Analysing flats, darks etc..

[Chefgage]

What's the best way to analyse the darks, flats etc.. I have taken. Obviously I can just stack them all with the lights in dss and process in pixinsight as normal and see what the resultant image looks like. But is there a way to analyse the flat frame to see if it has been taken correctly. Same with the dark frames?

Edited January 28, 2022 by Chefgage

[Adam1234]

One thing you could do with the flats is to use image statistics in pixinsight to check that the flat is exposed well and you have not black clipped or white clipped any data (i.e. no pixels with value 0 or 1).

[Chefgage]

1 hour ago, Adam1234 said:

One thing you could do with the flats is to use image statistics in pixinsight to check that the flat is exposed well and you have not black clipped or white clipped any data (i.e. no pixels with value 0 or 1).

Thanks. I will have a look at that tomorrow.

[The Lazy Astronomer]

5 hours ago, Chefgage said:

What's the best way to analyse the darks, flats etc.. I have taken. Obviously I can just stack them all with the lights in dss and process in pixinsight as normal and see what the resultant image looks like. But is there a way to analyse the flat frame to see if it has been taken correctly. Same with the dark frames?

I'm of the belief that the only meaningful test of calibration frames is to use them to calibrate lights (any clear defects like clipped pixels notwithstanding) Stretch the calibrated stack to oblivion and see if you can see defects. 

[Chefgage]

12 hours ago, The Lazy Astronomer said:

I'm of the belief that the only meaningful test of calibration frames is to use them to calibrate lights (any clear defects like clipped pixels notwithstanding) Stretch the calibrated stack to oblivion and see if you can see defects. 

What defects should I be on the look out for?

[The Lazy Astronomer]

1 hour ago, Chefgage said:

What defects should I be on the look out for?

Clipped pixels (value of 0 or 1), as mentioned above, and ideally a histogram peak that sits in the linear response range of the sensor - generally speaking, somewhere in 30 - 60% range. 

[vlaiv]

Testing out flats:

- take two set of flats with different settings - calibrate them one against the other and observe result - it should be perfectly flat uniform noise (no variation in brightness). Don't forget flat darks for each master flat.

Testing out bias:

Examine average ADU in sequence - they should be fairly uniform (very small differences on second or third decimal place)

Take set of bias, power down camera and computer, do another set of bias. Stack them using simple average each to their own stack and subtract two stacks - you should get perfectly uniform noise with mean ADU value of 0.

Darks - same as with flats.

With darks there is one more test - mean ADU value should raise linearly with time for given set point temperature

You can see if you have proper bias by taking set of bias, set of darks at one exposure and set of darks on double that exposure.

Then you stack each and produce

(long darks - bias) / 2 - (short darks - bias)

This frame should again be uniform noise with mean ADU value of 0.

 

[The Lazy Astronomer]

1 hour ago, Chefgage said:

What defects should I be on the look out for?

Oh sorry, I just realised what you meant! By defects, l mean can you see any residual dust mote shadows, or vignetting that hasn't been fully corrected, or a brightening of the edges caused by over correction (inverse vignetting).