Superbias with vertical and horizontal lines

[llavaud]

Hello all,

I have a master bias with vertical and horizontal lines, so i have used SuperBias with "columns and rows" orientation but it generate two images file.

How i must use these two images for light calibration ? Normally i must have only one master bias file.

Thanks in advance for your help.

Regards.

[Carlos Milovic]

Are you using a DSLR camera? Is it horizontal banding, instead of horizontal lines?

[llavaud]

Are you using a DSLR camera? Is it horizontal banding, instead of horizontal lines?

Yes i use a Canon 40D, and here are the images:

Master Bias:


Superbias vertical:


Superbias horizontal:

[Carlos Milovic]

How many images are integrated in you master bias? I ask, because I've seen that horizontal lines are more random. So, if you use a lot of images, they tend to cancel each other, and only the vertical lines matters.
If they are not random, I would either not use superbias, or try to use it in two steps. First, apply it to the vertical banding, to obtain "super bias 1" (let's call it sb1 for short). With PixelMath, substract sb1 to your master bias, and add a pedestal (0.5 for example... it doesn't matter as long as you remember wich one it is, and you don't get truncated values). Now, to this remanent bias signal, apply super bias again, in the horizontal direction. Use PixelMath to add this to sb1, and remember to substract the pedestal. The result is going to be your super bias, with both components.

[llavaud]

How many images are integrated in you master bias? I ask, because I've seen that horizontal lines are more random. So, if you use a lot of images, they tend to cancel each other, and only the vertical lines matters.
If they are not random, I would either not use superbias, or try to use it in two steps. First, apply it to the vertical banding, to obtain "super bias 1" (let's call it sb1 for short). With PixelMath, substract sb1 to your master bias, and add a pedestal (0.5 for example... it doesn't matter as long as you remember wich one it is, and you don't get truncated values). Now, to this remanent bias signal, apply super bias again, in the horizontal direction. Use PixelMath to add this to sb1, and remember to substract the pedestal. The result is going to be your super bias, with both components.

I have integrated 300 bias so i dont think it is some random noise

i will try you method and let you know if it is working

thanks

[MikeOates]

To be honest, if you have used 300 bias frame to make the master, just use it as that, don't use the SuperBias. You should get good results with that number of bias frames.

Mike

[llavaud]

How many images are integrated in you master bias? I ask, because I've seen that horizontal lines are more random. So, if you use a lot of images, they tend to cancel each other, and only the vertical lines matters.
If they are not random, I would either not use superbias, or try to use it in two steps. First, apply it to the vertical banding, to obtain "super bias 1" (let's call it sb1 for short). With PixelMath, substract sb1 to your master bias, and add a pedestal (0.5 for example... it doesn't matter as long as you remember wich one it is, and you don't get truncated values). Now, to this remanent bias signal, apply super bias again, in the horizontal direction. Use PixelMath to add this to sb1, and remember to substract the pedestal. The result is going to be your super bias, with both components.

here is the superbias with the two components (row + columns):



now i subtract the superbias with only columns from my master, we see that stay residual horizontal line



now i subtract the superbias with the 2 components (row + columns) from my master and it seems to be good, i see only random noise



so your method seem good,  thanks for your help !

[llavaud]

To be honest, if you have used 300 bias frame to make the master, just use it as that, don't use the SuperBias. You should get good results with that number of bias frames.

Mike

yes a 300 bias frame is quite very good but if i can get a better one, why not ?

[-Amenophis-]

FOr CCD camera (Atik One), should you use the column and line method or choose a column?

[Carlos Milovic]

Usually columns are enough. But, before using SuperBias, take a lot of bias frames, average them (+50) and see if there is a pattern, either way.

[-Amenophis-]

Thank you
I use approximately 500 BIAS frame.