I am new to DSLR processing in PI

[malcolmpark]

Hello,
Using the Keller book, I integrated master dark, bias and flats, calibrated, debayered, aligned and integrated lights into the attached
Question for others processing Nikon DSLR astro images, is it typical to see ugly looking output to begin with, and now my job is to learn how to process it? Or does it appear that my pre-procesing has gone awry, with this useless output the result. (ie try calibrating again)?

thanks for your helpful and supportive comments in advance
Malcolm

[pfile]

seems normal. if you unlock the channels in STF you'll see something that looks better. but yes, you need to neutralize the background and do color calibration to get the colors into shape.

rob

[malcolmpark]

thanks!

[malcolmpark]

its a start!!!!!!!

[pfile]

nice!

rob

[dhb2206]

I'm imaging with just DSLR (but Canon in my case). I started out with darks and flats but have found that I get no real benefit from them.

Potted workflow from raw is:

Debayer
Register (I may use the subframe selector prior to registration if I'm feeling time generous)
Integrate
Drizzle integrate (if I remember to check the box at registration and add the drizzle files at integration)

Images always look really ugly at this point - either awfully green or red

Dynamic background extraction (threshold set to 2, auto generate 10 rows) - shunt (or delete) any samples that lie on nebula or stars, flit through the samples one by one - zoom the image so you can clearly see the full sample and use the blue arrow pointing right to center (focus) on the current sample. Once all are okay reduce the threshold to 1 and click the resize button, this may cause the odd sample to turn red, move or ignore those. Set subtraction as mode and check all three of the sub selections (using normalise makes it look like nothing much has happened, but it has!)
Background neutralise (watch for backgrounds >0.1, I've had a few of these, Lord knows why, change the upper limit if you have)
Image Solver script - I tend to do this to check that photometric colour calibration has a chance of working
Photometric colour calibration using the solved image coordinates (or colour calibration with the background sample and the whole image as white reference - and watch for that background threshold again)
Noise reduction with multiscale linear process (NR selected on layers 1 to 4, threshold at 5, 4, 3, 2 respectively) and a lightness mask (extract lightness and stretch using histogram transformation)
Stretch with histogram transformation (don't over cut the blacks - 0.05 clipping max and watch for overstretching - once the colour is into the background you are stuffed!)
Re-run multiscale linear with reduced thresholds (3, 2, 1, 0.5) and a fresh lightness extract as a mask
At this point I tend to vary, but if the image looks reasonable I'll just dive into the curves; make a slight S out of RGB-K, increase lightness, increase saturation

This will probably get a load of groans, but it works for me! 

Davy

PS Got one of 46P myself - one clear night before it went between Pleiades and the Aldebaran asterism.