[Sticky] Astro Pixel Processor 2.0.0-beta23 release notes

Astro Pixel Processor 2.0.0-beta23 release notes

• IMPROVED PERFORMANCE

  Due to several improvements, beta23 is again a couple of % faster when compared to beta22, especially when using FITS files.

• IMPROVED macOS FILE/DIRECTORY CHOOSER, CMD-A now works !
  

  We have improved the macOS FILE/DIRECTORY CHOOSER by using a new version of the file/directory chooser that now supports the macOS shortcuts with the CMD key. CMD-A to select all files now works again! The chooser does not look completely native unfortunately, the native chooser that was used from beta7 to beta22 for some reason does and will not support CMD-A and we were not able to get it working despite many attempts. We think that the new chooser works great however and having CMD-A working was a must which most users will agree we think. This new version also has a clear title/text on the window which the native chooser also was lacking, so that is another improvement.

• IMPROVED Windows & Linux FILE/DIRECTORY CHOOSER

  The file choosers for windows and linux still look the same and are native choosers. But as you can find in these release notes as well, 2 bugs were reported related to the file chooser on windows 10 & windows 11. This forced us to analyse the code and then rewrite it which solved both reported bugs. The filechoosers now have correct titles and taskbar icons as well and will popup a bit faster as well.

• IMPROVED FITS READ SPEED

  We upgraded our FITS dependency to nom.tam.fits 1.18.0 which gives a nice performance boost for reading FITS images. This is more noticable on big FITS images, clearly. 45MegaPixel images now read about 25-35% faster depending on hardware resources. The performance boost for bigger images than 45MP will likely be larger and vice versa.

• IMPROVED DATA ANALYTICS EFFICIENCY

  Internally we made some optimalizations that reduces memory usage and thus increases speed whenever data is analysed for background, scale, noise or Signal to Noise Ratio. Apart from the normalization engine, in many different modules of the application data is analysed for these metrics and thus the application overall has improved performance from this. Performance increase when running a dataset from calibration to integration will be about 1-5% from this depending on the data and hardware resources. The larger the images are to be analysed, the more performance gain should be expected.

• IMPROVED & FIXED BAD COLUMN COSMETIC CORRECTION

  OSC/CFA Bad column correction was not working, it threw a Null Pointer Exception in CosmeticCorrectionHotColumnCFACalibratorWorker. This is fixed and thus works now properly. In addition, bad column correction on already demosaiced color data also works now, in previous versions, this was not even enabled. In addition, the code to detect bad columns in mono/CFA/Color is greatly improved and the code to perform the actual data correction is also greatly improved. The cosmetic correction is also faster now and in the general console you can clearly see the amount of bad columns that are corrected. The following screenshots should illustrate the improvements compared to previous APP version:

• FIXED INTEGRATION CROP COMPOSITION MODE CONFIRMATION

  A bug was found as reported here: https://www.astropixelprocessor.com/community/main-forum/drizzle-integration-error-the-image-will-be-downscaled-with-a-factor-o-0-00/ . When you select the crop composition mode in 6) Integrate, you would need to draw a crop that represents the integration field of view from the reference frame and then click on the Apply Crop button. This button used to be red and visible in the 6) Integration menu, but if you did not click on this button and proceeded directly by clicking on the integrate button, this error would occur. This is solved now, by only showing the Apply Crop button in menu 6) Integrate, forcing the user to apply a crop before being able to proceed with integration.

• FIXED INTEGRATION BAYER DRIZZLE BUG

  If OSC data that still needs to be debayered is loaded and the user want to integrate with Bayer Drizzle ,but has set the demosaic algorithm in 0) RAW/FITS to no interpolation, an error was thrown in the Integrator stating an ArrayIndexOutOfBoundsException Index 1 out of bounds for length 1. The no interpolation algorithm is for users with OSC camera's that have the CFA filters removed from the sensor. The algorithm enables them to shoot monochrome narrowband with a camera that was previously a color camera. This exception is now caught properly with the following message and adjusted tooltip for the demosaic algorithm no interpolation:

• FIXED INTEGRATION BUG for CFA data with demosaic algorithm set to NO INTERPOLATION

  Similar to the previous fixed bug. If OSC data that still needs to be debayered is loaded and the user want to integrate with Drizzle or data interpolation ,but has set the demosaic algorithm in 0) RAW/FITS to no interpolation, an error was thrown in the AlignAndResampler stating java.lang.ArrayIndexOutOfBoundsException: Invalid coordinates. This bug is fixed and it all works now.

• FIXED FILE CHOOSER ISSUES

  As reported here: https://www.astropixelprocessor.com/community/appreleases/2-0-0-beta22-program-freezes-during-loading-master-flats/ and here : https://www.astropixelprocessor.com/community/appreleases/2-0-0-beta22-ui-glitch/ 2 bugs were reported whenever the file chooser was used. One was related to our development platfrom, upgrading the development platform fixed that issue. And the other issue was fixed by rewriting the file chooser code and adding more error checking into it which has fixed that as well.

• UPGRADED OPENGL DEPENDENCY

  We have updated the external OpenGL library to JOGL 2.5.0 from 2.4.0.

• FIXED image viewer issue where OpenGL encounters an error or runs out of memory

  An error was thrown if OpenGL encounters an error and APP tries to show the image in the non-OpenGL = CPU image viewer where it then fails... This error will not be shown again and the image will be showed correctly in the CPU image viewer. The error that was thrown was "Encountered error in module: FloatsToUshortBufferedImageConverterWorker Error message: java.lang.NullPointerException: Cannot invoke "java.awt.image.BufferedImage.getHeight()" because "this.a" is null etc..."