4aiman
9ceadbc7bbStreamFX pulled the plug, obs-shaderfilter is the way. Had to fix implicit casting
27 lines
1.3 KiB
GLSL
27 lines
1.3 KiB
GLSL
uniform int shadow_offset_x;
|
|
uniform int shadow_offset_y;
|
|
uniform int shadow_blur_size;
|
|
uniform float speed;
|
|
uniform float4 shadow_color;
|
|
uniform bool is_alpha_premultiplied;
|
|
|
|
float4 mainImage(VertData v_in) : TARGET
|
|
{
|
|
int shadow_blur_samples = int(pow(shadow_blur_size * 2 + 1, 2));
|
|
float4 color = image.Sample(textureSampler, v_in.uv);
|
|
float2 shadow_uv = float2(v_in.uv.x - uv_pixel_interval.x * shadow_offset_x,
|
|
v_in.uv.y - uv_pixel_interval.y * shadow_offset_y);
|
|
float sampled_shadow_alpha = 0;
|
|
for (int blur_x = -shadow_blur_size; blur_x <= shadow_blur_size; blur_x++)
|
|
{
|
|
for (int blur_y = -shadow_blur_size; blur_y <= shadow_blur_size; blur_y++)
|
|
{
|
|
float2 blur_uv = shadow_uv + float2(uv_pixel_interval.x * blur_x, uv_pixel_interval.y * blur_y);
|
|
sampled_shadow_alpha += image.Sample(textureSampler, blur_uv).a / shadow_blur_samples;
|
|
}
|
|
}
|
|
float t = elapsed_time * speed;
|
|
float4 final_shadow_color = float4(shadow_color.rgb, shadow_color.a * sampled_shadow_alpha* (1 + sin(t)) / 2);
|
|
float4 color_two = final_shadow_color * (1-color.a) + color * (int(is_alpha_premultiplied) * 1 + int(!is_alpha_premultiplied) * color.a);
|
|
return float4(color_two.r, color_two.g, color_two.b, color_two.a);
|
|
}
|