APP & PI calibration, registration and stacking routines

[tomato]

I am trying to get to grips with PI starting with the raw data, so I have been following the Light Vortex Astronomy tutorial to the letter to create master calibration frames, apply them to the lights, then register, normalise and stack the calibrated frames. I have achieved a reasonable result except something is clearly not right with the dark frame correction, based on the final stretched image  (LH frame in the attached screenshot). Out of interest I loaded up the same integrated frame from APP (derived from the same raw lights, bias, darks and flats) and applied the PI autostretch to it, APP seems to have calibrated the lights without creating the artefact created by the amp glow?

I'm way more familiar with the APP software so it's unfair to do a direct comparison, but superficially the stretched APP Lum stack appears comparable to the PI stack. Obviously my first step in taking this further is to sort out what is causing the artefact in the PI stack, but is there any advantage in doing the pre-processing exclusively in PI if you intend to continue processing with this package? I can certainly generate the stacks quicker in APP, but if there is an edge to staying with the same software throughout, I'll happily do that.

I must stress I am not trying to start a "Which package is better?" debate, I am just curious if other folks successfully use stacked frames created outside of PI for onward processing with this software.

[fireballxl5]

Good question, I'd also like to know if there is any advantage in using PI throughout. Only occasionally, and mainly with very wide field images, do I use PI for calibration and stacking as it seems to be better at handling very busy star fields.  Most of the time I do this pre- processing in MaxImDL and then take these stacks into PI.

[Laurin Dave]

You need to have the “optimise” setting turned off for Darks in PI calibration routine..  otherwise it messes things up..  assume you are using matched darks and flat darks and no bias ...  

I nearly always calibrate in PI, integrate and colour combine in APP then move back to PI (sometimes use APP for background extraction) for a bit of stretching (ArcSinh, Histogram and Curves) then Photoshop for the rest.   

Edited March 20, 2020 by Laurin Dave

[tomato]

56 minutes ago, Laurin Dave said:

You need to have the “optimise” setting turned off for Darks in PI calibration routine..  otherwise it messes things up..  assume you are using matched darks and flat darks and no bias ...  

I nearly always calibrate in PI, integrate and colour combine in APP then move back to PI (sometimes use APP for background extraction) for a bit of stretching (ArcSinh, Histogram and Curves) then Photoshop for the rest.   

Thanks for the pointer, I suspected the “optimise” option may be where the problem is, as I wasn’t using matched darks for the flats, but was relying on this feature to do the job.  Does APP avoid this by not having a darks optimise feature, and is applying the bias frames instead?

I like the look of your workflow, PI does have a vast array of tools but having said that I am also very impressed with Startools v 1.6. So many choices...

[Laurin Dave]

I don’t use bias so not sure about that..  I think the difference is that in APP you have to tick a box to turn on optimisation whereas in PI you have to untick it to turn it off..  you want it off so with APP the default is correct..  it’s all quite  tricky remembering what to do so I always calibrate and integrate my flats and darks (I don’t calibrate darks) separately in PI to make sure they’re ok and use masters in APP and Pi..  in PI there’s a file format conversion script in the “Scripts” menu "Batch Processing" "Batch Format Conversion"  if you need to convert .xisf files to .fits..  

also note that the image integration parameters you need for flats in PI are not the default ones..  (you need to use multiplicative scaling , equalise fluxes etc) whereas APP knows it’s integrating flats 

hth

Dave

 

Edited March 20, 2020 by Laurin Dave

[geeklee]

I added some comments in another thread recently, being a new user to PixInsight - along the lines of potentially going back to APP for all pre-processing then moving over to PI for all the post processing... and this is what I've  shifted towards - I just find APP very usable and efficient for pre-processing.  With the latest version (1.077) there's another layer of "automation" in the integration settings to make life easier if required.