Page 1 of 1

"Partially" mixing RGB and HA

Posted: Mon Nov 13, 2017 7:50 pm
by micheleorsini
Now that I have a beautiful M31 image I'm proud of (see gallery), I was wondering if it was the case of trying to get HA frames too

This opens a problem that has been lying in the back of my head for a while, I want to share my thoughts with you: the most common technique to mix Ha and RGB data is to use the former as luminance and the latter as colors. The underlying assumption is that, in the whole image, the most significative contribution to luminance comes from ionized hydrogen.

Well this cannot always be the true, and the first "meh" came when I was preparing my Veil HaRGB image: I'm quite sure there were blue/green regions which cannot be represented in luminance by HA data.

M31 is one of other such cases: one wants to enhance HA regions and so takes both RGB and Ha and ... then?

there should be a way to use HA luminance to enhance just some parts of the RGB but I cannot figure out how.

Is it possible to use layer module somehow to achieve this? A sort of mixing relative values is needed: given a point in the image, use as source the image that has the the maximum relative luminance value (it is not possible to compare absolute values, isn't it? different ISO, integration time, different nights ...)

I spent some time in other forums and I found this guy https://www.cloudynights.com/topic/549498-m31-hargb-cc/ made a great job (see his reply at 09 September 2016 - 03:20 AM).

It seems to me that he used a mask made with Ha - R (absolute values? relative values?) so that mask is active only when HA signal is greater than Red signal, then he made something similar in L (again: how it is possible to compare values?)

I suspect that something similar can be done using Layer module but I have no idea about how to perform it, anyone else?

Michele

Re: "Partially" mixing RGB and HA

Posted: Tue Nov 14, 2017 10:39 am
by Guy
Hi Michele,

Here are a few ideas you might want to consider:
  1. Making a synthetic luminance frame using the Ha and RGB - see 'Creating a synthetic luminance frame' in the Special Techniques section of the Layer Module Use topic. Using this we know we are using all the luminance - not just where there is Ha. We will need the total exposure time for each frame, and the ISO used (if different), to weight the images.
  2. Create a difference mask which masks out where RGB<Ha. I think this can be done in the Layer Module. Try using the Brightness Mask Mode, Synth Lum in background, RGB in foreground, 'Where fg is light use Bg'. Areas over a certain luminosity will be considered active in the mask - you may need to experiment with Brightness Mask Power. Save the mask. Use this mask in the Layer Module to let through the RGB unchanged - see later.
  3. Use the Mask Module Auto mode to create a mask from the RGB - setting the Selection mode of 'Highlights>Threshold' or 'Light Feature + Highlights>Threshold' and experiment with Threshold values & Mask Fuzz. Use this mask in the Layer Module to let through the RGB. Finding the right threshold will take experimentation.
  4. Use the Ha with a mask to create an Ha frame with just the elements you want. Use the Layer Module, Set the Mask you made earlier, Synth Lum in background, RGB in foreground. Layer Mode 'Blend'. The mask should (I think!) let through the RGB unchanged and the Ha blended with the RGB around it. Adjust Blend Amount as needed to change the relative strength of the Ha.
  5. Or, to use a completely different approach, use the Synthetic Luminance frame combined with the RGB frame to create a Synthetic Luminance frame which is limited by the luminosity of the RGB frame. Try using the Brightness Mask Mode, Synth Lum in background, RGB in foreground, 'Where fg is light use Bg'. Layer Mode 'Blend'. Experiment with Blend Amount and Brightness Mask Power. Save the result an use it as a Synth Lum with the RGB providing the colour. See 'Combining luminance and RGB images using the Layer Module'
  6. In the Layer module, use the Ha in the background and the RGB in the foreground. Try using the Brightness Mask Mode, Ha in background, RGB in foreground, 'Where fg is dark use Bg'. Layer Mode 'Blend'. Experiment with Blend Amount and Brightness Mask Power. Increasing the Brightness Mask Power increases the dark threshold so more of the foreground shows.
All this is 'off the top of my head' - I haven't tried it - so I may have missed a step/got something wrong. If you want any help I'd be happy to - just upload the data.

Others may have more experience of doing this sort of thing - I'd be interested to see how others deal with this.

Regards,

Guy

Re: "Partially" mixing RGB and HA

Posted: Wed Nov 15, 2017 6:00 am
by micheleorsini
wow, a wirlwind of ideas! thank you!!

alright, I will upload the data, for sure ... once taken! currently here is snowing but I have a chance (last for M31, I think) to take a few shots on M31 in the next days

in the meantime, I'm going to read everything carefully and make a few attempts with Veil data

Veil HA: https://www.dropbox.com/s/t958yzn6g0t6e ... A.fts?dl=0
Veil HA processed with Startools: https://www.dropbox.com/s/uiw1yyfd8q9wr ... .tiff?dl=0
Veil RGB: https://www.dropbox.com/s/3ijc5uphsr4ql ... l.fts?dl=0
Veil RGB processed with Startools (and corrected for elongated stars): https://www.dropbox.com/s/6wsbsgbbdmw7w ... .tiff?dl=0

Michele

Re: "Partially" mixing RGB and HA

Posted: Fri Nov 24, 2017 8:21 am
by Guy
Hi Michele,
I've just had a quick look at this again to see the best way of doing this . From the FITS data It looks the total time for the RGB was 4hrs and the Ha was 5 hrs. Is that right? Also, what were the ISO settings used?

Guy

Re: "Partially" mixing RGB and HA

Posted: Fri Nov 24, 2017 7:40 pm
by micheleorsini
Hi Guy,
first I apologise for having done *nothing* in the last 10 days, also, no new frames unfortunately, here humidity has been above 80-90% all days, it is impossible even to have a decent guiding, I made a few attempts and all ended with wet lenses.

Then:

HA 5 hours at 1600ISO
RGB 4 hours at 400ISO

Michele

Re: "Partially" mixing RGB and HA

Posted: Fri Nov 24, 2017 11:51 pm
by Rkonrad
Hi Guy,

From what I"ve learned so far about astrophotography is that iso has no bearing on the actual exposure of the image. The same number of photons are collected no matter what the iso is. I've been told that a higher iso doesn't increase the sensitivity of the sensor (unlike film's asa). It simply amplifies the signal (which processing puts back to normal). Higher iso's have less read noise (though dslrs are pretty good that way) and lower ones increase dynamic range. I hope I'm understanding correctly what you were trying to convey.

I mention this as I noticed you wanted MIchele's exposure times and iso's.
We will need the total exposure time for each frame, and the ISO used (if different), to weight the images.
Here is a good article that explains what I'm trying to say better http://dslr-astrophotography.com/iso-ds ... otography/. There's also a chart that shows the optimal "read noise" to "dynamic range" trade off for different cameras. My canon 600d is 800-1600. I believe some nikons are actually 100.

Cheers Richard

Re: "Partially" mixing RGB and HA

Posted: Sat Nov 25, 2017 8:02 am
by Guy
Hi,

Michele, thanks - don't worry - I've only just found the time to have another quick look. I hope things become a bit drier for you soon.

Richard, you are right regarding the ISO setting in relation to choosing an optimum setting to use for photography.

I was asking because, as you say, it affects the amplification applied to the signal (and its dynamic range). I was thinking of what weighting to apply to each (unprocessed) image if combining into a synthetic luminance frame. If the ISOs were the same then you usually combine based on exposure time - however that is not the case here.

If anyone has an approach to weighting in these cases I'd be interested to hear about it.

Guy

Re: "Partially" mixing RGB and HA

Posted: Sun Nov 26, 2017 7:24 am
by micheleorsini
Hi!

Guy, I like your approach, I'm thinking at the same solution!

Richard: I'm shooting at 400ISO even if it is not optimal for my canon so that I can take 4 minutes long frames, there is too much light pollution for shootint at higher ISOs and I'm giving priority to long exposures to collect more photons. By the way, the strong light pollution, also, becomes the strongest source of noise. Of course I could use higher ISO in HA imaging.

Michele