PixInsight Forum (historical)

PixInsight => General => Off-topic => Topic started by: Ken Pendlebury on 2010 June 20 16:08:36

Title: Star Damage Question
Post by: Ken Pendlebury on 2010 June 20 16:08:36
Hi Folks,

This has happened in at least 3 images that I have processes with PI.  At Some time during the processing, I seem to be "damaging" larger stars, but only in the cores.  The star halos still look OK.  Tracing this back is a little difficult, because the damage seems to manifest itself as a small "lighter" spot in the star's core during the beginning steps, but is readily apparent when any attempt to saturate the image is made.  HDR Wavelets seems to enhance the core spots, but I am not sure if they are the root cause.

Here are a couple stars from an image of M101 that are representative of the damage I describe above:
(http://www.crossfitgeorgetown.com/ken/CN/m101_star1.jpg) (http://www.crossfitgeorgetown.com/ken/CN/m101_star2.jpg)

If there is any interest or information that can be gleaned by inspecting the full image, the wide field can be found here:
http://www.flickr.com/photos/kenastro/4715393083/sizes/o/

and the crop here:
http://www.flickr.com/photos/kenastro/4715393179/sizes/o/

I have attached my processing steps for M101 and a basic outline is as follows:


Thanks for any help,
Ken
Title: Re: Star Damage Question
Post by: Silvercup on 2010 June 20 17:14:46
Hi Ken:

HDRWavelets tends to produce artifacts on bright stars, I would try it with a starmask.

Best, Silvercup.
Title: Re: Star Damage Question
Post by: Jack Harvey on 2010 June 20 19:02:28
I usually do use a inverse star mask to mask my stars when using HDRWT also.
Title: Re: Star Damage Question
Post by: Niall Saunders on 2010 June 20 23:31:18
Hi Ken,

I have been struggling with this same problem recently on a bunch of M82 data - and I had decided that it was down to over-exposure in the RAWs. I might need to just re-think and re-examine the data and, if I see something esle (other than over-exposure) I'll get back to you.

Cheers,
Title: Re: Star Damage Question
Post by: georg.viehoever on 2010 June 21 01:41:32
Ken,

I have seen this in my images with stars that are close to saturation (i.e. have RGB values >0.9). I mainly see this when I boost colors using the V-Transform (discussed in http://pixinsight.com/forum/index.php?topic=1689.0), because the star's halo has a color that is different from the star core. My Canon EOS40D also is no longer linear in this range. Try star masks. Using a HDR combine of different exposures also helps.

Georg
Title: Re: Star Damage Question
Post by: Ken Pendlebury on 2010 June 23 20:28:23
Thanks for all the tips... I really appreciate the input.

Since these posts, I have reprocessed the image using a star mask for processes that are not kind to starts (i.e. HDRWT).  This has not yielded much better results.  In order to trace this issue back to its root, I simplified the processing of M101  for experimental purposes.  This simplified process consisted of the processing of a single uncalibrated frame in order to eliminate any calibration or stacking issues as suspects.  The steps I took, were debayer, Masked Stretch and then adjustment of the saturation curve.  Using only these three steps, my stars still exhibit the damage I showed in my original post.

So...   I am either damaging the stars during capture with inappropriate settings (gain and offset) or I the debayer / masked stretch is the culprit...

More to come as I investigate these remaining items.