Understanding how calibration frames are used

Hi Gene  @genehelsel,

Thanks, that makes me happy 😉

APP does subtract the master dark from each flat.

APP (since version 1.060) will automatically check if a suitable masterdarkflat or masterbias can be subtracted from each flat.

• If not, APP will warn you about this. The warning would mention that flat-field calibration is not possible if a masterbias or masterdarkflat is not present to calibrate the flats.
  
• If so, it will do this automatically when the flats are being integrated. So the Masterflat will be created out of calibrated flats as they should.

Also if I have dark flats for each channel, is it recommended that I also take bias frames (done just with Lum)?

Darkflats are fine ;-), adding bias frames for the calibration of the flats is not needed or recommend since the darkflats are sufficient.

If you have however darkflats with different exposure times or temperature as your flats, then adding bias will enable you to use dark scaling on the MasterDarkFlats. So in that case, adding bias is helpful to be able to scale the MasterDarkFlat(s). (Scaling requires bias ! )

Let me know if this answers your questions.

Mabula

OK so which is better to have: bias frames (super easy) or dark flats (kind of a pain to collect). For instance PI does just fine with Bias so no need for taking the dark flats.

My overall goal is to provide APP with the best data needed to create a registered, integrated image. So I need to know exactly what to "feed" APP to get to my goal. Recommendations for number of Darks, flats, and bias (or dark flats) would be super helpful.

Thanks,

Gene

Hi Gene @genehelsel,

Well to use bias or dark flats is more a matter of sensor characteristics,  versus  software characteristics ;-).

For most sensors, indeed, only bias is just fine and the difference with dark flats will be very hard to see it any at all.

It is known that for certain CMOS camera's like the ASI1600 you will get better results with dark flats versus only bias calibration of the flats.

Recommendations are very sensor specific, what sensor do you have in what camera?

regarding darks: if darks are needed ( for instance amp-glow, fixed pattern noise in the dark current, very strong dark current) make plenty, more is always better.  Only 10-20 darks can downgrade the results significantly. You want the dark to be of very high quality.

flats: in my experience make exposure of at least 0.1 seconds to prevent a majority of possible flat frame issues. And 20-40 is usually all that is required. By illuminating the flats a bit longer (by lowering ISO/gain for instance or damping the light source), they will be of higher quality and will therefore give better calibration results. Just make sure that the histogram peaks aren't close to the right side of the histogram.

bias: only needed to create 1 per gain/iso value. Shoot 200 bias or even 400 bias to create a very good MasterBias, you will be able to use it for a long time so I wouldn't  skip here.

Kind regards,

Mabula

Hi Mabula,

Thank you for the information - very helpful. I have 3 camera's 1ccd, 2CMOS

ccd:  QHY 16200 - APS-H

CMOS: QHY 128C - full frame same sensor as Nikon D600

For these sensors I like to know which would benefit from dark flat frames and which would benefit from bias frames

Thank you,

Gene

CMOS: QHY247  - APS-C same sensor as Nikon D5500

Hi Gene @genehelsel,

Perhaps you have tried already, but for the mentioned camera's I would think bias are just fine for flat calibration, especially the CCD.

I have a Nikon D610 and bias were fine.

In cases like this, if it's not fine, you will end up seeing failed Flat-field calibration on the integration files. Sometimes the problem is very small and you will not see it visibly on a single calibrated frames. Only when you start to integrate a lot of frames, the problem becomes apparent.

So if that's the case, try making some dark flats for the particular camera and redo the calibration and integration 😉 and check if the result is better.

To determine what workflow is better for a particular camera, it's best to experiment a little bit yourself, because camera's from the same model can even behave slightly differently due to manufacturing changes or even differences in the sensors supplied to a camera maker.

So don't take my word it... 😉

Mabula