MadMapper “Materials” are generative animated visuals. They are directly rendered on the screen by the GPU (graphics card). They are composed of a fragment shader file and optionnaly a vertex shader file.
MadMapper Materials are based on ISF specification: http://interactiveshaderformat.com/
Because they are not rendering in a texture, but directly on the output, the material shader code is mixed with the surface shader (for different surface types: Quad, Surface3D, Lines or Fixtures). So the “Materials” for MadMapper are not standard ISF shaders. But they use the same jSon header format, same builtin variables (such as TIME), same macros (IMG_NORM_PIXEL etc), they can use texture images the same way etc.
It’s almost an ISF shader except that you don’t implement the function “main”, but a function called materialColorForPixel that will return a color depending on a 2D coordinate.
The Material consist of a folder with a thumbnail (file name must be “thumbnail.png”), a fragment shader (.fs) and optionnaly a vertex shader (.vs).
The vertex and fragment shaders are the one mixed with a MadMapper internal shader that is specific to the surface type (Quad, Surface 3D, Lines, Fixture etc.) so they must implement a function with a defined name that will be called by the surface shader. Using the vertex shader is sometimes good for performances, if you have a computation to do that is the same for all pixels, better doing it once per vertex (6 times for a simple quad) than once per pixel (2 million times in HD resolution).
Shaders are compiled using GLSL version 150 core to be sure they work on a wide range of computers.
Simple example of fragment shader:
// This is a Json header following the ISF specifications for defining parameters
// that will appear as a slider, a checkbox, a combox... in MadMapper Material settings
/*{
"CREDIT": "MM Team",
"CATEGORIES": [
"Demo"
],
"INPUTS": [
{
"Label": "Red",
"NAME": "red",
"TYPE": "float",
"MIN": 0.0,
"MAX": 1.0,
"DEFAULT": 0.5
}
{
"Label": "Green",
"NAME": "green",
"TYPE": "float",
"MIN": 0.0,
"MAX": 1.0,
"DEFAULT": 0.5
}
{
"Label": "Blue",
"NAME": "blue",
"TYPE": "float",
"MIN": 0.0,
"MAX": 1.0,
"DEFAULT": 0.5
}
]
}*/
// This is the function MadMapper surface shader calls
// texCoord is the texture coordinate (position in the input media for this pixel)
vec4 materialColorForPixel( vec2 texCoord )
{
return vec4(red,green,blue,1);
}
Simple example of vertex shader:
// Declare an output variable we can use as input in the fragment shader
// to avoid doing this computation for each pixel!
out float myPrecomputedData;
// This function makes a pseudo random value
float makeRandomValue() {
return fract(sin(dot(co.xy ,vec2(12.9898,78.233))) * TIME * 43758.5453);
}
// This is the function MadMapper surface shader calls
// uv is the texture coordinate for this vertex (position in the input media)
void materialVsFunc(vec2 uv) {
myPrecomputedData = makeRandomValue()
}
MadMapper provides extensions to the spec. And some restrictions.
We also provide in this GitHub libraries that were written by Simon Gelfius for MadMapper, they are open source, check file headers for more info.
You will find here all details about writing Materials and using MadNoise.glsl and MadSDF.glsl libraries.
It is very easy to convert a standard ISF to a Material. In general renaming the function “void main()” with “vec4 materialColorForPixel(vec2 texCoord)” and use “texCoord” instead of “isf_FragNormCoord”
Example:
void main() {
vec2 uv = isf_FragNormCoord.xy;
vec4 color;
...
gl_FragColor = color;
}
Is replaced with:
vec4 materialColorForPixel(vec2 texCoord)
{
vec2 uv = texCoord;
vec4 color;
...
return color;
}
MadMapper accepts including glsl files. Ie
#include “MyCommonFunctions.glsl”
Or
#include “…/MyCommonFunctions.glsl”
The path in the include statement must be valid relatively to the file containing the include statement, except for the few libraries we bundle with application (for Noises, Signed Distance Field & MadCommon):
#include “MadCommon.glsl”
#include “MadNoise.glsl”
#include “MadSDF.glsl”
The specification of ISF only allow declaring 2D texture. We added a TYPE field to allow loading cube map and 3D textures. Also we added texture filtering and texture wrapping settings.
Here is a complete example.
/*{
"CREDIT": "mm team",
"CATEGORIES": [ "Image Control" ],
"INPUTS": [
...
