|
Author |
Thread Statistics | Show CCP posts - 8 post(s) |
|
CCP LogicLoop
C C P C C P Alliance
1064
|
Posted - 2013.10.08 00:16:00 -
[1] - Quote
I know at some point the "intention" is to have fauna. When that will arrive, or how, I have no answer to. However, to speak on your idea, these are things that concept art, environment art, and level design would flesh out in the initial phase of new terrains or game-play areas if the feature should arrive. |
|
|
CCP LogicLoop
C C P C C P Alliance
1298
|
Posted - 2013.11.18 00:32:00 -
[2] - Quote
So we could put in trees in to the maps. However, this counts towards our mesh limit on terrain (which last I heard is about 200 meshes). Trees will not work in the current "plant" system you see now. They have to be placed as a static mesh. I did clear this info up finally with a Tech Artist last week.
So it is a matter of convincing Team Daddies (CCP Stiffneck) of sacrificing other meshs for some trees and to make them.
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1298
|
Posted - 2013.11.18 01:36:00 -
[3] - Quote
It would require a duplicate of the game mode it is changing, then that game mode would have to be the one that randomly loads. We can not dynamically swap them in a single game mode layer.
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1337
|
Posted - 2013.11.18 07:13:00 -
[4] - Quote
Mark Crusader wrote:CCP LogicLoop wrote:So we could put in trees in to the maps. However, this counts towards our mesh limit on terrain (which last I heard is about 200 meshes). Trees will not work in the current "plant" system you see now. They have to be placed as a static mesh. I did clear this info up finally with a Tech Artist last week.
So it is a matter of convincing Team Daddies (CCP Stiffneck) of sacrificing other meshs for some trees and to make them. That's 200 mesh object-types, or instances?
200 total static meshes on the terrain. We can go over that but not by much. That is our safe count.
One trick around this is "combining" a group of meshes around each other into a new single mesh. However, that doesn't alter the mesh material count. The different materials are still in use and that has a limit as well. It's all a really fine balancing act in the long run.
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1382
|
Posted - 2013.11.19 00:18:00 -
[5] - Quote
Aeon Amadi wrote:CCP LogicLoop wrote:Mark Crusader wrote:CCP LogicLoop wrote:So we could put in trees in to the maps. However, this counts towards our mesh limit on terrain (which last I heard is about 200 meshes). Trees will not work in the current "plant" system you see now. They have to be placed as a static mesh. I did clear this info up finally with a Tech Artist last week.
So it is a matter of convincing Team Daddies (CCP Stiffneck) of sacrificing other meshs for some trees and to make them. That's 200 mesh object-types, or instances? 200 total static meshes on the terrain. We can go over that but not by much. That is our safe count. One trick around this is "combining" a group of meshes around each other into a new single mesh. However, that doesn't alter the mesh material count. The different materials are still in use and that has a limit as well. It's all a really fine balancing act in the long run. When you say materials are you referring to a texture or..? Also, in the case of 'forests' couldn't you put together multiple meshes in the single mesh approach you've brought up and just use the same materials for each? I mean, would be hard to manage with the terrain elevation changing but you technically could put large tree trunks with all the actual floral bits at the top - much like how the Installations work, what with their elongated spikes.
Yes. Materials are textures.
So we are now delving a little further into territory I am not 100% up to speed on. Mainly because of differences between versions of the Unreal Engine. For example if I recall later versions of UDK support converting all the materials in a multi mesh merging into a single material. Prior to that particular version though, the materials for each mesh merged into a single mesh are still counted as an individual material for each one. So if we combined 15 meshes into one, we would end up with a single mesh using 15 instances / draws / calls of the material, even though it is the same material. So the mesh that Unreal has made actually puts in the mesh 15 separate material id's calling for the same material.
Also, if we were to just build say one mesh in max ready made with 100's of trees or even 25, and spread them out how we like, we still have to worry about the LOD. If we spread it across the whole map, then the trees would never use a low LOD. It would always be loaded at full LOD because the player is always close enough to the "single" mesh.
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1385
|
Posted - 2013.11.19 02:24:00 -
[6] - Quote
We would love to do them as well, believe me. But as you can see, we have to balance a lot of stuff. So many things have to go into consideration for every aspect of a game. It's a really interesting balancing act. This is why we have to set technical budget limits. If we just did what ever we wanted we would have all sorts of problems.
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1385
|
Posted - 2013.11.19 03:02:00 -
[7] - Quote
Aeon Amadi wrote:CCP LogicLoop wrote:We would love to do them as well, believe me. But as you can see, we have to balance a lot of stuff. So many things have to go into consideration for every aspect of a game. It's a really interesting balancing act. This is why we have to set technical budget limits. If we just did what ever we wanted we would have all sorts of problems. I'm sure it's already been considered but you could use sprites to decrease performance cost in some place. It'd be cheap as **** but there is the potential to replace long distance visual effects (smoke and the like) with them. World in Conflict used -ridiculous- amounts of sprites and it still was considered the most realistic looking smoke effects for an RTS in it's time. Albeit, this is a suggestion from someone who isn't in game design (yet) but I blame you for saying in the dev blog to throw out all ideas even if you think they're silly =P
Sprites are definitely a cheap thing to use (in terms of trying to keep in budgets).
You can catch more flies with honey than with vinegar.
|
|
|
CCP LogicLoop
C C P C C P Alliance
1397
|
Posted - 2013.11.20 00:17:00 -
[8] - Quote
Aeon Amadi wrote:Robert JD Niewiadomski wrote:Can you make trees like you do grass and shrubs now and put them in "an invisible box" to prevent us passing through them? Would it be any different (from the Unreal Engine on PS3 technology limitations perspective) than making 3D models of trees? With meshes, textures and all "what comes with that"? More times than not, grass and shrubs are usually just a few polygons interlaced together to make a grassy look from the side, but when you stand over them you see something like this: http://preview.turbosquid.com/Preview/2012/03/11__04_05_35/th2.jpg42328553-a3fa-4e76-94d6-4d5eadae8657Large.jpgDo it enough times and you get this: http://preview.turbosquid.com/Preview/2012/03/11__04_05_35/th1.jpg6f48f306-9335-4ed4-ad07-2e51fd8ee21dLarge.jpgA low poly tree wouldn't be too bad of a performance loss ( like this one, which has a polycount of 277) but when you try to make a small wooded area of say, three dozen trees, that polycount suddenly jumps to 9,972. That's almost as much as your typical FPS character (fun fact: Kiril in Battlefield 3 had 15,818). That's just the polycount as well, not including animations (assuming that they are swaying, or static) texture budgets, etc. When something as simple as Nanohive/Drop Uplink spam bogs down the system - adding a bunch of other stuff to the picture suddenly looks a lot less promising But, as previously stated, I'm an amateur when it comes to this stuff - so don't take my word for it!
Pretty good summary. Our concerns are the actor counts, polygons, and materials. The animations could be done a couple of ways either with real animations or with a material shader that simulates swaying (which is what you commonly see).
Indeed, it does all add up, and as stated, it is an interesting balancing act.
You can catch more flies with honey than with vinegar.
|
|
|
|
|