I've started this thread since I need to get motivated on the new project I've started working on. The whole plan is to remake the whole city of Balmora in TES: Morrowind. I do this to boast my portfolio that I'm currently working on, but also to get loads of experience and an offer to work at Bethesda Softworks..
haha.
I'm currently deciding on which engine I'm going to use as well as the creation pipeline. The whole thought is to extensively use tileable maps together with low custom Ambient Occlusion maps, polypainting for variety.
This is what I got so far:
- References. (from Morrowind)
- Landscape mesh 20k tris (Needs fixing and crit).
- Idea for prop creation pipeline
- Great spirit.
This is what currently needs to be decided on:
- Engine.
- Creation pipeline needs to be set in stone.
Here's a quick Technical Sheet that I've done to illustrate my idea of pipeline:
http://dl.dropbox.com/u/15123461/Balmora_Reborn/TechnicalSheet01.jpg
Critique is really golden right now since I haven't really taken on this kind of projects before. What's most important is probably if the pipeline I've constructed is workable and if you have any opinions/advices regarding it and choice of game engine.
I've been looking at UDK and also have some experience in it, but I haven't really been able to figure out a good way to assign Ambient Occlusion to meshes/objects. I'm not really sure if that's the engine to choose for this type of project.
I'm going to keep working on some other prop while I wait for input.
Replies
If someone has actually done a pipeline like this in a real production setting, I'd love to see some documentation and screenshots. [edit]- a current game that actually looks nice
But if you decide to keep going that route, the only way i can think of would be a pain in the ass.
it would require you to make a seperate material for each texture that you have tiling on your mesh. In each of those materials you would multiply your ao map you created in the second uv channel over you diffuse. So in the end instead of having just one material with a custom texture, you'll need 3 *using your sign for example* and a custom ao map. headache and a organization nightmare.
As I stated, this is my idea and I don't really see why it can't be executed nicely. It might of course be heavy performance wise but that's a different matter altogether and if that's the case I'd retreat. You will however get a lot more detail going with tileable assets and of course some blending with other tileable maps and of course polypaint.
Well I figured that there might be a engine that could handle AO assigned to a mesh/object in the same way UDK handles lightmaps, one map per asset.
So in my example:
3* tileable Materials (seperate materials)
1* AO map assigned to asset in the same manner as lightmap usually is.
And I have to state that I love your feedback and you are most definitely correct regarding that this is indeed not the most common way of working, I'm fully aware of that. And I can see how it's seems like a pain in the ass from your perspective.
Edit: I know that GSC Game World uses tileable maps to a wide extent.(S.T.A.L.K.E.R. series).
From my experience I would say the main thing that matters is defining a clear silhouette - if you can do that in 10 poly's then do it in 10..
for something like this I would say you don't really need many more than 20 poly's for LOD 3, and go upward from there..
Of course I'm not much more than a newbie myself so I'm sure someone else can offer a much more technical opinion.
I see you mention tileable maps and "this prop has 3 different materials applied",
so if you're saying it's got 3 tilable materials applied to it that means 3 draw calls - a big no-no for such a small prop, I'd cut you some slack if that was for a whole building.
A small prop like this should be unwrapped uniquely with one texture, depending on how it's used it would probably use the same texture page as a bunch of other tavern/shop props.
update: all is not lost, you can build a low poly, use your current sign as the highpoly and bake the diffuse to your low poly.
The lowest LOD shouldn't exceed 100 polies.
Seems more reasonable to scrap the whole bolt idea altogether.
I've seen it in games I've played and though it was cool to have it as geo but I could probably make it look just as good with out it.
I had another version using the more of a standard pipeline before I created the one I showed you:
This is good stuff, I'll remember that from now on! :thumbup:
I made a quick LOD3 and it ended up at 40 tris.
I must be pretty off since I have 416tri for the chains only..
Got it! Will follow your advice :thumbup:
Do you guys think that a LOD 1 at 894 tris is acceptable or should I continue to butcher the mesh? I know that it might be hard to say a certain number but it would really make it much easier for me since I don't really want to be laughed at again
Edit: Added replies that I added before going to bed.
Haha, sure. I'm never going to be able to finish Balmora on my own in all seriousness. Once I'm done with creating a few props and feel a bit more settled with creating game assets I'd love to collaborate on Balmora
TES rules!
Keep it up.
Haha, can't wait to see more.
*Scampers off to take screenshots from another major city in Morrowind, so he can revamp it*
from http://www.polycount.com/forum/showthread.php?t=74502
I think 800-900 is okay for the detail model, but you could probably bring it down to 500 without much problems, got a wireframe to show? Preferably with the LOD's as well?
As for numbers, budgeting is a huge guessing game anyway, it's not as important to stick to them 100%. If a shape turns out to be pretty complex, you can spend some extra on them. Just make sure that you spend your polygons where they are most effective. By that I mean chamfering the bolts would be ridiculous, spend those polies on rounding out the round stick instead or just save them for another object.
Oh also. We're not just going to laugh at you for not knowing something. We're just trying to help you improve and work as efficiently as possible. We might be blunt sometimes, but we do that because we love you, honestly.