],
"IMPORTED": {
{ "NAME": myImage",
"TYPE": "2D",
"PATH": "image.png",
"GL_TEXTURE_MIN_FILTER": "LINEAR",
"GL_TEXTURE_MAG_FILTER": "LINEAR",
"GL_TEXTURE_WRAP": "REPEAT"
},
{ "NAME": "myCubeMapWithFile",
"TYPE": "cube",
"PATH": ["posx.png","negx.png","posy.png","negy.png","posz.png","negz.png"]
},
{ "NAME": "myTexture3d",
"TYPE": "3D",
"PATH": "colorGrading.png",
"DEPTH": 32
}
}
}*/
“myImage” is the name of the uniform sample (“uniform sampler2D myImage;” is inserted in the final shader)
Parameters:
“TYPE”: "2D, “CUBE_MAP”, “3D”
“PATH”: path of the file (or folder for cube map)
“GL_TEXTURE_MIN_FILTER”: refer to https://www.khronos.org/opengles/sdk/docs/man/xhtml/glTexParameter.xml
possible values: “LINEAR”, “NEAREST”
“GL_TEXTURE_MAG_FILTER”: refer to https://www.khronos.org/opengles/sdk/docs/man/xhtml/glTexParameter.xml
possible values: “LINEAR”, “NEAREST”
“GL_TEXTURE_MIN_FILTER”:
possible values: “LINEAR”, “NEAREST”, “LINEAR_MIPMAP_LINEAR”, “NEAREST_MIPMAP_NEAREST”, “LINEAR_MIPMAP_NEAREST”, “NEAREST_MIPMAP_LINEAR”, “LINEAR_MIPMAP_LINEAR”
“GL_TEXTURE_WRAP”: refer to https://www.khronos.org/opengles/sdk/docs/man/xhtml/glTexParameter.xml
possible values: “REPEAT”, “CLAMP_TO_EDGE”, “MIRRORED_REPEAT”
“DEPTH”: depth of the 3D texture. This parameter is optionnal.
If not defined, we except a texture with Width = Height * Height, ie 256x32 pixels composed of 32 squares of 32x32 pixels next to each other.
If specifying a depth, image width must be a multiple of Depth. If DEPTH is 64 and the full image is 256x32, it will make a 16x32x64 texture (width=16, height=32, depth=64)
With shaders, we can’t store a value that we’ll reuse when processing next frame (except using multiple render passes and storing a values in pixels of a persisten buffer, but it makes things a bit complicated to solve a simple problem and MadMapper doesn’t support render passes for materials since if you have 400 surfaces using the same material, you wouldn’t want to do 400 more render passes just to handle a proper animation…)
This problem is inherant to the fact that the shader code is executed in parallel for each vertex/pixel.
To animate something in a shader, you might write something like:
float linePosition = sin(mat_speed * TIME);
Where mat_speed is an “INPUT” uniform controlled by user & TIME is the time in seconds (as defined in the ISF spec).
The problem with that is that when changing the speed, we got jumps in the animation, ie:
frame N: TIME=120.0, speed=1 => linePosition=sin(120.0)=0.866
frame N+1: TIME=120.16, speed=1 => linePosition=sin(120.166)=0.864
That’s ok, now user sets speed to 0.2
frame N+2: TIME=120.32, speed=0.2 => linePosition=sin(24.064)=0.407
The guy just wanted to reduce animation speed and the animation moved to a very different place, not good !
To solve that we added “GENERATORS”. It doesn’t only solve this case, it allows adding stuffs to filter an INPUT uniform (ie Damper).
GENERATORS are objects with parameters. Their output is a floating point value. For instance:
"GENERATORS": [
{
"NAME": "anim_time_base",
"TYPE": "time_base",
"PARAMS": {"speed": "noise_speed", "speed_curve": 2}
}
]
Let’s see each GENERATOR type, why they are useful & their parameters:
"time_base" is a generator that increments its output value (the uniform "move_position" in following example) depending on its "speed" parameter and elapsed time. Between each rendered frame it does something like: "output_value = output_value + speed * elapsedTime"
But there are more parameters, here is a complete example:
{
"NAME": "move_position",
"TYPE": "time_increment",
"FILTERS": {"speed": "automovespeed", "reverse": false, speed_curve":3, "strob": "automovestrob", "bpm_sync": "bpmsync"}
}
Parameters are all optionnal:
"speed": speed of the time base (floating point value)
"reverse": invert speed value (boolean)
"speed_curve": the input value will be interpreted as pow(INPUT_VALUE,speed_curve). Using a high speed curve allow to define precisely a low values and to be able to access high values at the same time. If you set a curve of 2 ot other even number, we'll take care to keep the sign (a negative number will stay negative)
"strob": the output value will not be updated on each frame, but each "strob" seconds, ie if strob=1, each 1 second, if strob=0.1, 10 times per second.
