mirror of
https://github.com/godotengine/godot.git
synced 2024-12-15 10:12:40 +08:00
4226d56ca9
As of clang-format 6.0.1, putting the `/* clang-format off */` hint around our "invalid" `[vertex]` and `[shader]` statements isn't enough to prevent a bogus indent of the next comments and first valid statement, so we need to enclose that first valid statement in the unformatted chunk.
175 lines
4.8 KiB
GLSL
175 lines
4.8 KiB
GLSL
/* clang-format off */
|
|
[vertex]
|
|
|
|
layout(location = 0) in highp vec4 vertex_attrib;
|
|
/* clang-format on */
|
|
layout(location = 4) in vec2 uv_in;
|
|
|
|
out vec2 uv_interp;
|
|
|
|
void main() {
|
|
|
|
uv_interp = uv_in;
|
|
gl_Position = vertex_attrib;
|
|
}
|
|
|
|
/* clang-format off */
|
|
[fragment]
|
|
|
|
//#define QUALIFIER uniform // some guy on the interweb says it may be faster with this
|
|
#define QUALIFIER const
|
|
|
|
#ifdef USE_25_SAMPLES
|
|
const int kernel_size = 25;
|
|
/* clang-format on */
|
|
QUALIFIER vec2 kernel[25] = vec2[](
|
|
vec2(0.530605, 0.0),
|
|
vec2(0.000973794, -3.0),
|
|
vec2(0.00333804, -2.52083),
|
|
vec2(0.00500364, -2.08333),
|
|
vec2(0.00700976, -1.6875),
|
|
vec2(0.0094389, -1.33333),
|
|
vec2(0.0128496, -1.02083),
|
|
vec2(0.017924, -0.75),
|
|
vec2(0.0263642, -0.520833),
|
|
vec2(0.0410172, -0.333333),
|
|
vec2(0.0493588, -0.1875),
|
|
vec2(0.0402784, -0.0833333),
|
|
vec2(0.0211412, -0.0208333),
|
|
vec2(0.0211412, 0.0208333),
|
|
vec2(0.0402784, 0.0833333),
|
|
vec2(0.0493588, 0.1875),
|
|
vec2(0.0410172, 0.333333),
|
|
vec2(0.0263642, 0.520833),
|
|
vec2(0.017924, 0.75),
|
|
vec2(0.0128496, 1.02083),
|
|
vec2(0.0094389, 1.33333),
|
|
vec2(0.00700976, 1.6875),
|
|
vec2(0.00500364, 2.08333),
|
|
vec2(0.00333804, 2.52083),
|
|
vec2(0.000973794, 3.0));
|
|
#endif //USE_25_SAMPLES
|
|
|
|
#ifdef USE_17_SAMPLES
|
|
const int kernel_size = 17;
|
|
QUALIFIER vec2 kernel[17] = vec2[](
|
|
vec2(0.536343, 0.0),
|
|
vec2(0.00317394, -2.0),
|
|
vec2(0.0100386, -1.53125),
|
|
vec2(0.0144609, -1.125),
|
|
vec2(0.0216301, -0.78125),
|
|
vec2(0.0347317, -0.5),
|
|
vec2(0.0571056, -0.28125),
|
|
vec2(0.0582416, -0.125),
|
|
vec2(0.0324462, -0.03125),
|
|
vec2(0.0324462, 0.03125),
|
|
vec2(0.0582416, 0.125),
|
|
vec2(0.0571056, 0.28125),
|
|
vec2(0.0347317, 0.5),
|
|
vec2(0.0216301, 0.78125),
|
|
vec2(0.0144609, 1.125),
|
|
vec2(0.0100386, 1.53125),
|
|
vec2(0.00317394, 2.0));
|
|
#endif //USE_17_SAMPLES
|
|
|
|
#ifdef USE_11_SAMPLES
|
|
const int kernel_size = 11;
|
|
QUALIFIER vec2 kernel[11] = vec2[](
|
|
vec2(0.560479, 0.0),
|
|
vec2(0.00471691, -2.0),
|
|
vec2(0.0192831, -1.28),
|
|
vec2(0.03639, -0.72),
|
|
vec2(0.0821904, -0.32),
|
|
vec2(0.0771802, -0.08),
|
|
vec2(0.0771802, 0.08),
|
|
vec2(0.0821904, 0.32),
|
|
vec2(0.03639, 0.72),
|
|
vec2(0.0192831, 1.28),
|
|
vec2(0.00471691, 2.0));
|
|
#endif //USE_11_SAMPLES
|
|
|
|
uniform float max_radius;
|
|
uniform float camera_z_far;
|
|
uniform float camera_z_near;
