SPCC and ColorSaturation - why does changing color saturation of an image not get changed by SPCC?

A

aschiano1

Active member
I've gotten SPCC to work nicely. But I'm trying to understand some behavior I have seen from using it that doesn't make sense to me. Specifically, if I take an image and use SPCC on it - especially if it is really badly color calibrated - and it's now 'SPCC calibrated' and then use the ColorCalibration process to increase it's saturation (upping the saturation input in the process control box) a second use of SPCC doesn't change the image. Huh? It's as if varying color saturation levels of an image don't change the color indices (R/G and B/G). That doesn't make sense to me.

Try it. Take an image with bad color calibration and use SPCC and 'fix' it. Then use ColorSaturation and change the saturation of the image so you can see it visibly in PI. The run SPCC on the new image. No real change. The output graph looks about the same from the first and second SPCC usage. Why? Changing saturation in an image changes R/G and B/G. No?

What am I missing?
 
I think I just figured it out before anyone here responded :) But it would be good to have a confirmation :)

Indeed when you change the saturation of an image with ColorCalibration - say to increase it - you change the 'color indices' - R/G and B/G for each star or object on the image. But SPCC looks at many stars on the image and compares them with the catalog stars on the image. It creates the white balance data shown on the graph it outputs if you want to see it. Then it does a linear regression to compute a linear function of image indices versus catalog indices. The fit line on the graph shows how it will adjust the white color balance of the whole image.

Well if you increase the saturation of the image ALL the stars spread out vertically on either the R/G or B/G indices on the image only, not the catalog stars, of course! A linear fit gets worse in fit with it being fatter vertically. But in the end the slope and central point determine the changes to image that SPCC makes. So bottom line changing saturation in an image DOESN'T affect the SPCC color calibration. So both image are spectrophotometrically accurate across the average of all stars and objects. But it does mess things (change the color indices) up for any one star :)

Look at the graphs before and after I changed the saturation higher to follow this convoluted story. In total, if you mess with color saturation the overall average color of your image won't change to SPCC but any one object, like a star, will change.
 

Attachments

  • spcc graph on raw.pdf
    58.4 KB · Views: 60
  • spcc graph after sat1.pdf
    57.6 KB · Views: 50
The initial images were not stretched. Yes you are right, I mistyped meaning to say 'ColorSaturation' process.
 
After application of color saturation, the image is NOT linear any more.I don't know why you are claiming that.

A linear stretch would be one of the type:
I' = a * I + b

How can such an operation saturate color?

Bernd
 
Yeah, I use SPCC to get the colors "right" often after using NB Assistant. Then use the Curves to increase saturation. Think sat increase better in linear or too much noise...

I am on an Odyssey to improve star colors. One thing I have noticed is that even with SPCC my stars have too much green in them, which I think is attributable to my filters not really being balanced, Green may let in too many photons...

www.cloudynights.com

LRGB exposure ratios, SNR balance and color balance for QHY 600m - Experienced Deep Sky Imaging - Cloudy Nights

LRGB exposure ratios, SNR balance and color balance for QHY 600m - posted in Experienced Deep Sky Imaging: In another thread on CN https://www.cloudyni...x492-in-qhy294/, the quantum efficiency (QE) of the IMX455 sensor on the QHY600m was discussed. CN member Norup did some direct...
www.cloudynights.com www.cloudynights.com

But, you would think SPCC would fix this....

You can verify the RGB ratios by holding the cursor over the star and left clicking.
 
That’s the fun part. It doesn’t change it! I guess that’s the reason to not use it on a stretched image.
 
I retract my statement. SPCC on a newly saturated image that was 'SPCCed' before DOES SUBTLY change the image. The white balance factors do not stay close to 1. It's trying to fix the saturated change back to match the reference stars but it makes a small change that not easily perceptible to my eye.

Bottom line , as they suggest, use only on a linear image before you stretch (including changing the saturation).
 
You must log in or register to reply here.
Back
Top