"bpm_sync": that's a handy one. The application has a global BPM, so if you set this parameter to "true", the speed and strob values will be interpreted as a divider or multiplier of the beat. For instance:
If your speed is 1, the output value should be increased by 1 each second
If BPM sync is on:
If speed is 1 and BPM is 60, the outut value will also be increased by 1 each second
If speed is 0.5, speed will by quantized on a "power of 2" or "power of two dividor" of one beat: 1/64, 1/32, 1/16, 1/8, 1/4, 1/2, 1, 2, 4, 8, 16, 32 ... and outut value will be increased by "quantized_speed * elapsed_time * BPM/60"
"link_speed_to_global_bpm": links the speed to MadMapper Global BPM (in Master settigs tab). If this is set to "true", speed will be multiplied by global BPM / 120. It has no effect if "bpm_sync" is activated.
"animator" is a shape generator (Smooth, In, Out, Linear, Cut) with many parameters. We made this case to avoid duplicating a lot of code in many materials, and so optimize shader code. It is based on the "time_base" generator above, so it has all the "time_base" parameters plus two others. It generates a value between 0 and 1.
Parameters are all optionnal:
Parameters of time_base define how the animation goes
"shape": possible values:
"Smooth": sinusoidal shape each cycle
"In": value goes from 0 to 1 in a cycle, then starts again
"Out": value goes from 1 to 0 in a cycle, then starts again
"Linear": value goes from 0 to 1 in half a cycle, then back to 0 in half a cycle
"Cut": value stays half a cycle at 0 then half a cycle at 1
Usefull to get a channel value (ie BPM value, BPM position, a MIDI/OSC/DMX channel ?) as a uniform, use pass_thru filter with the right channel URL as input.
"GENERATORS": [
{
"NAME": "bpm_position",
"TYPE": "pass_thru",
"PARAMS": {"input_value":"/custom/BPM/bpmPos"}
}
{
"NAME": "bpm_value",
"TYPE": "pass_thru",
"PARAMS": {"input_value":"/custom/BPM/bpm"}
}
]
"INPUTS": [
{
"Label": "Width X",
"NAME": "mat_width_x",
"TYPE": "float",
"MIN": 0.0,
"MAX": 1.0,
"DEFAULT": 0.5
}
],
"GENERATORS": [
{
"NAME": "mat_width_x_damped",
"TYPE": "damper",
"PARAMS": {"input_value": "mat_width_x", "hardness": 0.1, "damping": 0.5}
}
]
Parameters for damper: "hardness", "damping", "bypass", "size"
Useful one, like in audio application for keyboard->audio synth response, but without sustain cause that would have no sense in our case (no one is holding a key ;-). Parameters:
"attack": defines how fast output value can increase. If set to 0, when input value is higher than output value, output value will directly take the input value. If set to 1, it will need 10 seconds to go from 0 to 1. The maximum value increase at each computation is: maxIncrease = 10*pow(1-(0.01+attack*0.99),3) * elapsedSeconds
"decay": when the input value in decreasing, time in seconds before we start decreasing output value
"release": defines how fast output value can decrease. If set to 0, when input value is lower than output value (and decay time elapsed), output value will directly take the input value. If set to 1, it will need 10 seconds to go from 1 to 0. The maximum value increase at each computation is: maxDecrease = 10*pow(1-(0.01+release*0.99),3) * elapsedSeconds
We added some parameters compared to ISF spec. In the ISF spec, an “INPUT” of type “audioFFT” will be a texture with the audio spectrum. An “INPUT” with type “audio” will contain the waveform.
{
"NAME": "waveform",
"TYPE": "audio",
"SIZE": 512,
},
{
"NAME": "spectrum_16_decay",
"TYPE": "audioFFT",
"SIZE": 16,
"ATTACK": 0.0,
"DECAY": 0.4,
"RELEASE": 0.0
},
{
"NAME": "spectrum_16",
"TYPE": "audioFFT",
"SIZE": 16,
"ATTACK": 0.0,
"DECAY": 0.0,
"RELEASE": 0.2
}
Additionnal settings compared to the spec:
"SIZE": defines the size of the texture (number of values it will contain). 512 is the maximum. The texture resolution is SIZEx1 (height is always 1 pixel)
"ATTACK" / "DECAY" / "RELEASE": on spectrum (audioFFT) you can set those parameters and we'll apply an ADR filter for each value of the spectrum texture
When you add an INPUT, it creates a widget with a label. You can organise those widgets into “Group Boxes” by setting your label to “GroupName/Parameter Label”. The group order will be define by their appearance order in ISF header. Example:
"LABEL": "Global/Line Width",
"LABEL": "Global/Smooth",
"LABEL": "Global/BPM Sync",
"LABEL": "Noise/Noise",
"LABEL": "Noise/Amount",
"LABEL": "Noise/Speed",
"LABEL": "Move/Auto Move",
"LABEL": "Move/Size",
"LABEL": "Move/Speed",
"LABEL": "Move/Strob",
"LABEL": "Scale/Auto Scale",
"LABEL": "Scale/Size",
"LABEL": "Scale/Speed",
"LABEL": "Scale/Strob",
"LABEL": "Scale/Shape",
Here we’ll have 4 groupboxes: Global, Noise, Move, Scale, in this order. And widgets inside each are also organized by the order of appearance in the ISF header.
Generate as define: for optimization purpose, some “INPUT” uniforms can be handled differently by MadMapper. If you set the flag “generate_as_define” on an INPUT, we’ll compile the shader with a #define added to specify the actual value of the INPUT. It’s especially usefull with INPUTS of type “long” (enumeration). Then in the shader code, instead of writing “if (myValue == 0) {…} else …” you write “#ifdef myValue_IS_FirstPossibleValue … #endif”. Check the LinePattern material:
/{
“INPUTS”: [
{
“NAME”: “animation”,
“LABEL”: “Animation”,
“TYPE”: “long”,
“VALUES”: [“Cross”, “Centered Filled”,“Single”,“Filled”],
“DEFAULT”: “Cross”,
“FLAGS”: “generate_as_define”
}
]
}/
vec4 materialColorForPixel(vec2 texCoord)
{
#if defined(animation_IS_Cross)
…
#elif defined(animation_IS_Centered_Filled)
…
#elif defined(animation_IS_Single)
…
#else // defined(animation_IS_Filled)
…
#endif
…
}
You can also do it on a bool attribute:
{
"Label": "BassFX/Bass FX1",
"NAME": "bass_fx1",
"TYPE": "bool",
"DEFAULT": false,
"FLAGS": "generate_as_define,button"
}
...
#if defined(bass_fx1_IS_true)
...
#endif
Button: if you declare a “bool” INPUT, it will make a check box widget. Setting the “button” flag will make a Push Button.
{
“NAME”: “autorotateactive”,
“LABEL”: “Rotate/Auto Rotate”,
“TYPE”: “bool”,
“DEFAULT”: false,
“FLAGS”: “button”
},
SpinBox: if you declare an “int” or “float” INPUT, it will make a slider. Setting the “spinbox” flag, MadMapper will make a SpinBox.
{
“NAME”: “int_value”,
“LABEL”: “Int Parameter”,
“TYPE”: “int”,
“DEFAULT”: 0,
“MIN”: 0,
“MAX”: 10,
“FLAGS”: “spinbox”
},
MadMapper is delivered with 2 major libraries developped by Simon Gelfius: MadNoise and MadSDF. Here is a quick explanation. MadNoise.glsl is a library to generate 2D and 3D noises in different ways. MadSDF.glsl is a library to make 2D shapes easiely. SDF stands for Signed Distance Field.
MadCommon.glsl provides a few common constants & functions: PI, rgb2hsv conversion etc. To use them, just add
#include "MadCommon.glsl"
Full Header:
//! Constants
#define PI 3.1415926535897932384626433832795
//! Returns the luminance of a RGB color
float luma( vec3 color );
//! Returns the luminance of a RGB color
float luminance( vec3 color );
//! Luminance based rgb8/linear conversion
vec3 linearToRgb( vec3 color, vec3 range );
//! Luminance based rgb8/linear conversion
vec3 rgbToLinear( vec3 color, vec3 range )
//! Contrast, saturation, brightness, For all settings: 1.0 = 100% 0.5=50% 1.5 = 150%
vec3 applyContrastSaturationBrightness( vec3 color, float contrast, float saturation, float brightness )
//! Returns HSV values for RGB input color
vec3 rgb2hsv( vec3 color );
//! Returns RGB values for HSV input color
vec3 hsv2rgb( vec3 color );
//! Returns the Hue value for RGB input color
float rgb2hue( vec3 color );
//! Usefull shortcuts
#define saturate(x) clamp( x, 0.0, 1.0 )
#define lerp(x,y,t) mix( x, y, t )
MadNoise.glsl provides many noise functions. For instance, vnoise(xyPos) return a float depending on a vec2. To use them, just add
#include "MadNoise.glsl"
Before your shader code, after ISF header. Most of them can be optimized using a Noise Lookup texture. To use this optimization, just write
#define NOISE_TEXTURE_BASED
#include "MadNoise.glsl"
And add the Noise lookup texture in the imported textures of your material:
/*{
"CREDIT": "mm team",
"CATEGORIES": [ "Image Control" ],
"INPUTS": [
...
],
"IMPORTED": [
{"NAME": "noiseLUT", "PATH": "noiseLUT.png", "GL_TEXTURE_MIN_FILTER": "LINEAR", "GL_TEXTURE_MAG_FILTER": "LINEAR", "GL_TEXTURE_WRAP": "REPEAT"}
]
}/*
Full Header:
//! Returns a 2D value noise in the range [0,1]
float vnoise( vec2 p );
//! Returns a 3D value noise in the range [0,1]
float vnoise( vec3 p );
//! Returns a 2D value noise with derivatives in the range [0,1]
vec3 dvnoise( vec2 p );
//! Returns a 3D value noise with derivatives in the range [0,1]
vec4 dvnoise( vec3 p );
//! Returns a 2D simplex noise in the range [-1,1]
float noise( vec2 p );
//! Returns a 3D simplex noise in the range [-1,1]
float noise( vec3 p );
//! Returns a 2D simplex noise with derivatives in the range [-1,1]
vec3 dnoise( vec2 p );
//! Returns a 3D simplex noise with derivatives in the range [-1,1]
vec4 dnoise( vec3 p );
#if ! defined( NOISE_WORLEY_ASHIMA_IMPL )
//! Returns a 2D worley/cellular noise in the range [0,1]
float worleyNoise( vec2 p );
//! Returns a 3D worley/cellular noise in the range [0,1]
float worleyNoise( vec3 p );
#else
//! Returns a 2D worley/cellular noise F1 and F2 in the range [0,1]
vec2 worleyNoise( vec2 p );
//! Returns a 3D worley/cellular noise F1 and F2 in the range [0,1]
vec2 worleyNoise( vec3 p );
#endif
//! Returns a 2D worley/cellular noise with derivatives in the range [0,1]
vec3 dWorleyNoise( vec2 p );
//! Returns a 3D worley/cellular noise with derivatives in the range [0,1]
vec4 dWorleyNoise( vec3 p );
//! Returns a 2D simplex noise with rotating gradients in the range [-1,-1]
float flowNoise( vec2 pos, float rot );
//! Returns a 2D simplex noise with rotating gradients and derivatives in the range [-1,-1]
vec3 dFlowNoise( vec2 pos, float rot );
//! Returns a 2D Billowed Noise in the range [0,1]
float billowedNoise( vec2 p );
//! Returns a 3D Billowed Noise in the range [0,1]
float billowedNoise( vec3 p );
//! Returns a 2D Billowed Noise with Derivatives in the range [0,1]
vec3 dBillowedNoise( vec2 p );
//! Returns a 3D Billowed Noise with Derivatives in the range [0,1]
vec4 dBillowedNoise( vec3 p );
//! Returns a 2D Ridged Noise in the range [0,1]
float ridgedNoise( vec2 p );
//! Returns a 2D Ridged Noise in the range [0,1]
float ridgedNoise( vec2 p, float ridge );
//! Returns a 3D Ridged Noise in the range [0,1]
float ridgedNoise( vec3 p );
//! Returns a 3D Ridged Noise in the range [0,1]
float ridgedNoise( vec3 p, float ridge );
//! Returns a 2D Ridged Noise with Derivatives in the range [0,1]
vec3 dRidgedNoise( vec2 p );
//! Returns a 2D Ridged Noise with Derivatives in the range [0,1]
vec3 dRidgedNoise( vec2 p, float ridge );
//! Returns a 3D Ridged Noise with Derivatives in the range [0,1]
vec4 dRidgedNoise( vec3 p );
//! Returns a 3D Ridged Noise with Derivatives in the range [0,1]
vec4 dRidgedNoise( vec3 p, float ridge );
//! Returns a 2D Curl of a Simplex Noise in the range [-1,1]
vec2 curlNoise( vec2 v );
//! Returns a 2D Curl of a Simplex Flow Noise in the range [-1,1]
vec2 curlNoise( vec2 v, float t );
//! Returns a 2D Fractal Brownian Motion sum of Simplex Noise in the range [-1,1]
float fBm( vec2 p );
//! Returns a 2D Fractal Brownian Motion sum of Simplex Noise in the range [-1,1]
float fBm( vec2 p, int octaves, float lacunarity, float gain );
//! Returns a 3D Fractal Brownian Motion sum of Simplex Noise in the range [-1,1]
float fBm( vec3 p );
//! Returns a 3D Fractal Brownian Motion sum of Simplex Noise in the range [-1,1]
float fBm( vec3 p, int octaves, float lacunarity, float gain );
//! Returns a 2D Fractal Brownian Motion sum of Simplex Noise in the range [-1,1]