|
|
uniform float unit_size;
|
|
uniform vec2 dir;
|
|
in vec2 uv_interp;
|
|
|
|
uniform sampler2D source_diffuse; //texunit:0
|
|
uniform sampler2D source_sss; //texunit:1
|
|
uniform sampler2D source_depth; //texunit:2
|
|
|
|
layout(location = 0) out vec4 frag_color;
|
|
|
|
void main() {
|
|
|
|
float strength = texture(source_sss, uv_interp).r;
|
|
strength *= strength; //stored as sqrt
|
|
|
|
// Fetch color of current pixel:
|
|
vec4 base_color = texture(source_diffuse, uv_interp);
|
|
|
|
if (strength > 0.0) {
|
|
|
|
// Fetch linear depth of current pixel:
|
|
float depth = texture(source_depth, uv_interp).r * 2.0 - 1.0;
|
|
#ifdef USE_ORTHOGONAL_PROJECTION
|
|
depth = ((depth + (camera_z_far + camera_z_near) / (camera_z_far - camera_z_near)) * (camera_z_far - camera_z_near)) / 2.0;
|
|
float scale = unit_size; //remember depth is negative by default in OpenGL
|
|
#else
|
|
depth = 2.0 * camera_z_near * camera_z_far / (camera_z_far + camera_z_near - depth * (camera_z_far - camera_z_near));
|
|
float scale = unit_size / depth; //remember depth is negative by default in OpenGL
|
|
#endif
|
|
|
|
// Calculate the final step to fetch the surrounding pixels:
|
|
vec2 step = max_radius * scale * dir;
|
|
step *= strength; // Modulate it using the alpha channel.
|
|
step *= 1.0 / 3.0; // Divide by 3 as the kernels range from -3 to 3.
|
|
|
|
// Accumulate the center sample:
|
|
vec3 color_accum = base_color.rgb;
|
|
color_accum *= kernel[0].x;
|
|
#ifdef ENABLE_STRENGTH_WEIGHTING
|
|
float color_weight = kernel[0].x;
|
|
#endif
|
|
|
|
// Accumulate the other samples:
|
|
for (int i = 1; i < kernel_size; i++) {
|
|
// Fetch color and depth for current sample:
|
|
vec2 offset = uv_interp + kernel[i].y * step;
|
|
vec3 color = texture(source_diffuse, offset).rgb;
|
|
|
|
#ifdef ENABLE_FOLLOW_SURFACE
|
|
// If the difference in depth is huge, we lerp color back to "colorM":
|
|
float depth_cmp = texture(source_depth, offset).r * 2.0 - 1.0;
|
|
|
|
#ifdef USE_ORTHOGONAL_PROJECTION
|
|
depth_cmp = ((depth_cmp + (camera_z_far + camera_z_near) / (camera_z_far - camera_z_near)) * (camera_z_far - camera_z_near)) / 2.0;
|
|
#else
|
|
depth_cmp = 2.0 * camera_z_near * camera_z_far / (camera_z_far + camera_z_near - depth_cmp * (camera_z_far - camera_z_near));
|
|
#endif
|
|
|
|
float s = clamp(300.0f * scale * max_radius * abs(depth - depth_cmp), 0.0, 1.0);
|
|
color = mix(color, base_color.rgb, s);
|
|
#endif
|
|
|
|
// Accumulate:
|
|
color *= kernel[i].x;
|
|
|
|
#ifdef ENABLE_STRENGTH_WEIGHTING
|
|
float color_s = texture(source_sss, offset).r;
|
|
color_weight += color_s * kernel[i].x;
|
|
color *= color_s;
|
|
#endif
|
|
color_accum += color;
|
|
}
|
|
|
|
#ifdef ENABLE_STRENGTH_WEIGHTING
|
|
color_accum /= color_weight;
|
|
#endif
|
|
frag_color = vec4(color_accum, base_color.a); //keep alpha (used for SSAO)
|
|
} else {
|
|
frag_color = base_color;
|
|
}
|
|
}
|