as something thats not worth their time. It's much easier to post images and wire shots and let people figure out the rest for themselves.
As well - and this goes for everyone - when you post WIP's, if you break things down you actually learn a bit yourself in additional to showing others a thing or two.
Alright, cool. I figured as much Thanks.
The idea behind this process is this:
- Even though I will have 1 full size and 1 high size piece for this fence, everything I'll need can be achieved with just the full size being made in highpoly.
- The shot below has the model completely made. This is simply for pimping reasons. When I go to bake it, the model will be broken down to its "lowest common denominators" and baked to a low poly of the same sense.
- I'll post pictures of this tomorrow.
- Don't forget this is a portfolio piece in addition to whats happening here on Polycount. I don't mind spending the time on something being shown in its complete form if the end result is decent enough to show on the portfolio.
As for the highpoly itself:- All soft edges are typically softer (read: rounder) than I'd normally do for a realistic piece. I'd like the thicker sizes to lend themselves to the style I'm trying to go for.
- I wanted a couple of subtle comical elements to be present in this piece.
- Dr. Q purposely has the razor wire lowered to face level as it'll increase the chances of doing facial damage to his opponents. Since he himself has his face in a glass bubble he'd like others to be apprehensive over their face as much as he is.
- The drawback being how easily this can be bypassed, but how can one think of that when there are faces to be ripped?
- The gap between the fences are big enough for a person (read: Spy, Special Agent(s), etc.) As the image says, he really should fire his head of security over this one.
- I won't do a lot of local detail or damage in the highpoly. If this was a realistic piece I'd subdivide it evenly, bring it in to Mudbox, and get some local details happening. But for now, for this particular style, I'm happy with achieving that inside of Max.
- The lowpoly will be a fence set of a few pieces (1 full, 1 half, and variances of each) that are all derived from this one highpoly mesh.
- Which we'll see when I post pictures tomorrow.
Enough chatter, here's the highpoly. Check it out unmarked and in JPEG format here.Can't wait to see your next update! I need to go back through this thread and write down some notes. There's a lot of good tips and techniques in here that would be very useful!
Sorry I didnt read the whole wall of text, it was all about technical stuff.
polycount needs a yearly vote to ban one user.
I made some chain link fence pieces like that a while ago, the idea is to make part of the chain that you can just duplicate over and over to make the whole fence. Its basically one cylinder that bends forward then down in a V then bends back and goes up, I could whip up a quick image tutorial.
Keep up the good work.
ZacD - Feel free I'll do up my own either tonight or tomorrow. I'm watching a video on Imbalu at the moment which is quite interesting, we'll see if I feeling like opening Max when its over.
Im sure there are tons of ways to make this, but here is how I would do it.
The scene is looking great Adam, it's is very cool that you are continuing to provide such detailed updates. I am still anxiously awaiting the appearance of the Owl turret .
Every time i visit this thread there is new found awesomeness to be had!
Even for a totalitarian regime, i think masses of face level razor wire is particularly harsh...i guess that just how the Evil Genius Rolls. ^^
Great work adam
Look at nrek's reply, you shouldn't need to smooth it beyond that.
But there's actually something wrong with nrek's fence, its more of a weaving pattern with the chains going forward and back, I'll show you what I mean.
Chain fences have a high point and a low point, the weaving only happens on the corners.
The blue is the lowest part, the red is the highest.
Looks like adam did it the right way.
gotta love splines XD
i understand how you did the rocks but i'll have to try it to fully grasp it...fence is sexy
U can also use antiportals. It can be quite efficient considering that u have many hills and "rock walls" wich can contain antiportal inside them Though i haven't checked if they kept them in UnrealEn 3... Should be still there i hope.
Btw, looks great, but grass planes don't seem to be well integrated with ground surface. there is not enough smooth transition between ground and grass planes. Making alpha not straight on the bottom could help.
! +1
Quick noob question, what do you use for rendering those shots such as the fence?
On a serious note, this whole project is a great piece of study on workflow and a lot of other aspects of environment art for many of us! Thank you Adam for sharing this with us! I've been following this since the start, and have jotted down shitloads of matter in the process.
Cheers!
Woo!