Jump to content


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Not having fun trying to process images.

  • Please log in to reply
26 replies to this topic

#26 gdd



  • -----
  • Posts: 2327
  • Joined: 23 Nov 2005
  • Loc: Lynnwood, WA (N/O Seattle)

Posted 06 August 2014 - 05:14 PM

I remember another thing I was curious about when I made the attempt at posting earlier today. 

Why does the image look so much better in DSS prior to saving the file? You can see a lot more detail in the image?


DSS does a crude stretch for the displayed version of your image, but the DSS author recommends you use other processing tools to produce a better quality stretch. The image you save should not apply the crude stretch that DSS made or that you made while in DSS. This will leave the image in the desired linear format which will appear very dark like in your post.




#27 JoLo


    Ranger 4

  • -----
  • Posts: 353
  • Joined: 09 Feb 2010
  • Loc: Highland, IL

Posted 07 August 2014 - 01:40 PM

I have been using PI for about 6 months now and have completely ditched my other software (mainly Maxim and PS).  It was a huge learning curve, but well worth the effort.  I would make a couple of notes on the comments above.


I use noise reduction in several areas of the workflow, both linear and non-linear.  I find TVGDenoise to be too powerful for most astrophotos...I use it on daytime photos and it works wonders.  I use a very light MMT noise reduction in the linear state (no mask), then ACDNR in the non-linear state with a mask in place.  The beauty of ACDNR is you can do luminance and chrominance NR separately; also, the lightness mask built into ACDNR usually works quite well, without having to build a separate mask, but not always.  Use a light touch with NR and do it in steps, you will end up with a nicely smoothed image that doesn't look like it was smoothed, if you know what i mean.


The deconvolution algorithm in PI is excellent, but it just doesn't work on some images, it is trial and error.  Images with a lot of faint, surrounding nebula is where the background tends to the curdled cheese look, but it generally does wonders on galaxies and bright nebula with a good mask in place.  Master the DynamicPSF process for Decon, and use a star mask for deringing support, 30 to 50 iterations is usually the sweet spot.


PI has too many great tools to mention here, i would strongly recommend joining the PI forum.  Lots of good workflow and processing examples, lots of good Q&A, like a mini-CN.  Some of the more exotic (but not necessarily complicated) tools such as the DarkStructureEnhance script, ExponentialTransformation, HDRTransformation, MaskedStretch, and MorphologicTransformation are all worth the time and effort to learn.  Some of these have really pushed my processing to a new level.


I usually process my photos in PI, then in Maxim/PS.  8 times out of 10 I prefer the PI version.  Warren Keller's DVDs on PI processing are worth every penny.

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics