Added vibrance filter (for increasing saturation of muted colors)

[melchiar]

Added a Vibrance filter for increasing the saturation of an image's more muted colors. Formula based on Caman.js

https://jsfiddle.net/melchiar/oq4eydrj/

[github-actions]

Code Coverage Summary

> [email protected] coverage:report /home/runner/work/fabric.js/fabric.js
> nyc report --reporter=lcov --reporter=text

-----------|---------|----------|---------|---------|-----------------------------------------------
File       | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s                             
-----------|---------|----------|---------|---------|-----------------------------------------------
All files  |    83.3 |       77 |   85.99 |   83.01 |                                               
 fabric.js |    83.3 |       77 |   85.99 |   83.01 | ...,29780,29905,29985-30050,30173,30272,30489 
-----------|---------|----------|---------|---------|-----------------------------------------------

[github-actions]

Code Coverage Summary

> [email protected] coverage:report /home/runner/work/fabric.js/fabric.js
> nyc report --reporter=lcov --reporter=text

-----------|---------|----------|---------|---------|-----------------------------------------------
File       | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s                             
-----------|---------|----------|---------|---------|-----------------------------------------------
All files  |    83.3 |       77 |   85.99 |   83.01 |                                               
 fabric.js |    83.3 |       77 |   85.99 |   83.01 | ...,29781,29906,29986-30051,30174,30273,30490 
-----------|---------|----------|---------|---------|-----------------------------------------------

[asturur]

ok is written here, but we can put this also over the vibrance property

[melchiar]

done

[asturur]

Should we add a test?

[asturur]

i can setup a test suite for all the filters maybe since you don't have examples.
We could use that image as reference

[github-actions]

Code Coverage Summary

> [email protected] coverage:report /home/runner/work/fabric.js/fabric.js
> nyc report --reporter=lcov --reporter=text

-----------|---------|----------|---------|---------|-----------------------------------------------
File       | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s                             
-----------|---------|----------|---------|---------|-----------------------------------------------
All files  |    83.3 |    76.98 |   85.99 |   83.01 |                                               
 fabric.js |    83.3 |    76.98 |   85.99 |   83.01 | ...,29789,29914,29994-30059,30182,30281,30498 
-----------|---------|----------|---------|---------|-----------------------------------------------

[melchiar]

yeah, I guess we don't really have any visual tests for filters yet, do we

[github-actions]

Code Coverage Summary

> [email protected] coverage:report /home/runner/work/fabric.js/fabric.js
> nyc report --reporter=lcov --reporter=text

-----------|---------|----------|---------|---------|-----------------------------------------------
File       | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s                             
-----------|---------|----------|---------|---------|-----------------------------------------------
All files  |    83.3 |    76.98 |   85.99 |   83.01 |                                               
 fabric.js |    83.3 |    76.98 |   85.99 |   83.01 | ...,29789,29914,29994-30059,30182,30281,30498 
-----------|---------|----------|---------|---------|-----------------------------------------------

[github-actions]

Code Coverage Summary

> [email protected] coverage:report /home/runner/work/fabric.js/fabric.js
> nyc report --reporter=lcov --reporter=text

-----------|---------|----------|---------|---------|-----------------------------------------------
File       | % Stmts | % Branch | % Funcs | % Lines | Uncovered Line #s                             
-----------|---------|----------|---------|---------|-----------------------------------------------
All files  |    83.3 |       77 |   85.99 |   83.01 |                                               
 fabric.js |    83.3 |       77 |   85.99 |   83.01 | ...,29797,29922,30002-30067,30190,30289,30506 
-----------|---------|----------|---------|---------|-----------------------------------------------

[melchiar]

changed this to 1 (100 makes no sense when the recommended range is -1 to 1)

[melchiar]

added a description to the variable saturation filter too since it was missing

[asturur]

I had this idea on my head for a bit:

this code here:

    /**
     * Return WebGL uniform locations for this filter's shader.
     *
     * @param {WebGLRenderingContext} gl The GL canvas context used to compile this filter's shader.
     * @param {WebGLShaderProgram} program This filter's compiled shader program.
     */
    getUniformLocations: function(gl, program) {
      return {
        uVibrance: gl.getUniformLocation(program, 'uVibrance'),
      };
    },

    /**
     * Send data from this filter to its shader program's uniforms.
     *
     * @param {WebGLRenderingContext} gl The GL canvas context used to compile this filter's shader.
     * @param {Object} uniformLocations A map of string uniform names to WebGLUniformLocation objects
     */
    sendUniformData: function(gl, uniformLocations) {
      gl.uniform1f(uniformLocations.uVibrance, -this.vibrance);
    },

is sort of repeated for each filter, but at the end is conceptually:

uniformsValeus = {
  uVibrance: function (filter) { return -filter.vibrance };
}

We still need a smart way to identify the type of the filter, in this case 1f

All the rest is wrapping code that could be in the base filter class.
i'm sure this would make the filters more easy to read and maybe save some byte here and there.

If not all, at least `getUniformLocation can be streamlined to a data declaration.