QHY163
С
10
The reason is that the Color balance is a ratio of the RGB sensitivity difference. So we use a ratio to
multiply the RGB value and get it done. But if there is a bias exist. The ratio will not be correct. For
example, the G sensitivity is the double of R sensitivity.
G=2R In order to get white balance. We multiply a ratio of 2 to R
R'=2R= G so we get R=G
When a bias exist. The bias is a constant added to each pixel. So the image you see is:
R''=R + bias
G' '= G + bias = 2R+bias
Now the ratio R'': G''= (R+bias) / (2R+bias) and it is not equal to 1:2. It shows the bias will affect the true
value of the R:G. And the ratio of R:G will vary when the image light changed. It is hard to correct with a
fixed ratio.
But for DSO capture, you should keep the offset above zero and avoid the background being cut off. A
background from 1000-5000 is a good value (16bit mode) for DSO imaging.
