The nice thing about deep space objects is that they don't change at a rate that there is any visible difference over timespans relevant to humans. This means that you can do very, very long exposures. It also means that you don't need to capture all colours at once.
The linked article talks about using a color camera with a bayer matrix that uses dyed glass to get color data and how the filters don't filter out light in the 800-1000nm range.
Lots of amateur astrophotographers don't use color camera's but monochrome cameras in combination with separate filters. These filters are of much higher quality than the dye based filters used on camera sensors. Instead of dyes they use interference filters[1]. These filters do not have the same issue as described in the article. For example the LRGB filter set I use only lets through a 100nm band for each color. Next to that you can use filter that only let through emissions from specific elements, like the HSO filters mentioned (Hydrogen-alpha, Sulphur-II, Oxygen-III). In my case I have filters with a bandwidth of 3nm around the emission line of each of these elements.
The bigger issue with making these photo's look like what they 'really look like' is that they really look like nothing at all. The light from these objects is so faint that even with a highly sensitive cooled camera you need many hours of exposure to grab enough data to make a nice picture. What they 'really look like' to the human eye is black.
The issue is worse when you add in the narrowband data, because that overlaps with the visible spectrum, but in that case you usually emphasise certain bands to stand out more in the final picture. It's not uncommon to combine the LRGB data with e.g. H-alpha data.
In the end the colours you choose are up to the photographer and their goals. Are you just going for a pretty picture or are you trying to visualize scientific data?
> It's not uncommon to combine the LRGB data with e.g. H-alpha data.
Yeah I do this too. Typically I do HOO to maintain as close to spectral accuracy as possible. I don't have an S filter.
It is, by the way, possible to get DSOs shot as a single image, it's tricky. See the 3rd image in this series:
https://www.instagram.com/p/CKXdL1YndQi/?img_index=3
I want to try to fine tune a diffusion model to make it possible to actually shoot these single shot with good fidelity.
Unfortunately I'm in a Bortle 5 to 6 area so I'd need quite a bit of integration time to get a decent SnR. There aren't really any proper dark sites in western Europe.
Yeah that would do it. I'm in the western US and there are lots of Bortle 2 sites I can access in a weekend, and Bortle 1 sites I can access on a long weekend.
There's even a Bortle 3 site I can get to after work just 2 hours from San Jose (Pinnacles National Park) which is exactly where I took that seagull nebula picture.