Chernobyl Window, searching for things to improve

hghtch

I'm pretty satisfied, but it would be nice if some trained eye could show me areas where I could improve it. It is my first real painter project and the first attempt at giving an object a history through texturing.

model

teodar23

The peeling paint is pretty random and it doesnt have continuity.
For me that is one big selling point in games and CG, the wear must take into account the object's shape. It cannot just be a tileable texture and call it a day. It has to be wherever it makes sense to be. In this example i suggest looking at some reference photos if you havent already. See where the paint chips and where it stays, where it cracks and where its just rubbed off.

hghtch

Well, I had not too exhaustive reference. Nonetheless I tried to follow a logic in placing the different textures.

I tried to arrange the wear in a way that would logically follow the water flow, with a usual weather direction from south east (relative to the window). The water runs down the vertical bar of the outer frame, leaving it mossy and nearly completely peeled. from there and other parts the water flows down and washes away the paint on the bottom bar of the frame. On the windowsill, the water accumulates and creates the damages on the inside, resulting in the voronoi patterned cracks.

Generally I tried to leave the parts more intact that are less exposed to rain, and let it peel from the edges where parts are exposed to rain.

Logically, the mixture of rubbed off paint and peeling doesn't make much sense. But I modeled the flakes and cracks after reference and figured in the process that it's not giving me the look I want by sticking fully to reference. In reality, the windows are much less rundown and only a few parts are very uniformly chipped away. It just doesn't read well as "run down apokalypse building". And with more removed, it didn't look right anymore. It looked like Tetris (which it still does on some edges. I guess less tetris is a paradigm I should follow), just digital or too plain. So I used the grungy rubbed off parts as a counterweight to the digital chipping to achieve a optical balance between organic and digital shapes.

What I learned from the reference is that there is either paint, or no paint. there is no "dithering" of flakes between paint and no paint, and the chipping edges are relatively continuous. I see this as the biggest point to improve on as this is IMO creating the impression of randomness that seems kinda off. The chips have something of this MARPAT camouflage look to me that breaks the coherence. There usually are no single flakes in the middle of a crack-island. The chipping always begins from edges where water has easy play to enter between the paint and wood. Hard edges are more prone than beveled edges, though the bevels are weakspots for cracks and will eventually result in the same.

The cracks are mostly found around the chipping edges, but there are cases where they didn't propagate into neighboring areas and all initial flakes are gone, resulting in crack-free edges. Other than that they accumulate in parts that soak in water and in the surrounding areas of these water pockets. I mostly followed the logical places where these pockets would acumulate (parts exposed to rain or collecting fluid) but also used them on spots you wouldn't directly identify to break the pattern a little.

So all in all, while randomness is concern, it is all but just a tilable texture (beside the parts that will be occluded by the walls. I didn't bother doing details that won't be seen later on.