4bfafa4d8e
Rewamp the scene api, add circle plots, add blender flips, fix caching issue in driver, fix bug in partialSvg. Former-commit-id: 5cb900521044fee13766dd7474fd9b72c7691686
1.6 KiB
On its own, data doesn't look like much. Often it is merely a wall of unintelligible numbers.
However, if we take each number and assign it a shade of grey, suddenly the data becomes understandable and the face of the monalisa stares out at us.
Using shades of grey is not the only possible color palette, though, and many colorful alternatives exists. Some of these colormaps have been around for a while, and Jet, one of the oldest, first appeared in the 1970s. However, in recent years, a lot of attention has gone into addressing the shortcomings of the early colormaps and create better standards for the future. Particularly, Matlab replaced Jet by Parula in 2014, Viridis became the default in Matplotlib by 2015, the research paper describing Cividis was published in 2018, and Google created Turbo as a spiritual successor to Jet in 2019.
So, why were so many new colormaps invented in this 5-year period? How are these colormaps created? Are they merely the favorite colors of their creators? How can a colormap be interesting enough to merit a research publication?
To answer these questions, first we have to explore a bit of color theory.
Visible light roughly ranges from a wavelength of 400nm to 700nm. If each combination of wavelengths gave rise to a unique color then creating a color space would be nigh impossible. Fortunately, most human eyes have just three types of light-sensitive cells that respond to ranges of wavelengths, and the space of colors is therefore reduced to three dimensions. The axes are called S, M and L because the corrosponding cones are sensitive to short, medium, and lone wavelengths respectively.