layout (location = 0) in vec2 a_pos; layout (location = 1) in vec2 a_decimals_ed; layout (location = 2) in vec2 a_normal2d; out vec4 v_color; layout (std140) uniform FillExtrusionDrawableUBO { highp mat4 u_matrix; highp vec2 u_pixel_coord_upper; highp vec2 u_pixel_coord_lower; highp float u_height_factor; highp float u_tile_ratio; // Interpolations highp float u_base_t; highp float u_height_t; highp float u_color_t; highp float u_pattern_from_t; highp float u_pattern_to_t; lowp float drawable_pad1; }; layout (std140) uniform FillExtrusionTilePropsUBO { highp vec4 u_pattern_from; highp vec4 u_pattern_to; highp vec2 u_texsize; lowp float tileprops_pad1; lowp float tileprops_pad2; }; layout (std140) uniform FillExtrusionPropsUBO { highp vec4 u_color; highp vec3 u_lightcolor; lowp float props_pad1; highp vec3 u_lightpos; highp float u_base; highp float u_height; highp float u_lightintensity; highp float u_vertical_gradient; highp float u_opacity; highp float u_fade; highp float u_from_scale; highp float u_to_scale; lowp float props_pad2; }; #pragma mapbox: define highp float base #pragma mapbox: define highp float height #pragma mapbox: define highp vec4 color void main() { #pragma mapbox: initialize highp float base #pragma mapbox: initialize highp float height #pragma mapbox: initialize highp vec4 color vec3 normal = vec3(a_normal2d / 16384.0, a_normal2d.x == 0.0 && a_normal2d.y == 0.0 ? 1.0 : 0.0); base = max(0.0, base); height = max(0.0, height); float t = mod(a_decimals_ed.x, 2.0); float z = t > 0.0 ? height : base; vec2 decimals = unpack_float(floor(a_decimals_ed.x / 2.0)) / 128.0; gl_Position = u_matrix * vec4(a_pos + decimals, z, 1); // Relative luminance (how dark/bright is the surface color?) float colorvalue = color.r * 0.2126 + color.g * 0.7152 + color.b * 0.0722; v_color = vec4(0.0, 0.0, 0.0, 1.0); // Add slight ambient lighting so no extrusions are totally black vec4 ambientlight = vec4(0.03, 0.03, 0.03, 1.0); color += ambientlight; // Calculate cos(theta), where theta is the angle between surface normal and diffuse light ray float directional = clamp(dot(normal, u_lightpos), 0.0, 1.0); // Adjust directional so that // the range of values for highlight/shading is narrower // with lower light intensity // and with lighter/brighter surface colors directional = mix((1.0 - u_lightintensity), max((1.0 - colorvalue + u_lightintensity), 1.0), directional); // Add gradient along z axis of side surfaces if (normal.z == 0.0) { // This avoids another branching statement, but multiplies by a constant of 0.84 if no vertical gradient, // and otherwise calculates the gradient based on base + height float fMin = mix(0.7, 0.98, 1.0 - u_lightintensity); float factor = clamp((t + base) * pow(height / 150.0, 0.5), fMin, 1.0); directional *= (1.0 - u_vertical_gradient) + (u_vertical_gradient * factor); } // Assign final color based on surface + ambient light color, diffuse light directional, and light color // with lower bounds adjusted to hue of light // so that shading is tinted with the complementary (opposite) color to the light color v_color.r += clamp(color.r * directional * u_lightcolor.r, mix(0.0, 0.3, 1.0 - u_lightcolor.r), 1.0); v_color.g += clamp(color.g * directional * u_lightcolor.g, mix(0.0, 0.3, 1.0 - u_lightcolor.g), 1.0); v_color.b += clamp(color.b * directional * u_lightcolor.b, mix(0.0, 0.3, 1.0 - u_lightcolor.b), 1.0); v_color *= u_opacity; }