In ResourceFunction["CompileColorFunction"][f], the function f should yield a valid color (as determined by ColorQ) when given real numbers between 0 and 1.

The compiled color function yields a vector of reals representing a color in the space specified by the ColorSpace option.

Examples

Basic Examples (3)

Compile a color function:

In[1]:=

Out[1]=

In[2]:=

Out[2]=

See the resulting color:

In[3]:=

Out[3]=

Compile a built-in color function from ColorData:

In[4]:=

Out[4]=

Apply it to some data:

In[5]:=

Out[5]=

Get the resulting image:

In[6]:=

Out[6]=

Use a compiled color function in a plot:

In[7]:=

Out[7]=

In[8]:=

Out[8]=

Scope (2)

Generate colorized images from tables of data:

In[9]:=

Out[9]=

In[10]:=

In[11]:=

Out[11]=

Create a fast color function from a slow one:

In[12]:=

In[13]:=

Out[13]=

In[14]:=

Compare the performance:

In[15]:=

$AbsoluteTiming[ Plot3D[ Exp[1 - x^2 - y^2], {x, -2, 2}, {y, -2, 2}, ColorFunction -> Function[{x, y, z}, slow[z]], PlotRange -> All, Mesh -> None]]$

Out[15]=

In[16]:=

$AbsoluteTiming[ Plot3D[ Exp[1 - x^2 - y^2], {x, -2, 2}, {y, -2, 2}, ColorFunction -> Function[{x, y, z}, fast[z]], PlotRange -> All, Mesh -> None]]$

Out[16]=

Inspect timing details:

In[17]:=

$ResourceFunction["EvaluationTiming"][ Plot3D[Exp[1 - x^2 - y^2], {x, -2, 2}, {y, -2, 2}, ColorFunction -> Function[{x, y, z}, slow[z]], PlotRange -> All, Mesh -> None]]["Summary"]$

Out[17]=

In[18]:=

Out[18]=

The compiled version contributes an insignificant amount of overhead:

In[19]:=

$ResourceFunction["EvaluationTiming"][ Plot3D[Exp[1 - x^2 - y^2], {x, -2, 2}, {y, -2, 2}, ColorFunction -> Function[{x, y, z}, fast[z]], PlotRange -> All, Mesh -> None]]["Summary"]$

Out[19]=

In[20]:=

Out[20]=

Options (13)

ColorSamples (2)

Gradients that have a high amount of variance can lose some detail when compiled:

In[21]:=

Out[21]=

In[22]:=

In[23]:=

Out[23]=

In[24]:=

Out[24]=

In[25]:=

Out[25]=

Use more initial color samples to increase the quality:

In[26]:=

Out[26]=

In[27]:=

Out[27]=

RuntimeAttributes (4)

By default, CompileColorFunction will use Listable in the RuntimeAttributes:

In[28]:=

Out[28]=

This allows the compiled function to be applied directly to an array of data:

In[29]:=

Out[29]=

In[30]:=

Out[30]=

If compiled without the Listable attribute, the function needs to be applied to individual values directly:

In[31]:=

Out[31]=

In[32]:=

Out[32]=

It works if mapped:

In[33]:=

Out[33]=

Parallelization (2)

By default, Parallelization is set to True for CompileColorFunction:

In[34]:=

Out[34]=

In[35]:=

In[36]:=

Out[36]=

Force the function to evaluate on a single thread:

In[37]:=

Out[37]=

In[38]:=

Out[38]=

CompilationTarget (2)

If a C compiler is available on the current machine, additional performance may be obtained by setting CompilationTarget to "C":

In[39]:=

Out[39]=

In[40]:=

$data = Table[(1/10) (5 + y + Sin[x^2 + y^2]), {x, -4, 4, 0.05}, { y, -4, 4, 0.05}];$

In[41]:=

Out[41]=

Use the default value of "WVM":

In[42]:=

Out[42]=

In[43]:=

Out[43]=

ColorSpace (3)

By default the values output by the compiled function will represent RGB values:

In[44]:=

Out[44]=

In[45]:=

$data = Table[(1/10) (5 + y + Sin[x^2 + y^2]), {x, -4, 4, 0.05}, { y, -4, 4, 0.05}];$

In[46]:=

Out[46]=

Specify a different color space:

In[47]:=

Out[47]=

In[48]:=

Now the ColorSpace needs to be specified for Image as well:

In[49]:=

Out[49]=

In[50]:=

Out[50]=

Properties and Relations (2)

Compiling a color function and applying directly to image data can improve performance over Colorize:

In[51]:=

Out[45]=

In[52]:=

Out[52]=

In[53]:=

Out[53]=

Compare to using an uncompiled color function:

In[54]:=

Out[54]=

Possible Issues (2)

In CompileColorFunction[f], if the function f does not produce a smooth gradient when evaluated from 0 to 1, the compiled function can have low quality results:

In[55]:=