Hi, Maybe a stupid question, but how long does a somewhat experienced 3D artist take for creating a map section? I guess it quite depends on the type of map and the detail level. Let's say, 'pretty' detailed. But for game purposes, not a CG movie or someting. 1- a street with some buildings, lampposts, trees, bench, signs (a GTA street) 2- an outdoor section. 1 acre of field/forest/hill (a Zelda field) 3- Indoor: a house (furniture, rooms, etc.) 4- Indoor: a cathedral (complex architecture) 5- Indoor: factory/high-tec/prison/hospital section 6- ...put in another way, how many 'scenes' did you create last year? How many hours/days/weeks would you need to create these 3D meshes in your favourite editor (Max, LightWave, Maya...) and then 'fill' them in another editor with textures, decals, objects, entities, rendering settings,...? Let's say the textures and common objects (tables, computer, bins, barrels) are already done. The reason I ask is because I'd like to make rough estamation. I'm creating a game concept and map, and I hope that I can attract artists some day (not anywhere soon, first finish engine/editors/game-concept !). But I'm afraid that my map quickly gets too large and too ambitious. I don't mind working a few years on a project, but if it turns out that I need 80 mappers and/or 20 years to finish it... A concept is just a concept, but nevertheless, I like to keep it somewhat realistic :) Greetings, Rick

Hey, thanks for that tip! I'd never think about that. 2/3 months... A little bit longer than I expected, although these maps are also larger than I had in mind. I guess the actual required time could be shorter, assuming that most of the creators are hobbyists that did it in their spare time. Some of them might spend hours and hours, others might be busy playing with their girlfriends and only do a few hours of modelling per week. I'm not really familiar with creating shooter maps, but I also guess that ~50% of the time is actually used to play & test the map, which is crucial for the multiplayer/shooter type of game.

Rick

The thing about art is that there isn't any set rules on how long something takes to make.

A rough estimation is putting it lightly. It's really really really really really really really rough.

What might take one guy 3 months to make, might take another guy 3 weeks... or another guy 3 years. You're going to get a wide array of answers from a wide array of artists, and of course the amount of time to make something changes depending on how you want it.

A hyper photo-realistic scene is generally going to take a lot more time to make than a simple cartoony environment. Plus the quality of anything is going to more than likely going to be determined by the amount of time spent on it.

For example, most of the animated shorts that Blur does for game trailers are like 3 minutes in length... yet they get 7-9 months to make them.... That gives them a lot of time to make things look really good. By comparison, some 3d cartoons have to get an entire season of episodes done in 7-9 months... which means the quality is going to diminish a lot.

I'm going to go ahead and just be honest with you and tell you that you're not going to find an ideal deadline for your project. Most games end up behind schedule and force everyone to work a whole lot of overtime just to get the thing finished. That's because you can only really give an extremely rough estimate of how long things will take.

But it is good to decide on a deadline, as this gives your team a goal. My advice is to just be sure to pick a good time frame that isn't too long (which means people won't be working as productively) or too short (which means it's just impossible to finish on time or that it will be really low quality)

It's a tough decision to be made but it's one that needs to be made. My advice in this situation is to guesstimate a reasonable amount of time and then add another month on it as a cushion.

But again, this is all just a matter of opinion. Art is very subjective. :P

Procedural generation works for outdoor games like Crysis, C&C or Farcry, but in my case it will play in a human-made world, most of it. But thanks for the tip!


I was already affraid it would be very hard to estamate. I'm not really looking for a formula to calculate we need X years to produce the whole thing, but I was worrying about fantasizing a huge world that is impossible to create with just a bunch of guys/girls. The thing is, I'm programming an engine, tools and a game-concept. When those are 'ready', I'd hope to collect a bunch of artists one day (not anywhere soon :) ). But as I can't offer a nice salary or whatsoever, I should be happy if I could get 1,2, maybe 3 talented people with some spare time. To attract talented people, I need to make sure the engine/tools and game-idea are tip-top, not "another hobby project #32421". So I'm really doing my best to work out all aspects before even trying to attract them. On the other hand, I shouldn't come up with Metal Gear Solid 5 or Halflife3 neither. That would be out of proportion, and experienced artists probably won't even start on it, as they know it will fail.


