Code Obsession Forum

Observatory => General Discussion => Topic started by: Sander Berents on January 23, 2018, 09:24:59 PM

Title: Observatory 1.2 arrives with enhanced image processing
Post by: Sander Berents on January 23, 2018, 09:24:59 PM
We have just released an update to Observatory. It enhances its nondestructive preprocessing workflow and adds eight new adjustments:

All of these adjustments work without the hassle of intermediate files.

(http://blog.codeobsession.com/img/ChromaticAlign@2x.png)

Other improvements and changes


(http://wwwcdn.skyandtelescope.com/wp-content/uploads/SkyMapper-MilkyWay-300px.jpg)

Full release notes here (http://codeobsession.com/observatory/releasenotes.html).

Observatory is available from the Mac App Store (https://itunes.apple.com/app/observatory/id1037215068?ls=1&mt=12).

Title: Re: Observatory 1.2 arrives with enhanced image processing
Post by: Sander Berents on January 23, 2018, 09:46:43 PM
Since this forum was set up in response to this thread (https://www.cloudynights.com/topic/536738-new-observatory-for-mac/page-2#entry8355950) on Cloudy Nights, here the part that followed the above post:

QuoteVery nice, I can't wait to try some of the new features. I would love to see a simple workflow that explains, importing, stacking, alignment/calibration, and updating image data across all the images of a set. I know some of this exists on the help website, but it would be nice to see it as a single workflow of how to work with the program. After owning it for a year, I still feel a little lost, and seem to be only able to get a few of the functions to work. But I think it stems specifically from not understanding the proper workflow to use with the application and my images.

Did you check the Getting Started article and image stacking video on the website? The videos quickly get outdated because new features are added all the time, but I would have thought it shows a typical workflow. Relevant changes since this video was made are


You are not alone in having trouble understanding the proper workflow for processing images with Observatory. This may be because Observatory's approach is very different.
Other image processing applications generally work like this:

Or, for some processing steps, they work like this:

Often it is a mix of these two approaches.

Observatory is at its core an image management application. When you create an Observatory library document with it (you can have many), that document will contain the references to the images you "import". These images do not have to reside in a single folder, but can be all over your disk, or even spread over multiple disks. Observatory stores the metadata of these images in such library (e.g. exposure date, duration, etc), but not the actual image itself.

When you select one or more images in Observatory, it will starting loading them in memory so they can be displayed, and it will display the normalized metadata. You can edit this metadata, also in bulk, and it will only affect what is stored in the library. Observatory will never alter your source images.

How you process an image in Observatory is by enabling "adjustments". Similar to for example Apple's Photos, none of these adjustments are destructive. The original image is never altered. Adjustments act like a recipe, and are in-memory only. You can change adjustment parameters, and Observatory will automatically recompute it all and display the resulting image. Most of the adjustments (e.g. calibration, cosmetic correction, debayering, flatten background, stacking, etc), are capable of doing the heavy lifting on the GPU, but Observatory still needs a few more additional enhancements to make that really fly (also, only use that option if you have a machine with at least two GPUs). On a CPU they will use all available cores and are very fast, but there too there is still room for optimizations.

There are a few things that may confuse some users:


What you need to do, is opening the stack (Focus on Stack), select all its images (choose ... button), and apply adjustments to the source images. An approach I often follow is:


Other important points:


btw, I am considering setting up a discussion forum on the Observatory website. I think that would be a better place to discuss it. What do you think?

QuoteThis is all extremely helpful. I'll have to apply some of what I learned here in the coming nights. I think a discussion group on the site might benefit your users more than it would here on this board where people would miss discussions due to the size and varied topics on CN.

I'll set it up.

Two more things: