Reddish images from Astromodified Camera

[lalou]

Any tips on how I can balance the colors on my images from my astromodified Canon350d? I already tried balancing the colors on the histogram when I edited these photos but I still can't remove the reddish tint (specially for the Omega Centauri). 

[Elp]

What does the histogram look like? You can also use an auto colour calibration in most astro software, Siril is free.

[lalou]

4 hours ago, Elp said:

What does the histogram look like? You can also use an auto colour calibration in most astro software, Siril is free.

Here are the histograms for  each photo. I'm thinking that the red hue/tint is because of the wider field of reds in the image. Not sure if I should clip it to get the original color of the target. 

[Rallemikken]

I've imaged with modified Canon's for several years. Firstly I'd suggest you try out Siril, and it's "Photometric color calibration" tool. If that fails, use the standard color calibration tool. This will most likely give you a neutral background. For fine-tuning I use the "levels"-tool in my image manipulation software (Gimp). Choose the channel you want to reduce or increase, and move the middle slider a little bit sideways. Easy does it.

[lalou]

11 minutes ago, Rallemikken said:

I've imaged with modified Canon's for several years. Firstly I'd suggest you try out Siril, and it's "Photometric color calibration" tool. If that fails, use the standard color calibration tool. This will most likely give you a neutral background. For fine-tuning I use the "levels"-tool in my image manipulation software (Gimp). Choose the channel you want to reduce or increase, and move the middle slider a little bit sideways. Easy does it.

I'll check out Siril. So far I'm only using DSS and Photoshop. For the levels, when you're choosing a specific channel do you clip the data if its more than the other colors? I.e. for my case should I clip the red data in the levels just to have them equal?

[sharkmelley]

Those histograms do not look like the histograms of linear data.  White balancing cannot be performed on non-linear data.  It raises the question about what kind of workflow you are following.  For instance, are you raw-converting your images before stacking or do you put your raw files directly into DSS?

[Elp]

Curves will give you more control on width of peak, or use selective colour.

[lalou]

1 hour ago, sharkmelley said:

Those histograms do not look like the histograms of linear data.  White balancing cannot be performed on non-linear data.  It raises the question about what kind of workflow you are following.  For instance, are you raw-converting your images before stacking or do you put your raw files directly into DSS?

I don't raw convert my images before stacking and I just put the raw files directly into DSS. Should I raw convert them first?

[sharkmelley]

9 minutes ago, lalou said:

I don't raw convert my images before stacking and I just put the raw files directly into DSS. Should I raw convert them first?

No, don't raw convert them first.  But white balancing should be one of the first operations you do after stacking - while the data are still linear.

[Elp]

Really learn Siril, I used DSS at the start but Siril is so much better, as it also has pre processing routines such as Dynamic Background Extraction (which will transform your images) as well as Photometric Colour Calibration if you can't align the peaks manually.

[lalou]

1 hour ago, Elp said:

Curves will give you more control on width of peak, or use selective colour.

This worked wonders!! Thanks for this suggestion. I believe this is better since I think this doesn't clip out the data but just lessens it compared to adjusting it via Levels

[lalou]

23 minutes ago, sharkmelley said:

No, don't raw convert them first.  But white balancing should be one of the first operations you do after stacking - while the data are still linear.

Thanks for this reco! I'll try this out. I guess I adjusted the curves already before I did the white balancing. 

[Rallemikken]

4 hours ago, lalou said:

For the levels, when you're choosing a specific channel do you clip the data if its more than the other colors?

No clipping. See image. If I touched the left handle here, I would start clipping the black immediatly. The righthand one could tolerate some adjustment, but for this purpose I  always use the handle in the middle. Zoom in heavily on a neutral spot in the background. 200-400%. Choose one channel, and move the slider in the middle a little bit while prewiev is enabled. I rarely move more than 0.05%. There is no right or wrong way in dealing with this issue; this is how I do it.

[Elp]

You generally don't want to touch the highlights level on the right as it's your brightest parts (ie stars) and if you're not careful you'll oversaturate them and any glow and cause them to bloat. The mid is the control.

It's usually better if you're trying to reveal faint signal to work on starless images, then recombine the stars after the stretching is done. This isn't so necessary when doing galaxy processing.

Edited April 19 by Elp

[TiffsAndAstro]

On 19/04/2024 at 16:54, sharkmelley said:

No, don't raw convert them first.  But white balancing should be one of the first operations you do after stacking - while the data are still linear.

My 600d is unmodded, I leave it on auto white balance.

All my imaging is controlled by Nina. Should I be doing something about my white balance?