Thanks guys, I hope to start texturing in two weeks or so, this building will provide me with most of the rF2 specific things I will need to learn I reckon
In preparation to the texturing stage I'm already thinking about a possible uv layout. Now to make sure everything is optimized as much possible I have a question concerning the windows. Right now I modelled everything in, but it is quite high poly so I am probably going to bake that detail into a normal map if I can't get the triangle count down. Now, for reflections, is it better to: - seperate the textures for the windows from the actual building and use a different shader (so two sets of textures and two shaders) or - include the windows in the building texture and use the same textures in the two shaders, omitting the reflection mask in the building shader (this way only having a pitbuilding.dds and a pitbuilding_normal.dds where in the first I store a reflection mask in the alpha channel, and in the second the specular information.) I've already decided to use two different shaders, because the windows which need reflecting are maybe 15% of the entire building at most, so when I use the two shaders the renderengine only needs to check for reflection on the stuff that matters. The small medical facility kicked my ass a bit, hardly any proper reference, but I got it sorted. With a bit of luck, I can texture the entire big-ass pitbuilding complex with only a handful of textures
I use separate shader for windows. Especially that I use more complicated shader for buildings (diff+spec+normal+AO) and only T1+cube map for windows.
Okay good so you are doing roughly the same. The AO is going to be a bit of challenge since it such a big building and I don't want to have a huge ao map... I suspect some tiling within my AO for some bits, but that is going to take some creativity The pitboxes themselves will probably be seperate meshes just copied over so ideally, they will share all maps. They will be combined for LOD purposes, and will also have seperate meshes for the smaller details to be culled when farther than xx meters away. This needs testing as I don't know how rF2 handles this at all. Hopefully before the end of the weekend actual modelling on the big stuff will be done Details in the pitboxes will be done at a way later stage, as they don't matter one bit right now. edit: Since the entire pitbuilding complex takes up so much real estate, and are basically two buildings joined by a smaller one, I will split the textures up. This is what I am planning to do: - pitbuilding A diffuse texture - pitbuilding A normal texture - pitbuilding A AO texture - pitbuilding B diffuse texture - pitbuilding B normal texture - pitbuilding B AO texture - windows diffuse texture - windows normal textures - tiling floor tiles diffuse texture* - tiling floor tiles normal texture* - tiling garage door texture - tiling stairs texture* - garage details atlas diffuse texture - garage details atlas normal texture * wip These are alot of textures unfortunately, but this combination of assets almost runs down the entire length of the pitlane so I guess that is okay. Some of these will be built in such a way that they can be used throughout the entire track. For example, the floor tiles and stairs, they will be used all over the place and for example the stairs will be modelled seperately and are not an actual part of the building mesh. This way I can cull them or build seperate LOD's for those Really pshyced about making the deadline I have set myself. At the end of this weekend, I want to have concluded all modelling work on the building(s). After that I will start optimizing the living crap out of this thing and start unwrapping it piece by piece and going through the first texturing pass. I hope to be able to share an update of this within a months' time. I will probably do an export test with solid colors first to see if everything imports without problems Something else, about the track. Earlier, I was talking about the track being distorted by google, and less so by bing. Yesterday I found out that which ever of the two I pick, they are both off, as I directly compared a Lidar scan from a certain track with it's google and bing reference. The track was only 85-90% accurate, due to slight skewing or scaling (in width) of the satalite images. Because of this, when the pitbuilding has had it's first pass, I will revisit the track. Doing so, I'll check how my track runs with the two reference images in there, and I'll just stick to the one which looks most likely to be the truest to life. And no, simracers won't have a clue if a corner is a meter off in radius or placement, so I will not dwell on this for too long, as I want to get this project going. Even better, after I check and fix the track surface where needed, and fix up some rumble kerbs, I will add in the AIW stuff needed to have basic functionality** in there, and will move it up to alpha status. In my case that will mean that all functionality is in there, so it can be packaged and played in single player. With that step, I hope to receive more meaningful feedback to learn what the rF2 community wants from a track. On to Zolder 0.1 ! ** 24 pit positions 12 garage positions 24 start locations left/right/fast path
Oh! I see there is an update Is this a non-dev mode or dev mode so that I try it tonight please? I'm really looking forward to this one mate as it's a great track! Thanks for the good work!
