Colour Balance and LRGB images

[Terry B]

Dear All
I am just starting to do some colour images with my CCD.
I am a bit confused with the images that i have seen posted here and elsewhere.
I have been trying to balance the colours. I have taken some imaged of a G2 star and using AIP for windows have found the colour weights for my filters and camera to be:
red 1.0
Green 0.7
blue 0.3
This means that to get even colours without multiplying the blue and green, their exposures need to be longer. I have compromised and just used a ratio of 1:1:2 respectively and just divided the green by 0.7 and the blue by 0.6.
I have noted that a lot of images use equal exposures for each channel or sometimes reduced exposures for the blue channel. So what gives?
Are the modern camera much more sensitive to blue than mine?
I run a KAF0401E which is suppose to be more sensitive in the blue than the original 0401 but it is still no where as sensitive as the red.
Thanks in advance.

[jase (Jason)]

May I ask what RGB filters you're using?

Generally, the ratios are calculated from the red filter. So if the R is 1, you then calculate the offset of G and B based on filter transmission and chip spectral sensitivity by imaging a G2V star - which you've done.

You can either perform the colour balance process in two ways - by exposure lengths or colour weights. The later is performed in when processing/combining the RGB image. Don't stretch the individual R, G, B exposures as this will through the weights out. Using colour weights in post processing allows you to take similar exposure lengths through each filter.

There are filters available that allow equal R, G, B exposures such as those manufactured by Astrodon. Equal exposure times are achieved through wavelength cross talk. In other words there is significant overlap between the B - G and G - R wavelengths. This simplifies acquiring colour data, but has a downside. Due to the cross talk, there is a lack of contrast between each individual exposure. As the red filter also passes through some of the green wavelength it is not ideal to use the R exposure as a luminance channel in an [R]RGB image. Also the wavelength cross talk lets in unwanted light pollution of sodium vapour street lights between the 570nm and 610nm (between the G - R wavelengths). This is the reason many astrophotographers don't do RGB images from heavily light polluted areas. Traditional RGB filters are considered broadband and do let in a lot of light that in some cases is unwanted. Narrowband imaging is far more appealing in these conditions.

[jase (Jason)]

Sorry, also forgot to mention that your colours will shift with L or Ha RGB images (particularly with Ha). In most cases these will wash out the colour saturation which is easily resolved in PS. Depending on the object it is sometimes easier to replace the R channel with the a blended R and L or Ha image.

There is a good article written by Neil Fleming on the topic - http://www.flemingastrophotography.com/lrgbshift.html

[Terry B]

Originally Posted by jase

May I ask what RGB filters you're using?

Generally, the ratios are calculated from the red filter. So if the R is 1, you then calculate the offset of G and B based on filter transmission and chip spectral sensitivity by imaging a G2V star - which you've done.

You can either perform the colour balance process in two ways - by exposure lengths or colour weights. The later is performed in when processing/combining the RGB image. Don't stretch the individual R, G, B exposures as this will through the weights out. Using colour weights in post processing allows you to take similar exposure lengths through each filter.

There are filters available that allow equal R, G, B exposures such as those manufactured by Astrodon. Equal exposure times are achieved through wavelength cross talk. In other words there is significant overlap between the B - G and G - R wavelengths. This simplifies acquiring colour data, but has a downside. Due to the cross talk, there is a lack of contrast between each individual exposure. As the red filter also passes through some of the green wavelength it is not ideal to use the R exposure as a luminance channel in an [R]RGB image. Also the wavelength cross talk lets in unwanted light pollution of sodium vapour street lights between the 570nm and 610nm (between the G - R wavelengths). This is the reason many astrophotographers don't do RGB images from heavily light polluted areas. Traditional RGB filters are considered broadband and do let in a lot of light that in some cases is unwanted. Narrowband imaging is far more appealing in these conditions.

My filter set is Astronomic L-RGB typ 2c
http://www.astronomik.com/english/eng_rgbt2.html
I think these filters would be considered "narrow" band.
I image from a pretty dark rural location 10km west of Armidale at 1100m altitude so light polution is minimal.
I balanced my images using the weights from the G2 star and didn't stretch the original images.
Should I bother taking a longer exposure in the blue or just let the software balance the images?
Thanks for the info.

[jase (Jason)]

Your 1:1:2 ratio is pretty close for the KAF0401E spectral response and the Astronomik filters. Though, doubling B compared to that of R seems a little excessive. I would suggest B be around 1.5

Personally, I would balance when combining. However having stated this, it is important to note that post processing colour balance isn't a substitute for lack of data acquisition. You should still take plenty of R, G, B exposures to improve your signal to noise ratio. Try to at least keep around the ratios for exposures, but accuracy is not needed when post processing colour balance is used.

The Astronomik LRGB type IIc filters are still considered broadband. By narrowband filters I mean specific emission line filters;
http://www.astronomik.com/english/eng_halpha6nm.html
http://www.astronomik.com/english/eng_oiii-ccd.html
http://www.astronomik.com/english/eng_sii-ccd.html