Why Do My Captured OSC Frames Appear to Be Color Balanced?

Discussion of using SharpCap for Deep Sky Imaging
Post Reply
Zubenelgenubi
Posts: 18
Joined: Thu May 24, 2018 5:26 pm
Location: Texas USA

Why Do My Captured OSC Frames Appear to Be Color Balanced?

#1

Post by Zubenelgenubi »

Flat Capture.jpg
Flat Capture.jpg (98.81 KiB) Viewed 678 times
I am attempting to do narrowband imaging with my OSC camera and I am seeing something that I find unusual when I capture flat frames in SharpCap. During the flat capture, the displayed histogram looks as I expect. The histogram peaks fall at the levels predicted by the camera's Bayer Filter response (i.e. when I use an Ha filter, the red channels is strong and the blue and green channels are weaker in proportion to the Bayer Filter gain.)

Image

However when I view the deBayered flat frame in SharpCap or other post processing tools, the histogram peaks are aligned as if a color balance process was performed. At first I thought that the color balance was for the displayed image only but when I look at the individual CFA channels before any deBayering process, those channels seem to be balanced as well. Is this what I should expect? It is my understanding that the master flat is applied to light frames before the deBayer process. It doesn't seem correct to color balance the flat frames. What am I not understanding?

Image
Attachments
Flat Display.jpg
Flat Display.jpg (81.15 KiB) Viewed 678 times
User avatar
admin
Site Admin
Posts: 13179
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Why Do My Captured OSC Frames Appear to Be Color Balanced?

#2

Post by admin »

Hi,

when you use the 'capture flat' functionality in SharpCap with a colour camera, you have the option of making a monochrome flat frame - see https://docs.sharpcap.co.uk/4.0/#Creati ... t%20Frames.

If you make a monochrome flat frame then SharpCap takes the individual flat frames, averages them, debayers and then converts the colour image to monochrome and saves as a monochrome image. If you use this flat image then it will not affect the colour balance of the captured images but the flat itself is a monochrome image (if you load it as RAW then the colour peaks will be aligned).

If you choose *not* to make a monochrome flate, then SharpCap captures and averages the individual flat frames, then saves the average. This flat frame will have a colour cast to it if you load it as RAW. However if you *use* this flat frame, it will have the effect of colour balancing your captured images, which may not be desireable (if the blue channel is weak in the flat then the blue pixels will be boosted more by the flat correction, effectively colour balancing the captured images).

Hope that helps explain what is going on.

cheers,

Robin
Zubenelgenubi
Posts: 18
Joined: Thu May 24, 2018 5:26 pm
Location: Texas USA

Re: Why Do My Captured OSC Frames Appear to Be Color Balanced?

#3

Post by Zubenelgenubi »

Thank you.
Zubenelgenubi
Posts: 18
Joined: Thu May 24, 2018 5:26 pm
Location: Texas USA

Re: Why Do My Captured OSC Frames Appear to Be Color Balanced?

#4

Post by Zubenelgenubi »

Robin,

I understand your explanation as to why it makes sense to color balance the color histograms in the master flat frame. However, in the case of my flats captured with my Ha filter, the green and blue histogram peaks are at a much lower intensity than the red peak. Doesn't aligning the green and blue peaks with the red increase the noise level in the master flat?
User avatar
admin
Site Admin
Posts: 13179
Joined: Sat Feb 11, 2017 3:52 pm
Location: Vale of the White Horse, UK
Contact:

Re: Why Do My Captured OSC Frames Appear to Be Color Balanced?

#5

Post by admin »

Hi,

that's an interesting question, and there are lots of things involved that might make the answer complicated...

For instance, converting the RAW flat frame to monochrome will very likely reduce the noise level - that's because the debayering process uses averaged values (bilinear interpolation) to fill in the missing colour values when creating RGB. Also, converting RGB to monochrome is averaging the 3 colour channels, so again a noise reduction due to averaging.

Now thinking about the case where the flat is kept in RGB, you are right that the less bright channels will have more noise - in a well exposed flat you would expect the shot noise to dominate, so the noise level in a channel that is 4 times dimmer will be half that in the brighter channel, meaning that the dimmer channel has a SNR that is 2x lower. If you brighten that dim channel up by a factor of 4 then you end up with more noise than the channel that was originally bright, but I don't think that really matters, because the SNR of the brightened channel is still the same, and it's the SNR that matters when it comes to thinking about how much noise ends up in the final image.

Overall, I think the mono flat would have less noise than the RAW flat (introduce less noise into the final image), but I haven't done the full maths to prove it. If your green and blue channels are *really* dim, then it might be best to make a flat based only on the red channel - ie capture in RAW, debayer to RGB, throw away the green and blue channels and convert the red to mono. You'd have to do that in an external processing application though, as SharpCap doesn't have that functionality built in.

cheers,

Robin
Post Reply