Yep there is, pretty soon a new update where I fixed the biggest issues between google and bing*, that might be the last Devmode one, being pre-alpha. V0.02 is still for devmode, practice only, no AI. Thanks, have fun, I can assure you already that when you drive the new corvette on here soon, it is a pretty sweet drive I'll probably post an update on the pitbuildings this weekend, first texturing and ambient occlusion of the entire thing. Bit of a hassle getting this optimized since it is so big, but I have found a way or two * again, that is just me nitpicking, not many will notice these changes straight away, but I will
For small details like pipes, cables, speakers, and so on you can bake AO into vertex. As for a main building, that will be hard to use one AO map for such a big building, you will probably end up with few materials, but you can find indentical parts and you can use the same AO map for those parts.
Yeah that could be an option when I get down to it, I'll probably have to cut in some extra edges to counter stretching across large polygons, but that will be something for the future. The smaller details will be a seperate gmt which will be added at a later time. Using a seperate gmt I can cull those at a certain distance, same with the pitbox interiors which have LOD objects, one with, and one without objects I'm still working on the pitbuilding, trying to use as little texture space as possible, is a real sport with this kind of building right have all the walls for the main pitbuilding in 1024x2048 map and they are still sharp Maybe an updated screenshot this weekend, who knows edit: feels, if you want a packaged version for testing, just ask. edit2: Finding identical parts is possible only for 50% of the building sadly, so all in all it will be a challenge to get the most detail out of it. I am aiming for a maximum resolution of 1k for the ao map, and that is already quite big.
Small update, don't have a new ingame shot of the building, as I have yet to complete the second half of the complex. That will be the goal this week, to have that done by the end of the weekend, and have the first textured version in the actual track.
I finally got an update ingame, the cubemap from Joesville gave me trouble.. Have a question about some decals I placed, they really have awful edges, is there any way to make these look better? The shader I am using is T1 alpha test, maybe there is a different one for decals? I don't really like the texture resolution I'm getting for the walls of the building, but this is optimized into an atlas texture, so I will keep it that way for now
I have been working on fixing the corners I talked about earlier. To be honest, seeing the differences again between reference, bing being much better image quality, in Darin Gangi's infamous words, I'm too frustrated to continue. What does this mean? I will completely redo the entire track surface, grass and kerbs. This will take more than I would like but currently it doesn't make me happy, especially for a portfolio piece. So, I have two options now. I have a packaged update for Single player, and a Devmode version with the latest fixes and an optimized / textured pit building. I'm only going to upload one if any so choose wisely. edit: To clarify, when redoing this, everything will be textured as I go along. This will speed up progress a bit, but I'll keep the project in alpha status till everything is in place again. edit2: changes already feel way better
Red=old Black=new Especially the first three corners flow much better. Yes, I delayed my work now, but I'm happy with the result. Everything I will add next, like kerbs, those will be immediately textured and finalized. I get the impression the devmode builds aren't very appreciated, so the next build I will release will be a regular singleplayer component, with everything textured between the guardrails, and the pitbuilding with a first texture pass.
Looking forward to this mate. Take as long as you need. Zolder is such a great circuit, I'm really pleased you are bringing it to rF2 and obviously taking so much care over this project.
Same thing here. I prefer rather to wait a little bit and get the chance to try something real good than something where even the creator is not really satisfied with..
Thanks guys ^^ I'm already working on the inner grass sections between road and guardrail. This is the second time I'm doing it and as usually the case that means it only takes half the time I inspected Limerock park yesterday and I really like how they did the edge of the road there, I could hardly spot any texture seams there. Overall a very nice track, even off-track as well. For now, I'm going methodic on this one, finish kerbs, finish rumblestrips, finish rumbleblocks, finish grasscrete, finish road*. The road comes last, cause when touching the kerbs and sides of the grass, you never know what might be found Kerbs, rumble and grasscrete will all have realroad shaders applied to them as well. Stay tuned