Image Calibration Workflow

[pfile]

vicent, if you are still reading this -

i have read that a 2nd reason for using flats is to correct for differences in QE from pixel to pixel. now, i have no idea if this is relevant on a DSLR given how much 'cooking' is done on even the "raw" files.

so yes, the vignetting will be the same at ISO100, but how about this QE normalization? any idea if this really applies to a DSLR? if so, it seems that you'd want to use the same ISO for your flats and lights.

in any case it seems the effect would probably be pretty minor, even for an astronomical CCD...?

[vicent_peris]

Hi,

The same situation happens on any electronic image sensor. In DSLR cameras, this is completely independent from the ISO "speed". This ISO is simply an amplifier, and this hasn't any to do with the photon sensitivity.


Best regards,
Vicent.

[pfile]

agreed, so just so i understand, you are saying that once the flat has been normalized it really does not matter what ISO it was captured at?

[vicent_peris]

agreed, so just so i understand, you are saying that once the flat has been normalized it really does not matter what ISO it was captured at?

Yes, it does not matter. But take into account that your flat frames will have better SNR at lower ISOs because raising ISO is a kind of limiting the electron well depth. Also remember that your flat frames will need bias at the same ISO value.

What do you mean by normalizing?


Regards,
Vicent.

[pfile]

right, yeah, usually i just calibrate the master flat with a master bias if i shoot it at ISO100. although sometimes the exposures are long enough at ISO100 to warrant a dark.

by normalization i guess i meant the multiplication by the mean of the master flat in juan's equation here:

Ical = (I - B - OptimizeDark( D, I, B )) * Mean( F ) / Max( F, tiny )

not really the right word i suppose.

[vicent_peris]

by normalization i guess i meant the multiplication by the mean of the master flat in juan's equation here:

Ical = (I - B - OptimizeDark( D, I, B )) * Mean( F ) / Max( F, tiny )

That's correct.


V.