WBPP Pipeline - Debayer Issues

Hi,
I wanted to try out the drizzle feature in WBPP because the latest video released by Pixinsight recommended drizzling if you have a color camera. I checked the enable drizzle box in WBPP and re-launched a previously successful script. A few seconds after entering the Debayer process in the pipeline, Pixinsight crashed. I then re-ran an existing instance of WBPP from 26 Aug with drizzle disabled (which had run successfully on that date), but it crashed too at the same spot in the pipeline. The strange thing is that if I run Debayer as a stand-alone process it runs fine. The only thing I can see that was different with my configuration was that Pixinsight released an update on 28 Aug. Has anyone else had problems with Debayer when running with WBPP?

Sorry for the ramble.
Jim

Comments

  • My suggestion is to launch from scratch... I fear you are using cached information??
    Just confirm you are not doing that...and then other ideas might be reasonable.
    -the Blockhead
  • Sorry, it took some time to run the test cases, but now I am really confused.

    I restarted Pixinsight and reset WBPP per your direction. Then I loaded the WBPP with the same files as my successful 26 August run. This time rather than crashing in Debayer, Pixinsight crashed in "Measurements". I should note that when I say crashes I mean the WBPP execution monitor continues to run and update its progress, but the Pixinsight desktop disappears and the monitor screen goes white. All of the events ran successfully and I clicked "done" twice and "exit", but there was nothing left of the Pixinsight desktop other than an "X" in the upper right of the screen to close the program. I have 32GB of RAM in my machine. After the crash I had 13GB available according to the system information app, so it doesn't appear to be a shortage of RAM.

    There was one change to my configuration and that is a USB-C SSD drive rather than a HDD that I use for the swap file. I forgot to connect the SSD for the run above, so I did a second run with the SSD drive connected. I followed the same procedure, but this time WBPP crashed in "Registration", so the error does not appear to be repeatable.

    In both cases the results appear to be good. I "blinked" the registration files and everything was aligned.

    Thanks for any help/suggestions you can provide.
    Jim
  • I am not certain I understand.
    WBPP will not generate any files on the desktop. Everything sounds successful.
    What do you think is the issue?
    -the Blockhead
  • I figured it would be hard to explain. Yes, the WBPP execution monitor runs successfully and generates what appears to be all of the files, but everything else that is Pixinsight crashes (i.e. everything that is on the Pixinsight workspace disappears along with Pixinsight intself). The execution monitor remains by itself. I hope that makes more sense.

    I think it is some weird memory allocation/access problem.

    Jim
  • Blockhead,

    I thought an image might help explain what I am trying to say. I did another test run and this time Pixinsight, including the execution monitor, crashed in cosmetic correction. I attached a screen shot
    Jim
    image
    Crash1.jpg
    1891 x 1184 - 49K
  • Getting a white screen while the execution monitor is running is NOT unusual. It might come back after a while or maybe not until it is actually complete. How do you know it has crashed and is not still simply running?
    (I am saying this is just a display thing.)

    -the Blockhead
  • Hi,

    My conclusion that Pixinsight crashed in the Cosmetic Correction module is based on several factors:

    1.) My computer never exhibited this behavior (with identical inputs to WBPP) before upgrading to build 1579.
    2.) According to Windows Task Manager, the hard drive I/O write percentage was less than 5%. I would think that the percentage would be quite high if it was writing all of the files.
    3.) No files were created in the cosmetic correction directory.
    4.) I terminated the process after 50 minutes, when it was still in the CC process. Normally this test run takes about 39 minutes.

    Jim
  • edited September 2023
    Yeah, that does sound like a crash.
    Many things can do this from the 256 file character limit to other hardware specific issues.
    I would not expect the new version to cause this issue.

    (around 90% of the time that people claim it is because of the new version... actually just find it is a regular issue that occurred after they obtained the new version of the software. There is room for version specific issues... but there isn't much evidence for them with respect to crashing... especially in WIndows)

    -the Blockhead

  • Thanks for the quick response. I will check my character limit and be on the lookout for unusual behavior.

    Jim 
  • Well, this is interesting. My machine just did an automatic Windows Update to "2023-09 Cumulative Update for Windows 11 Version 22H2 for x64-based systems (KB5030219)"


    The problem has gone away! WBPP runs successfully to completion.

  • Hmm... OK!
    -the Blockhead
Sign In or Register to comment.