Looking for suggestions on how to manage excessive red cast with modded DSLR

[pdarmody]

Hi, I am pretty new to PI (2 months) and relatively new to astrophotography (3 years).  I like the extra red sensitivity of my full-spectrum modded T2i (550D), but very often there is just too much red and it overwhelms the other colors in my images.  Nearly everything is red or pink!  You can see examples of this in my Astrobin gallery (link below).  Since my T2i is full-spectrum modded, I also need to use an Astronomik CLS-CCD filter for IR/UV cut.  It really helps with light pollution also.  Regarding white balance, I just leave the camera set to Daylight. I don't do any custom WB in the camera itself.

With PI, I have found one way to fix the excessive red issue- unlink the RGB channels in STF and copy the process settings to the HT tool.  That really helps (did it for my HH-Flame nebula and Orion nebula photos on Astrobin)!  Other than this fix, I pretty much follow the PI workflows published by Jason Brown and Edoardo Luca Radice.

However, sometimes STF is a little too aggressive (I get a lot of noise) and sometimes it reduces the red too much.  I am sure that there are better approaches in PI to deal with this issue and would appreciate any suggestions.  I have a lot of older photos that I would really like improve with PI.

Thanks in advance for any ideas.  I am really enjoying PI and its easy to see how a person can get addicted to using it!    8)

[Carlos Milovic]

It doesn't matter which WB setting you use. When you shoot RAW, they are not used.
Instead of using STF with the channels unlinked, try to use the color calibration tools while on the linear stage. Do a search on the forum, and you'll find pretty good references for this.

[pdarmody]

Carlos,

Thanks very much for the quick reply. Before trying the STF workaround, I actually had been using only the Color Calibration tools in the linear stage.  I understand that is where color correction needs to be done. The result on my HH-Flame nebula photo was that the Flame was still pink -not the orange-white I typically see in better photos.  I am still doing the recommended Color Cal prior to any stretching.  What I have noticed on the forums and tutorials (so far) is that the information is oriented toward CCD cams or non-modified DSLR's.

I will keep looking through the forum to see what I need to do differently in Color Cal. 

[Carlos Milovic]

I do my own imaging with the same T2i you have (although with the astrodon filter). BackgroundNeutralization and then ColorCalibration is the way to go. It works the same way as in ccd. If properly done, you'll end with a rich gamut of colors, in the stars and the object. Usually a follow the image data, and do not try to get results similar to other people's.
Also, remember that later, on the nonlinear stage you may finetune the colors and saturation with curves and all the tools in IntensityTransforms.

[pdarmody]

Thank you again, Carlos.  It's really helpful to hear from people who use similar hardware.  I will keep working with BackgroundNeutralization and ColorCalibration.  I am really enjoying using and learning about PI!  I am very new to PI, but I can tell that your team has done a great job with this platform, both from the development and support aspects.

[cdavid]

Just a question...I find that my full spectrum T3i will also show a heavy red cast even after performing DBE and BN.  I have in these instances split the image into RGB channels and then performed a linear fit of the  B and G channels to the R channel as reference.  I then recombine the RGB channels and proceed from there.  It seems to remove the heavy red cast....

is this not a correct thing to do?

Thanks
Carlos D.

[Carlos Milovic]

You just need to use ColorCalibration, after BackgroundNeutralization.

[guest11353]

I noticed before I started processing my image yesterday that it was extremely red when the channels were linked in PI. I had used DBE, BN and CC to address it but there still a lot of red in it.
 
A tutorial on Harry's Astroshed highlights a tool called Linear Fit and it did an amazing job in fixing the red bias - see below - before and after LF

Image: https://lh4.googleusercontent.com/-2emtBJlMh2g/VNpzIu8d_SI/AAAAAAAAXK8/VJp-PrP-H_4/w1358-h448-no/02_Linear_LinearFit.PNG
 
This is just in preview using STF to do a preliminary screen stretch (channels are linked in both) - can't believe how well it worked - and it was sooo easy. I remember working for a half hour in photoshop to address this sort of thing :-)

After using LF - DBE and the rest of the processes worked great.

Harry's Astroshed tutorial: http://harrysastroshed.com/pixinsight/pixinsight%20video%20files/2013%20pix%20vids/dbe/dbe.mp4

Hope that helps

David

[cdavid]

That's exactly what I do and was describing above. But Carlos M. Says we don't need to do that. I must be using DBE and BN incorrectly because if I don't use this linear fit method the heavy red cast remains.

Carlis D.

[guest11353]

That's exactly what I do and was describing above. But Carlos M. Says we don't need to do that. I must be using DBE and BN incorrectly because if I don't use this linear fit method the heavy red cast remains.

Carlis D.

Yes I noted that, and it could very well be that I hadn't used DBE or BN correctly when I made my first attempt but...when I tried DBE after Linear Fit processing it worked perfectly and I didn't need Background Neutralisation.

Please note I'm not saying this is the way to go - I've only been using PI for a week - Linear Fit really had a huge impact on the DBE process when I used them together :-)
Just my observation.

David

[cdavid]

...agreed it works well. I am just not sure why Carlos M. is recommending we do not.

[Carlos Milovic]

LinearFit is an unnecesary extra step.
ColorCalibration applies the same operations (rescalation mostly), but instead of achieving 3 channels with the same mean value, you are maximizing the hue information. With ColorCalibration you may calibrate using the stars, or extended objects like galaxies or nebulas. It is much more powerfull. Also, at the same time, CC ensures that your background (witch could be the neutral grey, or a custom color) remains with the same hue. This does not happen with LinearFit.

[cdavid]

Thanks Carlos.
I've tended to employ linear fit with red as my reference channel after DBE and BN. I then also proceed to color calibration after the LF.

So you are saying that by doing this I may be muting some of the natural reds in my image?

I have noted that my red nebulas will not be as extensive and are somewhat muted compared to other people's images. I always attributed that to not using heavy saturation boosts. Perhaps this linear fit step I have been using is the cause?

Thanks
Carlos D.

[jkmorse]

Carlos,

How does color calibration affect something like M51.  That has lots of blue and red in the galaxy and if you use the galaxy as the white point, aren't you at risk of muting those hues?

Thanks,

Jim

[Harry page]

Hi Carlos

I teach in my vids how to do dbe - colour calibration etc in the correct order , but with some images ( mostley from dslr cameras ) this will not give

good results , due I think to the very noisy nature of these images 

The only way I could find to give good repeatable results was to use Linear fit after this which always seems to work well  8)

If you are able to show a better way I am all ears

Harry