So... that's why I asked how long people take for making the stated map types on "normal/good" game-quality level (not cutscenes or cinematics). Trying to peel out the average time. Let's ask it in another way:
- How many (fulltime) mappers are working on a game such as Halflife2, Zelda or GTA IV?
- How many years does such a team need for these games?

Just trying to get a slight idea :)

Rick

hey,

ive dabbled a bit in 3d myself and dont get scared, big scenes arnt nessisarily harder/more time consuming than smaller ones! =D

also you have to take into accound how many instances can be reused. like trees, walls, lamps etc, so once its made u can just copy and paste it to the next location to save time ;) also it saves time on texturing as u only have to texture one of them and if your looking for extreem detail and being unique, you can then use Alpha maps with dirt and grunge on them to make the texture look different but not altering the overall base texture which also saves MB's of files because you dont have 30 or so of the same base texture slightly altered with different dirt maps etc..

also alot of the textures in game maps are just tiled over and over again, so theres usually not one big unique map for each building, there are only Alpha maps overlayed to make it unique.

so the map you describe shouldent take more than 2 weeks to fully model and a nother 2 or 3 weeks to texture (thats going by if hes a realitivly average to fast modeler. BUT not all 3D modelers do their own texturing! which can have its own problems espesh if the 3D mesh is a bit messy, it can prove hard to UV map cleanly. to be honest the texturing is always gonna take way longer than modeling it.

i hope i have somewhat helped you a little bit :)

take care,
Corey

2/3 weeks. That's a little bit more positive :) I'm trying to separate mapping,
modelling of reusable objects and texture-drawing into 3 different tasks. So
a mapper should be able to choose them from a library. That also means I need
a modeller and texture artist of course, but that's another story.

When it comes to making scenes that can reuse lots of stuff (Halflife) or very unique scenes (Resident Evil 1), I think I'm somewhere in between. When possible objects/textures should be reused. On the other hand the scenes should have their own little details. The game-pace is relative slow compared to shooters, so the player gets plenty of time to observe the area, and athmosphere is crucial (I'm aiming for a horror game). "Painting" the walls with dirt is a nice way to increase the unique feeling. Also small stuff like pipes, railings, cables and signs are important here. But I guess those little bastards are also making the development time longer :)



I'm splitting the mapping up into 2 stages. First the mapper creates the untextured mesh (without dynamic objects/entities) in Maya/LightWave or whatever kind of tool. I think that's still easier than making it in a game world editor such as Hammer, Quark or UnrealEd. Especially when it comes to complex shapes. Downside is that you have to re-import each time you changed something in the mesh.

When the mesh is ready, it can be imported into the editor. The editor will automatically create portals, and create default texture coordinates. The user can select faces and apply a texture, insert lightSources , place entities, modify rendering settings, and so on. As the editor is "what you see is what you get", you can directly see and test it in a "game modus". When the mesh has been changed and must be imported again, the editor tries to remain the texturing/entities as much as possible. Don't know if its the best way, but if I had to create my own modeller I need another 2 years coding. I'm sure the guys at LightWave/Maya/Max can do that alot better than me with years and years of experience :)

greetings,
Rick

Not that I don't believe you, but isn't procedural stuff used to create 'bulk' such as a forest? It's certainly possible to place roads, houses and even flats random on the map as well. But in my case the environment won't be huge and open such as GTA or Crysis. It's more like a puzzle, like in adventure games (Resident Evil, Alone in the Dark). You can't just walk everywhere, you have to 'unlock' area's. So I really have to think carefully about how to design the map, hard to lean on a random generator here :) Also the reason why I'm already sketching maps, and thinking "Nice and well, but will we ever be able to create all of this?"

Rick