Computer Generated Watercolor

Size: px

Start display at page:

Download "Computer Generated Watercolor"

Josephine McGee
5 years ago
Views:

1 Computer Generated Watercolor Loren Block Lots of research has been done trying to simulate effect used by artist. The watercolor simulation described here was designed by University of Washington, Stanford University, and Pixar Studios. I find this research interesting, not because I expect to ever use these watercolor simulation algorithms, but because I think looking at the steps taken by these researchers give a good example of how artistic effects can be simulated.

2 Watercolor The first step taken by the researcher was to study what characteristics of watercolor make it appealing to artists, and focus on those aspects in the simulation. The elements of watercolor are broken down so that they can be more easily analyzed: Watercolor Paper: Watercolor paper is usually made from linen or cotton cloth. The cloth is pounded to the point where the paper because mostly air, just a web of small fibers. A cellulose called sizing is applied to the paper so that it isn t so extremely absorbent. At the end of the process the surface of the paper is very rough. Pigment: The individual partials are about.05 to.5 microns in diameter. Density and staining power very greatly between different pigments. Two properties that are exhibited by some pigments are: granulation and flocculation. Granulation is when partials tend to settle in to crevasses in the paper, and flocculation is when partials draw together into clumps. Binder and Surfactant: Binder is added to the pigment to help it adhere to the paper and surfactant is added to help the water soak into the paper. The concentration of these elements in the paint mixture is fixed in this simulation.

Figure 1 The important characteristics of watercolor are demonstrated in Figure 1. Figure 1a is the day-brush effect where a dry brush only applies pigment to the raised portion of the paper.

3 Figure 1 The important characteristics of watercolor are demonstrated in Figure 1. Figure 1a is the day-brush effect where a dry brush only applies pigment to the raised portion of the paper. Figure 1b is edge darkening, where a wet brush is applied to a dry paper causing the pigment to move to the edge of the wet area. Figure 1c is an example of backruns where water spreads back into damp sections that have already been painted. Figure 1d is an example of granulation. Figure 1e is wet-on-wet painting where a wet brush is applied to a wet surface. Figure 1f is glazing, where strokes are overlaided to create a certain effect.

Fluid Simulations The simulation is separated into three different layers: the sallowwater layer is above the surface of the paper, the pigment-deposition layer is where the pigment is absorbed

4 Fluid Simulations The simulation is separated into three different layers: the sallowwater layer is above the surface of the paper, the pigment-deposition layer is where the pigment is absorbed and/or lifted from the paper, and the capillary layer is where the absorbed water is diffused. (Figure 2) Figure 2 The creators of this simulation method created six rules to realistically simulate the fluid dynamics: 1. The flow must be constrained so that water remains within the wetarea mask. 2. A surplus of water in one area should cause flow outward from that area into nearby regions. 3. The flow must be damped to minimize oscillating waves. 4. The flow must be perturbed by the texture of the paper to cause streaks parallel to flow direction. 5. Local changes should have global effects. For example, adding water in a local area should affect the entire simulation. 6. There should be outward flow of the fluid toward the edges to produce the edge-darkening effect.

5 Each cell will have a value for each for the variables below denoted by the subscripts i, j. (x i, j ) The values needed to simulate the shallow-water layer are: (M) The wet-area mask. 1 for wet, 0 for dry. (u, v) The velocity of the water in x and y directions. (p) The pressure of the water. (g k ) The concentration for each pigment k in the water. (h, h) The height and the slope of the paper. The height is just a value between 0 and 1. The angle of the paper is not considered. The values for viscosity and the viscous drag are constants and set to.1 and.01 respectively in this simulation. The values needed to simulate the pigment-deposition layer are: (d k ) The concentration of each pigment k in the paper. (ρ, w, y) Density, staining power, and granularity of the pigment. The values needed to simulate the capillary layer are: (s) The water saturation of the paper. (c) The fluid-holding capacity of the paper. This can be calculated using the equation: c = h * (c max c min ) + c min.

6 proc MainLoop(M, u, v, p, g1,, gn, d1,, dn, s): for each time step do: MoveWater(M, u, v, p) MovePigment(M, u, v, g1,, gn ) TransferPigment(g1,, gn, d1,, dn ) SimulateCapillaryFlow(M, s) end for end proc proc MoveWater(M, u, v, p): UpdateVelocities(M, u, v, p) Handles rules 1 through 4 from above RelaxDivergence(M, u, v, p) Handles rules 5 and 6 from above FlowOutward(M, p) end proc The function UpdateVelocities() uses the equations from Figure 3 to calculate the new value of u, v. The function RelaxDivergence() moves the water into neighboring cells, unless the velocity field is less then some set tolerance t. The FlowOutward() function simulates the way water evaporates more quickly when it is closer to the dry areas. This creates the edge darkening effect. Figure 3

7 The MovePigments() function moves the pigment based on the u, v values calculated in the UpdateVelocities() function for each cell. The TransferPigment() function is where the pigment is moved between the shallow-water layer and the pigment-deposition layer. The scalar values of, ρ, w, y affect how pigment is absorbed or lifted from the paper. Backruns and dry brush effects are special cases. SimulateCapillaryFlows() deals with the backruns. Figure 4 Figure 4 shows the watercolor effect created using the methods described above.

8 Applications Virtual painting Image watercolorization Rendering of 3D models Reference The paper that I used as reference is Computer-Generated Watercolor by Cassidy Curtis, Sean Anderson, Joshua Seims, Kurt Fleischer, and David Salesin.

Organization. Watercolor Painting. Watercolor Materials. Watercolor Materials. Computer-Generated. Watercolor SIGGRAPH 97.

Organization. Watercolor Painting. Watercolor Materials. Watercolor Materials. Computer-Generated. Watercolor SIGGRAPH 97. Computer-Generated Watercolor By Cassidy J. Curtis, Sean E. Anderson, Joshua E. Steins, Kurt W. Fleischer and David H. Salesin SIGGRAPH 97 Aditya Bhatia Computer Science McGill University Introduction