Wish you best of luck and fun on it, man.
You should mod it into Skyrim in November.
I loved Morrowind. (never liked Oblivion, though)
I'd suggest blocking out the map first. You could also bring in some of your own ideas and go slightly off the path of the original Balmora. For example, adding objects that aren't in the original town, because of the limited graphics back in the time of Morrowind. Hiring companies like seeing a little bit of your own creativity – properly dosed ofcourse.
Did you make a blockout of the town yet? That could be very useful to you before you start with details. It helps keeping the style coherent.
Thanks for all the help and support!
It would be awesome if you guys remade other cities!
We could have all done in time till Skyrim and just mod it in!
I'm still working away at the signs... Taking it slow and trying to get it right I guess. Still wip as you probably can tell.
That's awesome! I Would love to try that on my lower LODs.
Do I need to save a unique alpha map or could I just save my tga in 32bits?
I removed one of the corners of the chain and welded so I got it down to about 300, still not optimal but having it as geometry was important for me haha :poly136:
Would probably make it differently with hindsight.
I'm still working on the lower lods but the lod3 is finished for the first sign. How much would be okay for lod2? Would 50% of lod1 be okay or is there any rules/guidelines? Feels like I should scrap all the depth for lod2 and just cut out all the other geometry that I can and use a alpha for the chains.
Yeah I have a rough blockout, I also have some plans for some new props/ideas to add for the design.
I'll take some screengrabs of it and show you guys later to day(friday).
I've downloaded UDK now and was wondering if there's some good way to create the terrain? I have a heightmap but I also have the terrain as a mesh.
I'm currently reading as much info I can about it before I jump in but if someone has a link or hands-on tip I'd be more than happy!
The most awesome would be if I could use the heightmap in UDK and decide on polycount in the engine, but that's probably more of a silly dream...(?)
And then maybe we can have a sexy sexy port into skyrim.
Otherwise, I like where this is headed. But at the rate you're going with this, it may just take forever for you to finish Balmora altogether. It's not a small town! I would block out the scene first, and then finish with these panels. Also, the level of detail you put into these panels, is... well... a little extreme. Sure, it looks great! But also very time consuming for such a little thing that players only occasionally glance at.
Once I've normal mapped the materials I'll make sure that the wood pops the paint!
I know that it's going slow but I haven't done a project like this and I really want to get all the fundamentals down before I start working fast and in big scale. I also have a lot more stuff than I'm showing off right now. But once I get more settled I hope that the work speed will increase.
I'll show the blockout as soon as I can!
I've also made a rough blockout of the whole city:
I've had some serious hardware failure so the height is probably off by a lot. Will have to install Morrowind or try to recover the hdd. I know that some of the house shapes are off as well.
If you find something odd or that doesn't match when it comes to scale/else I'd be happy to hear it.
I haven't added any stairs yet, will probably start blocking them out now.
It can be playable in this case)
But once Skyrim is released I might try to bring it in to their new engine depending on what software support we'll get.
Haha yeah, the text is just a placeholder while I wait for my wacom pen to magically appear! The Mage's will just have a clean symbol as a sign as well.
On the flipside: I'm installing the newest udk and will start to bring in the blockout. Will grab some screens depending on how it goes! :thumbup:
Since part of your goal is to use this to specifically get a job at Bethesda, I'd use Oblivion or even Fallout 3 to create a playable version. That's up to you however. Who knows how much of the Gamebyro/GECK Creation Kit is carried over to Skyrim's Creation Engine.
The part about getting a job at Bethesda was more of a joke :poly136:.. Internship will do fine.. Nah, but I'll create it in UDK for now and if I'll get it done I might start thinking of bringing it into the Creation Engine. I mostly do it in UDK because I have some experience of the engine. I also think that using a up to date engine with lots of updated tutorials will help make the art justice.
Mostly just to get a feeling of the scale and how it will look like ingame.
Also have some things to fix, I think I'll start swinging from now on and just make everything fit.
The changes will be so minimal that most people won't even notice. The key is that everything fits and resembles Balmora on a concept level. I'll start adding some more geometry to the blockout and start thinking about how to divide the things that will be tileable/instanced.
http://dl.dropbox.com/u/15123461/Balmora_Reborn/_Screenshots01/needfixingwip01.jpg
Or else....
I used to think that the thing would kill me if I got too close! Long live the first days playing Morrowind!
"Speak quickly Outlander, I haven't much time."
"N'Wah!"
Oh how I missed the good old days. I remember that town gave my GeForce 2 GTS are hard time.
I'd love to have a go at the silth strider! I'm kind of weak at organic modelling/sculpting so it's on my to-do list:poly142: Haha, I remember walking into a random house in Balmora and just endlessly beat the crap of a NPC until I'd level up enough on my hand-to-hand, then I'd just walk out the door and pay the clerk at thieves guild and it was like it never happened!
It was like an hour of constant "ouch" and "ough". I still consider the woodelf males naked line to be the best of all time.. all time!
Also constantly jumping to level up the acrobatics. :poly136:
So much memories!