Author Topic: COULD PI IMPROVE ON THIS IMAGE OF M42?  (Read 296 times)

Offline ballyhoo

  • Newcomer
  • Posts: 21
    • View Profile
COULD PI IMPROVE ON THIS IMAGE OF M42?
« on: 2019 April 03 15:03:05 »
Hello.

This is about a 32 min integration of 2 min subs taken on an AS! 294 Pro, cooled at -10. Most everything was processed on Nebulosity, and I am just not sure whether there is more to be done here given the 30 min of integration.   I did use the screen transfer function. but that was about all?
Any recc's for some tools I might apply, to learn more about using PI?

Thank you

Online Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6537
    • View Profile
    • http://pixinsight.com/
Re: COULD PI IMPROVE ON THIS IMAGE OF M42?
« Reply #1 on: 2019 April 04 00:40:01 »
Judging from the picture you have posted, the background is severely clipped (lots of zero background values, the main histogram peak is clipped at the left side). Not doing this is probably the best thing you can do to improve the image. With that in mind, there is probably much more you can achieve if you start from the raw data from scratch.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

Offline ballyhoo

  • Newcomer
  • Posts: 21
    • View Profile
Re: COULD PI IMPROVE ON THIS IMAGE OF M42?
« Reply #2 on: 2019 April 04 14:53:07 »
I recall reading the famous Primer book, where he discusses clipping of the histogram. Not sure I understand it though.

How many minutes of sub time gernaly with like an 85mm refractor is required to get the optimal integration of M42? In my case I had about 30 minutes.

Thank you

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 319
    • View Profile
Re: COULD PI IMPROVE ON THIS IMAGE OF M42?
« Reply #3 on: 2019 April 05 10:41:52 »
Which gain was set in the camera?
Are you using calibration frames?
What is your calibration workflow?

Statistics and histogram of your JPG image look strange. In the R and G channel, 70 % of the pixels are clipped (i.e. = zero), in the B channel 63 % are clipped. This has to be avoided absolutely. In an astrophoto, the background (region without objects) shall be in the range of about 0.08 to 0.15. That means: in the histogram, the peak of the background pixels shall have its maximum in this region. The peak shall not be cut sharply at the left side but shall smoothly fall to the origin of the diagram.

I assume that your images are calibrated. If the clipping is already visible in the calibrated images, there is something wrong with your calibration master files or in the calibration workflow.

If the calibrated images are not clipped or if you are not calibrating the images (of course you should), the clipping results when you stretch the image: you are applying a too large 'Shadows' value. The HistogramTransformation is showing you which fraction of the pixels will be clipped with the current 'Shadows' setting, see the grey field (red arrow in the appended screen section): the number of clipped pixels and the percentage of clipped pixels will be indicated. Observe these values before performing the stretch and avoid values > 0.1 % if possible.

Since M42 is a very bright object, 30 min of total integration might last if you don't aim for bringing out faintest nebula surrounding M42.

Bernd

Offline ballyhoo

  • Newcomer
  • Posts: 21
    • View Profile
Re: COULD PI IMPROVE ON THIS IMAGE OF M42?
« Reply #4 on: 2019 April 05 16:46:18 »
I have some learning of terms etc before I understand everything you have posted on.  I did not add flat frames, though I will be re-doing and adding flat frames.  I iny had about  6 darks, which I will also some some in the re-do. I will also ad bias. The integration actially was done in Nebulosity. Isn't that the case that most PI users integrate in DSS first?  In any case, I will integrate in the PI script, and uise sigma clipping. But I am not sure how that will affect the channels. I really do not understand channels yet.


Edit, also the gain was set low, about 20, and I have since learned to raise the gain to about 60 for my ASI 294. Not sure about offset though I am open to suggestions. For whatever reason I use Nebulosity for capture rather than SharpCap. 

Offline bulrichl

  • PixInsight Old Hand
  • ****
  • Posts: 319
    • View Profile
Re: COULD PI IMPROVE ON THIS IMAGE OF M42?
« Reply #5 on: 2019 April 06 02:50:41 »
> Edit, also the gain was set low, about 20, and I have since learned to raise the gain to about 60 for my ASI 294. Not sure about offset though I am open to suggestions. For whatever reason I use Nebulosity for capture rather than SharpCap.

When using calibration frames it is important to use the same gain and offset settings for the lights and all calibration frames. Don't alter the default offset settings of the camera. I guess Nebulosity saves the frames in FITS format?

> I have some learning of terms etc before I understand everything you have posted on.  I did not add flat frames, though I will be re-doing and adding flat frames.  I iny had about  6 darks, which I will also some some in the re-do. I will also ad bias. The integration actially was done in Nebulosity. Isn't that the case that most PI users integrate in DSS first?

Oh no, you really should do the preprocessing in PixInsight, see the other thread. For more details on preprocessing, see my guide: https://pixinsight.com/forum/index.php?topic=11968.

Bernd

P.S. For some reason your image of M42 is mirrored.