Help with First Pixinsight Image

[dyfiastro]

Good evening everyone

Until now I have been using a combination of Photoshop and DSS. The main issue I have had is with colour correction and feeling like have not been able to get the most from my data.
After reading and watching many videos of people using Pixinsight I thought I would give the a trial version a go.
I have spent the last day trying to get to grips with it and watching tutorials.

The main issue I seem to be having at this stage is colour correction. I am sure I must be doing something wrong but cannot figure out what.
Everything has been done via manual calibration and integration and so far I have attempted the following:

DBE > Background Neutralisation > Colour Calibration
DBE > Background Neutralisation > Colour Calibration > SCNR
ABE > Background Neutralisation > Colour Calibration > SCNR
ABE > DBE > Background Neutralisation > Colour Calibration > SCNR

I have also attempted to split and recombine the image as described using linearfit before doing the above.

RGBWorkingSpace > Channel Extraction > Statistics > Linearfit (Lowest noise frame as reference) > Channel Combination

This seemed to be the best option so far but I am either getting very a very desaturated image which is hard to re-saturate or giving strange colour casts and blue stars.
I have provided a link to the Final FITS file (HERE) that I am working from and any advise would be great.
The image was taken over the last week so I am wondering if the fact that it was not completely dark may be throwing the calibration off.
The camera is a full spectrum 600D and this set of data was taken without any CLS or IR filter.

Any help would be great and thankyou in advance

Mark

 

[Scooot]

Hi Mark,

I'm a pixinsight novice but are you integrating CR2/raw files in pixinsight before you start to process them? I was wondering because pixinsight usually opens raw files in mono and you then have to debayer them to add the colour back in before any colour calibration routine. 

[dyfiastro]

2 hours ago, Scooot said:

Hi Mark,

I'm a pixinsight novice but are you integrating CR2/raw files in pixinsight before you start to process them? I was wondering because pixinsight usually opens raw files in mono and you then have to debayer them to add the colour back in before any colour calibration routine. 

Thanks for the reply.
No I am doing all the Calibration and Integration within Pixinsight using the manual method (Not Batch)
I have changed the RAW settings to enable "Pure RAW" and then debayer after calibration but before integration

[wxsatuser]

Colour calibration with a dslr, for me was always a nightmare with PI.

There is colour mottle in the final FITS and no doubt in the original files.

The very best I did was using dithering, this defeats the colour mottle and colour calibration works better.

[dyfiastro]

8 hours ago, wxsatuser said:

Colour calibration with a dslr, for me was always a nightmare with PI.

There is colour mottle in the final FITS and no doubt in the original files.

The very best I did was using dithering, this defeats the colour mottle and colour calibration works better.

Thanks for the reply. I am indeed dithering for this very reason. I did take another look at the image that I was using and think the colour may being thrown off by the fact that the sky was not completely dark.
I am going to go back over some older data and see if I get better results

 

Thanks

[wimvb]

Try this:

1. Dynamic crop to get rid of any stacking artefacts

2. DBE, manually placing samples in each corner and along the edges. Sample size 25 pixels. Increase the tolerance so that all samples are valid.

3. Create a preview that contains background and use that for BackgroundNeutralization

4. Use the same preview for background reference in ColorCalibration. Use the entire image as white reference.

This usually gives a good colour balance.

I'm downloading your image to have a look at it.

[wimvb]

Download finished.

The image has a gradient, but that is easy to remove with DBE. See sample placement in screengrab.

Background Neutralization with a small preview (preview1) taken from the lower left works ok

Colour calibration with the galaxy as white reference is also ok. (white reference is aggregation of preview2 and preview3)

(click image to enlarge)

As you noted, there is very little colour in the image. Did you stack in PixInsight? If so, are the Bayer matrix settings ok?

If you want to increase the colour saturation in the galaxy without affecting the stars, create a starmask and protect the stars with that. Then use colour saturation in the CurvesTransformation process to increas colour saturation. You can apply this several times.

A suitable mask for this can be created as follows:

Extract a luminance from the colour image.

Use this image to create a star mask:

      noise threshold = 0.35 (about), scale = 7 or 8, large scale = 1, small scale = 4, compensation = 2

Then use Pixelmath with the expression:

      $T*~star_mask

and apply this to the luminance image This will remove the stars from the luminance image.

Then use the luminance image as a mask for the stretched colour image and increase colour saturation of the galaxy without affecting the colour of the stars.

[dyfiastro]

11 hours ago, wimvb said:

Download finished.

The image has a gradient, but that is easy to remove with DBE. See sample placement in screengrab.

Background Neutralization with a small preview (preview1) taken from the lower left works ok

Colour calibration with the galaxy as white reference is also ok. (white reference is aggregation of preview2 and preview3)

(click image to enlarge)

As you noted, there is very little colour in the image. Did you stack in PixInsight? If so, are the Bayer matrix settings ok?

If you want to increase the colour saturation in the galaxy without affecting the stars, create a starmask and protect the stars with that. Then use colour saturation in the CurvesTransformation process to increas colour saturation. You can apply this several times.

A suitable mask for this can be created as follows:

Extract a luminance from the colour image.

Use this image to create a star mask:

      noise threshold = 0.35 (about), scale = 7 or 8, large scale = 1, small scale = 4, compensation = 2

Then use Pixelmath with the expression:

      $T*~star_mask

and apply this to the luminance image This will remove the stars from the luminance image.

Then use the luminance image as a mask for the stretched colour image and increase colour saturation of the galaxy without affecting the colour of the stars.

Thankyou very much indeed.

Yes the image was calibrated and stacked using the manual processes in the light vortex tutorial. I will take a another bash at the image this afternoon and let you know how I get on.

[dyfiastro]

I had another go this afternoon and even pre-processed the images again.
I got a better result this time using the settings but still not something I was happy with.
Looking at the RAW files I have a feeling the sky background is throwing my resulting colour off hence why I am losing some of the blues in the outer arms.

I am in the process of going over previous other shots from during the winter (with darker sky) and seems to be getting a lot better results.

Thanks again and I have made a note in regards the Pixelmath equation as well.

Regards

Mark