Blend modes (with a trick 😉)

[MarioMey]

I made a little shader that can mix images with different blend modes: Add, Screen, Multiply, Substract, Divide, ColorDodge, LinearDodge, ColorBurn and LinearBurn. You have to create a nested scene with both images to blend, one beside the other, like this:

And you use this code:

/* Blend modes
*/

#pragma shaderfilter set type__description Type +S*-/CdLdCbLb
// #pragma shaderfilter set type__description Type Add Screen Multiply Substract Divide ColorDodge LinearDodge ColorBurn LinearBurn
#pragma shaderfilter set type__step 1
#pragma shaderfilter set type__default 0
#pragma shaderfilter set type__min 0
#pragma shaderfilter set type__max 8
#pragma shaderfilter set type__slider true

#pragma shaderfilter set inf__description Influence
#pragma shaderfilter set inf__step 0.01
#pragma shaderfilter set inf__default 1.0
#pragma shaderfilter set inf__min 0.0
#pragma shaderfilter set inf__max 1.0
#pragma shaderfilter set inf__slider true

uniform float type;
uniform float inf;

float4 render(float2 uv) {
	float4 img1 = texture(image, vec2(uv.x / 2        , uv.y));
	float4 img2 = texture(image, vec2(uv.x / 2 + 0.5 , uv.y));
	
	// Add
	if (type == 0)
		{
		vec4 add = vec4(img1 + img2);
		return mix(texture(image, vec2(uv.x/2, uv.y)), add, inf);
		}
	// Screen
	else if (type == 1)
		{
		vec4 screen = vec4(1 - (1 - img1) * (1 - img2));
		return mix(texture(image, vec2(uv.x/2, uv.y)), screen, inf);
		}
	// Multiply
	else if (type == 2)
		{
		vec4 multiply = vec4(img1 * img2);
		return mix(texture(image, vec2(uv.x/2, uv.y)), multiply, inf);
		}
	// Substract
	else if (type == 3)
		{
		vec4 substract = vec4(img2 - img1);
		return mix(texture(image, vec2(uv.x/2, uv.y)), substract, inf);
		}
	// Divide
	else if (type == 4)
		{
		vec4 divide = vec4(img2 / img1);
		return mix(texture(image, vec2(uv.x/2, uv.y)), divide, inf);
		}

	// Color Dodge
	else if (type == 5)
		{
		// vec4 dodge = vec4((255 * img1) / ((255 - img2) + 1));
		vec4 colordodge = vec4((img2) / (1.0 - img1));
		return mix(texture(image, vec2(uv.x/2, uv.y)), colordodge, inf);
		}
	// Linear Dodge
	else if (type == 6)
		{
		// vec4 dodge = vec4((255 * img1) / ((255 - img2) + 1));
		vec4 lineardodge = vec4((img2) / (1.0 - img1));
		return mix(texture(image, vec2(uv.x/2, uv.y)), lineardodge, inf);
		}
	// Color Burn
	else if (type == 7)
		{
		vec4 colorburn = vec4(1.0 - (1.0 - img2)/(img1));
		return mix(texture(image, vec2(uv.x/2, uv.y)), colorburn, inf);
		}
	// Linear Burn
	else if (type == 8)
		{
		vec4 linearburn = vec4(img1 + (img2) - 1);
		return mix(texture(image, vec2(uv.x/2, uv.y)), linearburn, inf);
		}
}

... and voilá, you have something like:

Yes, you loose the half of width quality... but it's ok 🤷.