[faichai]

Whoah! I was just thinking "aren't these things all colourised by a human and largely fake", I was especially thinking about the starbusts and you just blew my mind. Thank you!

Do you know if the colourisation essentially just a frequency shift into visible light or is there some human artistic license involved?

[denton-scratch]

I believe this photo was taken using filters for three wavebands of IR, with each waveband presented as a different colour. That doesn't sound like "human artistic licence" to me.

But ultimately it's all just data transformed into a visible image; there's plenty of scope for artistic licence in the rendering, if that's what you want.

The striking thing for me is the two diagonal red spikes pointing in opposite directions; apparently a newly-formed giant star does this.

"Huge bipolar jets of molecular hydrogen, represented in red, dominate the image, appearing horizontally across the upper third and vertically on the right. These occur when a star first bursts through its natal envelope of cosmic dust, shooting out a pair of opposing jets into space like a newborn first stretching her arms out into the world."

I don't know how that works.

[etrautmann]

It is not a frequency shift - it’s a colormap applied to the imaging data and there’s considerable creative license in the coloration, most prominently with the selection of a colormap but also with other aspects of post processing. I’m not an astronomer, but thats my layperson understanding and I’d love to be corrected if anyone has more info.

[gliptic]

There's some "artistic license" that you could also call scientific license, because you want to pick colors that maximize contrast and highlights features you want to highlight, like different atomic elements.

[blincoln]

The "NIRCam Compass" version of the image includes the map of filter to colour that was used.[1] It's not as easy to find, but the documentation for the telescope also includes a complete list of all the filters and what they're used for.[2]

The number in the filter name is (the center wavelength in nanometers) / 10, or (the center wavelength in microns) * 100. An "N" at the end means "narrow" and "W" indicates "wide".

For this image, the team used a pretty straightforward mapping of five filters to more or less the colours of the rainbow in ascending order of filter wavelength, with the shortest wavelength being mapped to purplish-blue and the longest being mapped to orange-red.

When I first got interested in imaging outside the wavelengths we can see, like a lot of people I assumed that the most "accurate" way to represent the data would be to pick three bands in order of descending wavelength and map them to the red, green, and blue channels. That can certainly be a useful approach, but there are often better ways to process the data that make it easier to see things of interest in the result. In particular, doing a R/G/B mapping means that most of the detail in the blue channel is invisible to human eyes, because they're so much less sensitive to blue. TLDR: I think the approach used for this image strikes a nice balance between "intuitive to humans with typical colour vision" and "highlights areas of interest in the image in a useful way".

[1] https://webbtelescope.org/contents/media/images/2023/128/01H...

[2] https://jwst-docs.stsci.edu/jwst-near-infrared-camera/nircam...