Doom3world

these are the first images I`ve taken of this wip project. there of some floating (gravity defying) rocks I modeled in blender which the d3 player will traverse. I`ll ul some in game images tomorrow, if all goes well.

also uses one of my own rock textures, but in game I`m still using some local and height maps which came with d3, which I`ll replace with my own when I get round to modeling the local map in blender and painting the hieght map in psp7.

edit: updated links to use imageshack thumbs.

It does look good!

You should extract those rooms seperated, because you need visportals if you make large map!

Hurr!

thanks Zamideus, although you might be misjudging the scale, these are just rocks the player stands on and jumps between. I haven`t started yet modeling the lava pit room they are will go into but I intend to start on that today.

kit89 asked if I had some wire frame versions of the images so I`ll show you guys as well. I`ve tried to keep the polys as low as possible, and this is the first time I`ve modeled rocks using quads rather then tris, which kat had recommended I use quads where ever possible, and I can see why now, although I had to convert some of them to tris to control exactly where goofos .ase exporter would make the splits.

oh and I registered an account with photobucket.com yesterday but then their site went into some maintenance time, so I couldn`t link to the pics I was going to ul on there, in the mean time we`ll have to do without thumb images.

edit: updated links to use imageshack thumbs.

it appears photobucket.com is online again, but being a noob with using their service, let alone html code, how do I post a thumb image in here which links to the larger image?

dont know about photobucket but image shack gives you 5 of 6 different links for the different purposes-direct, thumb, large version etc

thanks voldemort and I might set up an account with them instead if I can`t find out how to do that with photobucket.

Watch out for the height of those 'pathway' blocks from the ground. You'll need to drop them or raise the ground level around the edges so that the flat faces of the boxes are NO MORE THAN 8 units from any given surface otherwise you hit a wall and travel along the edge instead of up and over it (as you would normally do) becasue of the way the collisions system works.

Don't be so stingy with the polycount either! Round off the sharp edges otherwise they'll look naff in game (that'll add a bit to the pCount if you do it on all the sharp/hard edges).

thanks kat, and the height of the "stepping stones" as I like to think of them, seems fine in game, as there is always a side of them which is only a few map units above the rock surface, and I`ve done lots of tests with the player walking, running and jumping on them, as well as comparing them to the ones from the original d3 hell maps, but I agree something we need to keep an eye on when placing these.

encouraging to hear you think I can increase the polycount on these rocks, which I would love to do if I can get away with it. I actually tried to keep about the same poly count as the originals, which I agree do look very angular in places.

saying all that, I think I`ll finish a basic layout of the first area or room that these are going into, then if the sp monsters and particles aren`t slowing the fps down too much I can add more polys

Ohhh if they're part of a larger scene then yeah you can still up the polycount a bit to get better smoothing in game but don't go crazy. I thought whta you've shown was a map in it's work right, hense saying you can increase the pCount, obviously don't do that if you've got other things going on.

np and I welcome any constructive input

I think the scale and basic concept will make more sense when I post my next images. atm, I`m working on the rock walls of the map which encase these floating rocks.

it took me almost as long setting up and tweaking the d3 lighting in this map as it did modeling in blender. at least now I should be able to do the lighting quicker for the rest of the map, this being just the first area, and still a wip so a several more models to be done for this area, as well as particle emitters for fire and smoke.

now using imageshack.us so we have thumb images

the two yellow lights which illuminate the glyphs on the small floating rocks, either side of the main floating rock, haven`t yet been bound to the func_bobbing map objects, which make those two rocks bob up and down slightly.

does anyone know is there any slow down with having moving lights instead of static ones, even if the geometry they cast light on is moving?

That is looking really nice, I like that sort of stuff!

Like Kat said, up the polys on the edges of the rocks to eliminate hard edges, you want to make the silhouette nice and smooth to make it look like it's got more polys than it looks.

