Correct calibration workflow for DSLR - Forums

Topic starter

Aug 25, 2020

Hello masters, I am looking for a help creating a correct calibration files for my DSLR workflow in Astro Pixel Processor. I checked 2 tutorials straight from APP developers, but I still don't understand couple of things. I shoot on my Canon 6D and Samyang 135 mm, mostly on ISO 1600 and f/2.5. I will try to explain from what I've learned and I need your help creating correct workflow for best possible calibration and I hope this won't be confusing:

In APP, we work with: Lights, Darks, Flats, Bias, Bad pixel map (BPM)Every session I shoot: Lights, Darks
And I have set of universal master: Flats, Bias, BPM
To create BPM, which should last for a year or so, they say "insertflats and darks then click create bad pixel map". but what kind of darks am I supposed to use, if they vary on exposure and temperature? Do they mean Dark Flats?

I shoot Bias at ISO matching my exposure of lights, maximum shutter speed, without lens (only camera cap)
I shoot flats at ISO 100, AV mode with aperture of my lights. Flats should be corrected with Dark Flats and Bias (ISO 100 BIAS), right?
I shoot lights and darks at 60" to 120", ISO 1600, f/2.5

The point: So should I create 2 sets of BIAS (ISO 100 for my flats and ISO 1600 for my lights), Set of Flats and Dark Flats (to create BPM as well) and then:

1. Calibrate Flats (ISO 100) with Dark Flats and BIAS ISO 100 to create Master Flat, then...
2. Calibrate Flats and Darks to crete Bad Pixel Map, then...
3. Calibrate Lights (ISO 1600) with Darks (ISO 1600), Master Flat (ISO 100), BIAS (ISO 1600) and BPM?, then...
continue with registration, normalization and so on

Please, is that correct? My current setup works in most photos which contains nebula, but when I shoot region, which contains only sars and few DSOs (such as M31 region, Virgo cluster or similar), my calibration files doesn't work and whole image is bad.

Thank you!

Helpful Engaging

Daniel Erickson

Aug 28, 2020

I hope someone here can help you, but I am going to PM you now with some thoughts…

CS,
Daniel

dkamen

Aug 28, 2020

Hi,

A BPM needs very long darks, five or ten minutes each. And the more the better. You prepare it separately and then reuse it in every real processing session, alongside that session's calibration frames. Unlike all other calibration frames which are applied pixel by pixel (eg pixel xy is divided by pixel xy from the master flat) the BPM tells APP which pixels should be ignored and interpolated from nearby ones instead.

That said maybe you do not need it at all. It does not work well with my Nikon D7500 and I have stopped using it.

Helpful

Frédéric Auchère

Aug 28, 2020

Hi Robin,

If non masters are allowed, here are a few thoughts:

I'm using a Samyang 135 and Astrodon refiltered 6D. I'm usually using it at ISO 3200 because I found it to be the lowest read noise setting on my copy, but that should not matter because at 135 mm, the read noise is usually small compared to the photon shot noise from the sky background. I'm not using APP but that should not matter, biases are biases and darks are darks

The bias on my 6D does not vary much from ISO100 to ISO3200. It is quite constant around value 2048. At 1600 or 3200 there are some column to column variations at the level of a few DNs, which is small compared to sky background shot noise. I've actually tried to use a constant value bias on my images and could indeed not tell the difference. That being said, you should in principle indeed apply a bias at matching ISO to your flats/darks/lights.

Are your dark flats taken at the same exposure time as your darks? I also take my flats at ISO 100. I'm using a lightbox and the exposures are a fraction of a second so I don't use darks to process the flats. With such short exposures there is negligible thermal noise, even at room temperature, so darks are indistinguishable from biases taken at the shortest possible exposure.

Actually from tests I made when I first got my 6D and played with it waiting for clear skies I found that up to 30 seconds (which is the longest exposure time I use) the thermal noise is very small and uniform. So I never use darks and rely on dithering to take care of the hot pixels.

But I suppose I should uses darks

and thus your workflow sounds similar to what I do. I don't use a bad pixel map though. Maybe the problem comes from that? Do you have an example?

Frédéric

Helpful

dkamen

Aug 28, 2020

Correct, a lot depends on what "bad" in the bad looking images actually looks like. It could even be unrelated to calibration. Darker fields are more unforgiving of minor tracking errors and dithering would actually help a lot.