Posts by jameswilddev2

To recap following my absence: I did about half a day's work, lost it, then didn't feel like doing anything at all for the following week. But I return...

That might be normal... I recall a lot of skies from EF being like that at least; the environment mapping stuff in the shader isn't supported by the editor, and no fallback texture has been provided.

Quote from PyroGXPilot

http://www.surfacegroup.org/tutorials/bump_mapping/index.html

That's very neat, I had no idea it existed, thanks!

This looks normal mapped, or maybe custom textures to fit the lights. How'd you do it? Looks good.

Working on the high poly of the "underpass" which runs under the diagonal corridor. In the background the eastern wing is visible.

[Blocked Image: http://i.imgur.com/y3nAlwi.png]

Also started thinking about how to build the background. Quake 3's skybox system is quite easy to produce textures for with Blender's camera system, but given some parts of the backdrop come quite close to the player and the sky cube has no depth, I'm going to have to include some parts of the backdrop as models.

[Blocked Image: http://i.imgur.com/c5OLEtf.jpg]

Began experimenting with sculpting the cliff the base sits on. I have only limited experience with sculpting and that is getting on for two years ago. I think the material is what is holding this back more than anything though. Will have to see how it bakes tomorrow.
[Blocked Image: http://i.imgur.com/5c8l5QD.jpg]

Quote from BJA

That looks like a great start! I like what you did with the layout and the curves. Though it's probably best to do this in the editor with simple patches.

Quote from Hipshot

I've seen others do that here for Q3, I would suggest not doing that. It might come back and bite you big time later on...

As someone working on a map built entirely from models, what kind of problems should I be looking out for? I imagine stock Quake 3 might not be very well optimised for this kind of map as meshes were used more for details and not for structural stuff; given the mesh gets baked into the map precomputed visibility might be done but I doubt it. Lightmapping might be tricky I guess; I recall having to make separate UV sets in UE3, but I'm showing my meshes fullbright.

Is that better? It was showing up before as a thumbnail here; not sure if that was a browser thing. They were showing up as links on my phone.

GTKRadiant has a 4-viewport mode; go to Edit->Preferences->Interface->Layout and it's one of the four layouts selectable along the top.

That's a good question actually - my entire map is being done as big static room-sized mesh chunks. At present, I'm exporting them in such a way that there's a big pile of meshes around 0, 0, 0, but the content of those meshes is usually about 2000 units away. If I understand from what I've been reading, misc_model is compiled into the map itself. I suspect, but cannot yet confirm, that this means the transform is baked into the mesh and the pivot point is lost completely.

I'm liking the planks; most Quake maps are very regular and lean heavily on patterns in their geometry. Is that BSP? How is it to run on?

Also, is that some bloom? Is that something stock Q3 can do? I recall Jedi Knight having bloom but I thought that was a custom extension.

[Blocked Image: http://i.imgur.com/kvEscXF.jpg]

You were 100% correct on it being a bit big Sprony. I like the height for how it plays but will bring some of the details closer to the floor (I'm partway up a jump here)

[Blocked Image: http://i.imgur.com/n9iWTZJ.jpg]

Baked to low-poly geometry. This is a relatively quick test bake with only 25 iterations; final render will be 150 at double the resolution.

[Blocked Image: http://i.imgur.com/CIBseHy.jpg]

It doesn't look like much progress but it's the fourth rebuild of the room. Beginning to learn the performance characteristics and limitations of Blender's Cycles engine. I'll be showing a test bake soon.

Hmm. It's not been going well. I've been at work this week so I haven't had a lot of time to put into this, but it seems the approach I was taking does not scale well; with only one room done, a bake takes a few hundred megabytes short of all of my RAM, and far too long to bake to be practical. I might need to make stylistic adjustments...

Quote from SpronyvanJohnson

Yes, nearing the last couple of weeks I'll rent a server so you all can playtest your maps.

Awesome, I look forward to playing some of that!

Began working on the high polygon models for one of the rooms; plan to have a test bake done by the end of tomorrow.

[Blocked Image: http://i.imgur.com/4T0g8na.jpg]

I'm leaning towards styling it with blue/green/sickly yellow dark interiors and fairly bright orange/red sunset outdoors for all that teal-orange goodness.

Looks very open plan! Like.

It is pretty big; I think I'm still in Unreal unit land. Have been cutting it down bit by bit, but there's going to be quite a bit of geometry on the walls so I'd rather it be a bit big rather than cramped right now. Thanks for your input.

Thanks, that helps. I'm only using the Quake 3 lighting for the greybox; my longer term plan is to high poly the whole map in Blender, raytrace it there and bake the whole thing to fullbright meshes. Though I will still need Quake 3 lights in the scene to illuminate players.

http://imgur.com/a/jffUL
Continued working on the greyboxing this morning. Got the whole map there, now to do refinements and see if any big changes need making.

Although the map is fairly big, it's pretty open making it very chaotic with the full 8 players. Finding 8 places to put spawn points where they didn't have line of sight with one another was difficult. It works a lot better with about 4 players at the moment.

Need to figure out how exactly to structure item pickups - I've got a couple of places about the map where the "big" guns are going to go that are fairly exposed, but I want to avoid putting any disk one nukes anywhere near spawns. For instance, I put some armour on the little shelf around the back of the ramp down from the diagonal corridor to hint that it's a path you can take, but that puts it right next to a spawn point and whichever bot spawns there tends to win.