Page 1 of 2

Colour in 1.6

Posted: Sun Oct 06, 2019 6:24 pm
by happy-kat
Hi, love the development in v1.6 thank you but I am not being successful.
Sharing my experience with using the new function in DSS and StarTools as I am having trouble with colour and have been unable to create a useable image.
This issue has been with my camera, I did a stack using same settings and a Nikon 5100 with the camera selected in colour module and the output was great.
Canon 1100d raws (plus calibration files) stacked using the StarTools preffered DSS options and selected:
No white balance processing
Bilinear interpolation
Bin 50% / crop / develop / colour and selected the Canon 1100d
The attached image is a composite to show I tried different options (I tried loads more), the two on the right are without DSS of No white balance processing.
colour.jpg
colour.jpg (298.41 KiB) Viewed 8709 times

Re: Colour in 1.6

Posted: Mon Oct 07, 2019 2:31 am
by admin
Hmmm...
I got a 1100D here to test with. Let me look into this. Thank you for reporting this!

Re: Colour in 1.6

Posted: Mon Oct 07, 2019 7:12 am
by admin
I've tested with some 1100D test images and they come out ok, so the issue must be somewhere else...
Would you be able to share the stack with me? Thank you!

Re: Colour in 1.6

Posted: Mon Oct 07, 2019 6:19 pm
by happy-kat
Hi Ivo, yes of course and thank you, but my tomorrow now.

Re: Colour in 1.6

Posted: Wed Oct 09, 2019 4:05 am
by admin
Thank you for sharing your dataset with me!
I think the issue may be that in your workflow, you did not run Wipe. Without doing so, the color ratios will remain heavily skewed due to the bias/gradient/skyglow not being removed.
As a result the Color module will recover the incorrect colors from the linear (non-wiped) data.
Does this help/make sense?

Re: Colour in 1.6

Posted: Wed Oct 09, 2019 4:16 pm
by happy-kat
Thank you Ivo for the feedback
I had not realised that wipe was a mandatory step to include in my workflow.
The image was from a pretty dark site, bortle 3, and was a heavy crop so had removed any vignetting potential, so had omitted wipe, now I am reminded not to.
I am struggling to get the same colour intensity with new stacks and also with stacks I had processed in the prior DSS version and now processing as linear (as wb not turned off as pre option) in StarTools.
It's the yellowy brown bits in the milky way that are different.
I'll have to do more digging on the original crop and dev parameters used.
Maybe first I will restack the data but using the new options.

Re: Colour in 1.6

Posted: Fri Oct 11, 2019 7:38 am
by admin
happy-kat wrote:I am struggling to get the same colour intensity with new stacks and also with stacks I had processed in the prior DSS version and now processing as linear (as wb not turned off as pre option) in StarTools.
It's the yellowy brown bits in the milky way that are different.
Let me know if there is any particular dataset you'd like me to have a look at that is problematic. Thanks!

Re: Colour in 1.6

Posted: Fri Oct 11, 2019 10:06 am
by happy-kat
It is the same data set really as you already played with. (same night just different slices of same cake)
If you are happy with the colour output then I will bow to your experience of what would appear to be an expected colour result.
I realise we also want more from the data we have and I am limited by what I can collect which does not help my dreams of what the data should reveal. Should being a very subjective word, lol.

Re: Colour in 1.6

Posted: Sat Oct 12, 2019 9:36 pm
by happy-kat
I am really struggling with this
DSS stacks with No white balance selected, have tried both Bilinear Interpolation and AHD and startools 1.6.376 and 1.6.378 and bin|crop|wipe|dev and then colour.
In the colour module I chose the matrix for the camera Canon 1100d and this is then adding quite a lot of green to the image.

Re: Colour in 1.6

Posted: Sun Oct 13, 2019 1:40 am
by admin
happy-kat wrote:I am really struggling with this
DSS stacks with No white balance selected, have tried both Bilinear Interpolation and AHD and startools 1.6.376 and 1.6.378 and bin|crop|wipe|dev and then colour.
In the colour module I chose the matrix for the camera Canon 1100d and this is then adding quite a lot of green to the image.
Though ST behaves as expected (see below), I'm struggling with that particular dataset too in terms of getting expected coloring. What does the dataset look like if you let DSS white balance the data?
Switching to a DSLR matrix in the Color module, resets the color bias increase/reduce ratios to the manufacturer-prescribed D65 white balance. E.g. the change to green is purely because the color ratio parameters are changed! :)