OK lets challenge Darktable 3.6 with Aurora Borealis

Well done to get the foliage just visible, but the dark blue sky is not right. It was moonless at midnight and the only illumination would be possible sparkles of aurora most likely greenish shaded.

IMG_4091 IMG_4091.CR2.xmp (13.5 KB)

That looks pretty good, but a couple of questions.

I notice you didn’t use the color calibration module, but just set the white balance module to “as shot” - is this just personal preference? In practice it didnt make much difference for this image, but color calibration gives a temperature 400K lower.

Regarding filmic - it looks like you chose White Relative exposure of 4EV, was that a personal choice? as pushing up to 6EV does reveal more of the fainter aurora

And lastly, I like the way you used the tone equalizer to darken the background night sky, and increase the contrast.

Just a word of caution regarding the various edits:

One should not change the colours of the aurora since the various colours are indicative of specific physical processes with atoms and ions.

Hermann-Josef

2 Likes

Although it looks like there is quite a range of colour frequencies that could result from each specific ion interaction.

Nice shoot and edits!

A technical note on the darktable 3.6 capabilities. The overexposed area in the lower-left could probably be saved to a useable degree with some better highlight reconstruction tools. It is a smooth gradient and only the green color channel is clipping. A possible caveat is if there is if the ionized color is not picked up at all by either the blue or red pixels and that they, therefore, don’t correlate with the green channel.

Yes I tried to rescue that green blob, but I think its pure green aurora, so not much to work with.

Looking at @Jade_NL 's edit I think you could clone or heal it out
rather than try to massage the highlights?? just a thought

No cloning or healing was done during the massacre of this RAW file :rofl:

Seriously though; This was an amazingly easy edit to do in RawTherapee:

  • Exposure (tad of negative exposure, bit of highlight compression, slight film-like curve),
  • Shadows/Highlight (highlights only),
  • L*a*b Adjustments (L channel, very slight negative curve),
  • White balance (just a tad warmer),
  • Softlight (default setting)

I did go wild with the noise reduction: I separately adjusted the blacks (darks), greens and purples. All done with Local Adjustments and 3 times the Blur/Grain & Denoise tool. Why? Because I could


1 Like

Hi Aliks,

To be quite honest, I don’t really dig too deeply into the modules; I mostly just fiddle with the sliders till I see something I like. To answer your questions:

Colour Calibration: The reason I didn’t use the colour calibration module is because I don’t really understand it! I didn’t change the white balance because the colours looked appropriate to me based on past viewings of northern lights.

Filmic: Definitely personal choice. I found that when more of the fainter aurora was visible, it reduced the dramatic contrast of the image. But I just popped back in to check out +6 and, well, it looks good too!

Glad you liked my edit!
Have a good one,
Matthew

hmm seen them several times here in Finland - but how cool or warm they look varies :slight_smile:
IMG_4091.CR2.xmp (16.1 KB)
an older one of mine - not great, just from back garden but an idea of the greens


Sorry, but I saw this photo and couldn’t help but play with it, using the current version of Darktable. ( 4.5.0~git773.e612b821-1+10669.1 ).

My version


IMG_4091.CR2.xmp (14.1 KB)

3 Likes

Nice but I think you lost the pink/purple on the right hand side lower edge.

Try this one which was taken 40 seconds later.
IMG_4094.CR2 (31.1 MB)

copyright licence CC-0
And here is my take, based on latest build and advice given on Aurora Photography, Clarkvision.com:

As you can see, it was quite a night for aurora!!

1 Like

This is my take of the first aurora using latest DT version and clarkvision advice on settings:

and my XMP:
IMG_4091.CR2.xmp (8.5 KB)

1 Like


IMG_4091.CR2.xmp (11.8 KB)