Dark scaling question

[Luigi]

When doing calibration for flat frames (for example), how does PI 'know' how to scale darks?

For example, I had a master bias and a master dark of 30 seconds that I am using to calibrate flat frames of differing exposure times. If I have a flat exposure, of say, 8 seconds, how does PI 'know' that the master dark is 30 seconds duration, since that information is not in the FITS header?

Should I even 'bother' with darks for flats? Should I just use master bias for calibrating  flats?

[vicent_peris]

Hi Louis,

The dark scaling factor is found in IC by purely numerical methods, so it doesn't need to know the exposure duration of your dark and flat files. This factor is found by looking for a minimum noise in the resulting calibrated image.

One advice: take as much bias and dark frames as you can. This method tends always to undercorrect the dark noise because the master dark has also some bias noise. The resulting calibrated image is always optimal in terms of noise intensity, but it may have some residual hot pixels that represent a very low percentage of the total pixel count.


Regards,
Vicent.

[Luigi]

Vicent,

Thanks for the explanation.

Of late, I have been using 16 bias and 16 darks to create my masters. I combine them according to the tutorial you wrote (thanks for writing that, by the way)

[ruediger]

The dark scaling factor is found in IC by purely numerical methods, so it doesn't need to know the exposure duration of your dark and flat files. This factor is found by looking for a minimum noise in the resulting calibrated image.

Hello Vicent,

for DSLR users, are there any experiences which dark file scales best when one or more parameters of "ISO", "exposure time" and "temperature" do not match exactly the light frames?

Let's say, I have a pretty large collection of dark frames, but no dark matches the light frames exactly. Should I prefer darks with same temperature or same exposure time?

Is it even worth the time to take darks with exactly the same exposure time as the light frames? Instead of, e.g., taking 10x6min to exactly match the light frames, I could also take 20x3min and scale them, which set would perform better?

Regards,
 Rüdiger

[pfile]

well i'm not vincent but this is my understanding:

darks should match the lights - in order of importance: ISO, temperature, time.

what i hear is that dark current is a linear function of exposure time but a nonlinear function of temperature. the IC dark scaling is linear only, so if your darks are going to differ from your lights, a difference in exposure time is best.

i would not stress too much if your master dark average temperature is within 3-5C of your light temperature though.

the problem with DSLRs and dark current is that the image processing software in the camera does a lot of tricks to reduce noise that can not be turned off. therefore it's hard to say if any two images taken under similar conditions are going to have similar noise characteristics.