Don't be affraid if you seen hits 100-150k tris and about 50k shadow tris, even cards like the 6600GT can handle that quite easily on high settings.

thanks zombie13, and so far I haven`t noticed any slow down with the models I`ve been adding to this map, although I`ve only been testing it with com_showfps, but I will start taking note of the tris if someone can remind me what the cvar to display them is?

and maybe because I don`t have tons of lights overlapping, but what ever the reason it looks like I will be able to increase the polycount on the parts of the map the player walks on at least

interesting suggestion you made about the silhouette edges being given higher polycounts, if I understand you correctly? so combined with the bump textures should give a good impression the over all polycount is higher then it really is, yes?

am I correct in thinking the d3 engine is very efficient at drawing particles, like used in smoke? and is that because they usually are just alpha or additive blended and not have bumps or specular texture stages to slow them down?

r_showprimitives 1 combined with con_noprint 1 will show you whats going on in your scene.

Every time a light is casting on a surface that surface has to be redendered again, so imagine what something like 8 lights would do? try and keep it to about 2-4 lights (I'm pretty sure there is a section here on D3W about lightcounts etc).

The D3 engine really isn't that efficient at drawing particles unlike some engines where it doesn't really bother it, but in saying that, having 3-4k tris worth of particles isn't going to hinder your scene much.

Yup having higher poly silhouette edges does make for nicer rock/terrain what ever.

understood and thanks for your help zombie13

not sure where to ask these questions so I`ll post them in here.

kat, I just read your tutorial on texture space and orientation

1. although I think I understand your tut, am I right in thinking it wouldn`t matter with rock textures and their local maps, as they would probably be acceptable in game, even if their orientation on the models and some of the local map axis were not aligned to their original up/down, left/right?

2. also do you have a preference over using .lwo or .ase model formats, both for the in game low poly models and the high poly local map versions?

3. and do all of the static model formats produce a smoothing seam where the uv maps have a seam in blender?

1 : What that tutorial/tip was really refering to was orientation of heightmap 'templates' when used to generate normalmaps. If you render out a normal map from the 1st image by then applying that verticle 'trim' to a model along the horizontal you'll create a 'flipped' channel becasue of the way the Nmap was created.

Rotating and moving UVWs don't make any difference in terms of 'fixing' the problem becasue it's a problem at source, on the image itself.

But using a 'correct' normal map has no effect when UVWmapped as the game 'compensates' for it (for want of a better term)

2 : I've not yet fully tested LWO models other than to solve that smoothing issue mentioned in the 'prepping LWO models tutorial'. I've got a multimaterial terrain mesh section to try out but have yet to do so.

3 : You mean an actual UVWmap split? A map physically broken into two sections (for example). Yes whereever there is a break in a UVWmap you get a seam in game, you can't avoid that (for terrain you can't use that tangentspace material parameter you see in character models).

If you're talking about Blender UVW (mark) seams then no those have no effect whatsoever one the exported model; it's basically just an 'internal' helper for Blender.

thanks kat,

1. understood and that`s good to hear, so I assume in blender I could even have my uv maps upside down or rotated 90 degrees on the terrain models, as long as the local map used in game as been generated using correct axis, yes?

2. understood and I`ll go read that tut of yours again as it was months ago I last read it.

3. understood and sorry, I ment a physical split where two or more meshes with different materials in blender are combined and exported as one .ase or .lwo which then produces a shadow seam in game.

and I remember in the past trying out the unsmoothedtangents global keyword on my blender terrain materials, but never got it working, do you know why it only works on character models?

kat, I`m adding some low poly details to the large cave, and wondered whether to keep it all as one .ase model or export it as several model sections, similar to how you appear to of built your d3 and q4 terrain maps?

and is there any advantage if the engine can`t see the model even if it`s in the same vis portal area? I did a quick search on this but wasn`t sure I understood what the other guys were saying.

with my first area of this map, I`m only intending to use brushes for the sky box, but should I split them and caulk the parts of the hull that the player can`t see, ie brush faces that would be behind the models?