Choosing colors for scientific 3D renderings

Deborah Schmidt

Head of Helmholtz Imaging Support Unit

Sep 25, 2024

Slides available at https://ida-mdc.github.io/workshop-visualization/2026-workshop/colors/

Color Representation in Digital Programs

RGB color representation: rgb(255, 87,51)

HEX color representation: #FF5733

HSL color representation: hsl(9, 100%, 60%)

Additive color mixing with Red, Green and Blue.

Transparency: rgba(255, 87,51, 255), #FF5733FF

Color palettes

https://coloors.co

Color and colormap theory

Harmonizing colors
- often found next to each other on the color wheel (like blue and green) , and they create a sense of balance in visualizations.
Contrasting colors
- like red and green, are on opposite sides of the color wheel and create visual tension, which can be useful for highlighting important data.
Perception of contrast varies
- we perceive contrast more strongly with some colors (like red) compared to others (like blue), making some colors more effective for highlighting key elements in visualizations.

Color and colormap theory

Choosing a colormap

Sequential colormaps
- are best for data that progress from low to high values (e.g., depth or intensity).
Diverging colormaps
- should be used when you have a critical midpoint in your data that you want to emphasize, such as positive vs. negative values.
Qualitative colormaps
- are ideal for categorical data with no inherent order (e.g., different materials or regions in a dataset).
Avoid using the Jet colormap
- as it can distort data, especially when used for sequential data. Instead, use perceptually uniform colormaps like Viridis or Cividis.

Color and colormap theory

Try it on real data

A smooth heightfield, colored by height. Smooth data is where a colormap’s faults show - on a specimen with hard edges you cannot tell whether an edge you see is in the data or in the palette.

Switch to Jet and look at the flanks of the hills. The rings are not in the data. Jet changes lightness fast through cyan and yellow and slowly through green, so it invents boundaries where the surface is smooth and hides real structure where it is flat. It is still the default in a lot of software.

Then switch Seen as to Greyscale. That is the print test, and it is also a test of whether the map has a single consistent ordering: viridis and the single-hue ramp survive it, jet collapses into nonsense.

Finally Deuteranope - roughly one man in twelve. Viridis keeps almost all of its range. Jet loses most of its.

The same heights, five colormaps. Jet’s rings are in the palette, not in the data.

Target audience expectations

Visualizing Science: How Color Determines What We See

Color accessibility

Fiji ▷ Image ▷ Color ▷ Simulate Color Blindness

Fiji Plugin: Simulate Color Blindness

Picking colors yourself and have fun!

Sometimes, the best color palettes can be inspired by your surroundings. Here are some fun and creative ways to choose colors for your scientific visualizations:

Video games: Many video games are beautifully designed with carefully chosen color palettes. Take a screenshot of a game with an atmosphere you like and use a color picker tool to extract the colors.
Photography: Take pictures of sunsets, landscapes, or even advertisements that catch your eye. These images often contain harmonious color schemes that can inspire your visualization.
Tools for Picking Colors: All major operating systems have color picker tools. You can use:
- macOS: The built-in Digital Color Meter.
- Windows: Tools like ColorPic or the Snipping Tool with its color picker extension (authors note: this is a suggestion from ChatGPT, I didn’t verify it).
- Linux: For example the Gpick color picker or KColorChooser.

Image editors or graphical design tools such as Gimp or Inkscape (staying in Open Source domain here) have built in color pickers as well.