Experiments with Material scans
After seeing demos of the new nodes for surface scanning in Substance Designer 6 I wanted to see if it was as good as advertised.
I did a few tests with different materials and for the most part it's pretty good with little effort. The "Smart Tile" node works pretty good although it's output greatly varies based on how random the texture patter is. All materials here are tiled between 2-4 times, some worked out better then others.
Photos were taken in RAW format , adjusted, and cropped square in Lightroom, then exported as 4k 16bit Tiff.
Tiffs were imported to Designer and put into the scanning template. Images were then rendered in Marmoset.
Objects scanned starting at top left:
-Cotton Sweatshirt
-Corduroy Button down shirt
-Wool knit scarf
-Vinyl fabric
-Smooth pine board
-Rough pine board
My scanning setup. Flash was raised about 1ft off the ground (not shown here). White tape shows light positions.
One of the sets of source images showing the position of the light source for each.
Substance graph setup. Most looked like this with some needing 1-2 "Material clone" nodes to remove blemishes or to help it tile.
Cotton sweatshirt
Cotton sweatshirt
Cotton sweatshirt
Corduroy button down shirt
Corduroy button down shirt
Corduroy button down shirt
Wool knit scarf
Wool knit scarf
Wool knit scarf
Vinyl fabric
Vinyl fabric
Vinyl fabric
Smooth pine board
Smooth pine board
Smooth pine board
Rough pine board
Rough pine board
Rough pine board
Some Cloth sims