Yesterday I increased the faction decal count to 16 and it works ok. 32 is probably possible but there’s only a few factions I couldn’t fit in to a list of 16. maybe I’ll leave those for a further expansion.
I tried using the blue value to control the hue of the camo and it works quite well, but in fact in game the main 16 camo types look fine. The hundreds of hue shifted variations just look strange and gaudy.
I’ll be using the third object color variable to control mist color then. I’m not sure why object color alpha cant be picked up by the nodes. That would give one more channel to work with.
Wow this is pretty cool! It looks very grungy and feels like a really old game, very well done
Is it going to be a strategy game were you compose different army setups and then go first person once the battle begins?
I am interested as well. Some quick experiments show me that I can do it using vertex color, but I can’t find a way to get object color into the node editor.
There’s an example file on page 2, the solution I found is perhaps not the most efficient way, but it seems to be the only viable way of working with rigged meshes which cant accept vertex colors.
To start with it needs a material node set to shadeless, with object color checked and no textures. That provides the object color, then another material node set to fully shaded without object color checked does the final shading.
A texture with an alpha channel lets the base color show through, either fully or blended. You need an extra texture if you want the final material to have transparency.
In my case the object color is used only to get uv scroll offsets, it doesn’t show in the final material.
This is inefficient because it requires an extra render pass for the object color material, but it’s no worse than having two materials on the same object.
It does look pretty good, I’m still thinking of whether to use fully simulated animation or just keyframed (normal) animation. The key is cpu workload.
It’s no good having a beautifully simulated animations if it runs at 5 fps. ^^
@ BPR
Have you stress tested it?
I just ran a test with 20 agents (exactly the same kind as I showed in the video) and it never went above 2.0ms of logic.
That’s going to go higher as I add more to the game (especially with mechs), but I’m still hopeful of having up to 16 agents in combat at once with no noticeable drain on resources.
The other factor is that all my mechs are already made, and they are all battletech cannon (see herefor some of the tabletop minis).
They are all rigged and ready to be added, so going back through and changing the logic and rebuilding them all would take too much extra work. Don’t worry, if I make a new game with an original IP, I’ll take a look at your design…
I just read the old Gray Death Legion books again, and there’s a section in the second book where Grayson is running around in the jungle with a bunch of converted industrial mechs as part of a guerrilla army… It always makes me want to make some mad max style conversions. But for now I have to concentrate on getting the base game up and running, then I can start adding some of the extra content I’ve already got hanging around, like the Clan mechs I made but never got around to adding.
For infantrymen crowds you can use a bit modified RVO2 + AI for targeting. This may allow to have ~100 infantrymen at once each shooting at it’s opponent side.
That kind of crowd simulation is actually very easy to do simply by summing vectors. Get a vector to where you want to go, get vectors from nearby agents you want to avoid, sum them together and reduce the length to the desired speed.
Some more work this week, starting o the particle system, also tried using a nice resource from the forums to add water (just because of BGMC ), but found the result to be too expensive for the quality of the result.
But for now I have to concentrate on getting the base game up and running, then I can start adding some of the extra content I’ve already got hanging around, like the Clan mechs I made but never got around to adding.
), but found the result to be too expensive for the quality of the result.