vec3 dfBm( vec2 p );
//! Returns a 2D Fractal Brownian Motion sum of Simplex Noise with Derivatives in the range [-1,1]
vec3 dfBm( vec2 p, int octaves, float lacunarity, float gain );
//! Returns a 3D Fractal Brownian Motion sum of Simplex Noise with Derivatives in the range [-1,1]
vec4 dfBm( vec3 p );
//! Returns a 3D Fractal Brownian Motion sum of Simplex Noise with Derivatives in the range [-1,1]
vec4 dfBm( vec3 p, int octaves, float lacunarity, float gain );
//! Returns a 2D Fractal Brownian Motion sum of Billowed Noise in the range [-1,1]
float billowyTurbulence( vec2 p );
//! Returns a 2D Fractal Brownian Motion sum of Billowed Noise in the range [-1,1]
float billowyTurbulence( vec2 p, int octaves, float lacunarity, float gain );
//! Returns a 3D Fractal Brownian Motion sum of Billowed Noise in the range [-1,1]
float billowyTurbulence( vec3 p );
//! Returns a 3D Fractal Brownian Motion sum of Billowed Noise in the range [-1,1]
float billowyTurbulence( vec3 p, int octaves, float lacunarity, float gain );
//! Returns a 2D Fractal Brownian Motion sum of Billowed Noise with Derivatives in the range [-1,1]
vec3 dBillowyTurbulence( vec2 p );
//! Returns a 2D Fractal Brownian Motion sum of Billowed Noise with Derivatives in the range [-1,1]
vec3 dBillowyTurbulence( vec2 p, int octaves, float lacunarity, float gain );
//! Returns a 3D Fractal Brownian Motion sum of Billowed Noise with Derivatives in the range [-1,1]
vec4 dBillowyTurbulence( vec3 p );
//! Returns a 3D Fractal Brownian Motion sum of Billowed Noise with Derivatives in the range [-1,1]
vec4 dBillowyTurbulence( vec3 p, int octaves, float lacunarity, float gain );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec2 p );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec2 p, float ridgeOffset );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec2 p, float ridgeOffset, int octaves, float lacunarity, float gain );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec3 p );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec3 p, float ridgeOffset );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
float ridgedMF( vec3 p, float ridgeOffset, int octaves, float lacunarity, float gain );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise in the range [-1,1]
vec3 dRidgedMF( vec2 p );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise with Derivatives in the range [-1,1]
vec3 dRidgedMF( vec2 p, float ridgeOffset );
//! Returns a 2D Ridged Multi-Fractal sum of Simplex Noise with Derivatives in the range [-1,1]
vec3 dRidgedMF( vec2 p, float ridgeOffset, int octaves, float lacunarity, float gain );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise with Derivatives in the range [-1,1]
vec4 dRidgedMF( vec3 p );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise with Derivatives in the range [-1,1]
vec4 dRidgedMF( vec3 p, float ridgeOffset );
//! Returns a 3D Ridged Multi-Fractal sum of Simplex Noise with Derivatives in the range [-1,1]
vec4 dRidgedMF( vec3 p, float ridgeOffset, int octaves, float lacunarity, float gain );
Check MadSDF.md for more info
Caution: you shouldn’t use those features except if you have very specifical needs and you’re doing it only for you - NOT FOR SHARING with other people
When writing a Material in MadMapper, you can test what kind of surface is actually using it using those definitions:
You can specify GLSL version in ISF header. By default we use version “150 core” (june 2017, version 3.0.0).
/*{
"CREDIT": "mm team",
"CATEGORIES": [ "Image Control" ],
"GLSL_VERSION": "150 core",
...
}/*