Hey guys from RAD, just wondering if you can answer a few questions about the material/texture pipeline you used. We're looking to test this idea of tiling swatches using masks like you did for potential use at our studio, so I was hoping you could answer some unknowns I couldn't find online.
1. Were the tiling textures you made for the ground created from multiple swatches with masks, or was it just the props that used the masking system and swatch library. For example, was a cobble stone ground texture just authored in Photoshop, or was that ground actually built from masks and multiple material balls plugged in.
2. I see that the masking and blending of materials could be previewed in Mari or Maya, was there any recompile time when plugging in a new tiling swatch, was their any bake time for materials or was it all real time if you decided to plug in copper instead of steel for example. I'm wondering if there was a long build time if you decided to make edits to the base copper after it had been referenced into hundreds of props for example.
3. Where did you find the base values for something like steel, I've checked online and there's not a lot of reference out there for a physically based chart, did you have to capture all this information yourself with a scanner?
4. Do the material swatches run live, or do you bake them into the UV space as a unique texture after you have masked all the layers together. For example the material balls/swatches were plugged in and could be tiled as much as you want I would assume, but did you then have to bake that out as a new diffuse to run the game at target fps, or if you zoom in would you actually see the detail is much higher rez than the 0-1 unwrapped mask texture for example.
5. We are finding the physically based chart goes out the window as soon as something is not lighting correctly, we've adopted the motto of start with physically based, then adjust in game if it just looks bad. Did you find everything just worked if it was the real world value, or did you find you have to edit a lot of stuff down the road to make it look good rather than being the pbr value from a chart?
6. This system makes a lot of sense for a metal object with paint on it and some rust, but an old brick wall with lots of grime layers seems really painful when I could just drop a grime photo from textures.com on top of my clean texture and change the blend mode to overlay, how did you tackle complex surfaces with lots of subtle variation in hue and value. Did you really create 8+ masks rather than using a blend mode and photograph.
7. For your previewing scene did you use just a single hdr image hooked up to an ibl, or did you have day, night, overcast and allowed the artists to switch between them. Also how did you choose the 3 hdr images, and are those images available online or did you shoot them yourself.
Thanks very much in advance if you answer these questions, the game looks amazing.
I'm puzzled by steal beams, are all welds made as floating geometry decals as on some places they do not match exactly underlying geometry. I'm trying to do the same as I need to tackle some steel structures (plenty of them) so good workflow for this would be nice! Anyway this artdump is of epic proportions. Well done, especially environment modeling. That I personally dig the most.
Replies
1. Were the tiling textures you made for the ground created from multiple swatches with masks, or was it just the props that used the masking system and swatch library. For example, was a cobble stone ground texture just authored in Photoshop, or was that ground actually built from masks and multiple material balls plugged in.
2. I see that the masking and blending of materials could be previewed in Mari or Maya, was there any recompile time when plugging in a new tiling swatch, was their any bake time for materials or was it all real time if you decided to plug in copper instead of steel for example. I'm wondering if there was a long build time if you decided to make edits to the base copper after it had been referenced into hundreds of props for example.
3. Where did you find the base values for something like steel, I've checked online and there's not a lot of reference out there for a physically based chart, did you have to capture all this information yourself with a scanner?
4. Do the material swatches run live, or do you bake them into the UV space as a unique texture after you have masked all the layers together. For example the material balls/swatches were plugged in and could be tiled as much as you want I would assume, but did you then have to bake that out as a new diffuse to run the game at target fps, or if you zoom in would you actually see the detail is much higher rez than the 0-1 unwrapped mask texture for example.
5. We are finding the physically based chart goes out the window as soon as something is not lighting correctly, we've adopted the motto of start with physically based, then adjust in game if it just looks bad. Did you find everything just worked if it was the real world value, or did you find you have to edit a lot of stuff down the road to make it look good rather than being the pbr value from a chart?
6. This system makes a lot of sense for a metal object with paint on it and some rust, but an old brick wall with lots of grime layers seems really painful when I could just drop a grime photo from textures.com on top of my clean texture and change the blend mode to overlay, how did you tackle complex surfaces with lots of subtle variation in hue and value. Did you really create 8+ masks rather than using a blend mode and photograph.
7. For your previewing scene did you use just a single hdr image hooked up to an ibl, or did you have day, night, overcast and allowed the artists to switch between them. Also how did you choose the 3 hdr images, and are those images available online or did you shoot them yourself.
Thanks very much in advance if you answer these questions, the game looks amazing.