The Markov's Dream Palette

When I created the work Markov's Window in 2004, I hand picked a simple colour palette, based on pure, fully saturated RGB colours.

• Blue: rgb(0, 0, 255)
• Green: rgb(0, 255, 0)
• Magenta / fuchsia: rgb(255, 0, 255)
• A black to white gradient: rgb(0, 0, 0) to rgb(255, 255, 255)

I have since returned to this palette in multiple more recent works.

https://harm.work/work/markovs-window

Markov's Dream from 2022

https://harm.work/nft/markovs-dream/1

Markov's Window 2023

https://harm.work/nft/markovs-dream/30

Stained Unravel (Markov's Dream) from 2026

https://harm.work/nft/stained-unravel/6

I have long wondered what attracted me to these colours the first time I picked them in 2004. Looking back - perhaps it's a stretch - I now think that the blue reminds me of a test screen, the green is very much "not printed matter" because in print green can only be constructed by mixing cyan, yellow and kali, which always makes it darker than what a screen can produce, and the magenta, on the contrary, is actually part of the CMYK gamut. The black to white gradient is so very easily rendered on a digital screen, but it needs dithering in printed matter. The palette is therefore very digital - aware of how it can and cannot be represented in printed matter using the CMYK colour space.