›

Fast Track, WBPP calibration stage, debayered lights w/white pixels w/blue halos, green artifacts

Cindy Yepez
in PixInsight
Hi, I am working my way through the Fast Track tutorial and using the sample data you provided. I searched to see if anyone else in the forum had posted about this but could not find anything.

I ran the smaller data set you recommended through WBPP, doing only the calibration per the tutorial. When I Blink through a subset of the debayered light files I am seeing quite a number of white pixels surrounded by a bright blue halo, also occasionally a green "plus sign" shaped artifact (or just a bright green pixel).  I tried repeatedly to upload a screenshot of my workspace but the file would not upload even after I reduced the size.

I went back and looked at the original raw data files, the dark files, also the master darks, everything that went into the final debayered lights but there are no hot pixels or other artifacts that correspond to these new ones. 

Is it normal to have a small number of these flaws with OSC files after debayering? Every light file I looked at had some, but they are in a different place every time.

I followed the tutorial closely with all the settings. I am running the latest version of PixInsight. I did choose not to use the built-in cosmetic correction but instead created the template like you explained from earlier versions of WBPP.

I am finding the lessons extremely helpful by the way. So many PI tutorials never explain the "why". The person just says this is how they do things.


Comments

  • BlockHeadBlockHead
    The funny thing is that in this FastTrack series...I am still quite restrained and not explaining too much "why" ...  lol

    Please put your image on a cloud server somewhere for me to see and share the link here (make certain the link is accessible to anyone that has the link).

    Debayering will not produce "flaws" in the sense you are indicating. So, I think there is something else to explain here. 

    -the Blockhead
  • Cindy Yepez
    So, I'm a little embarrassed. Silly beginner mistake. I went back to double check how I set up the Cosmetic Correction and I forgot to check the CFA box. After I watched the YouTube video on Star Alignment, Debayering, etc. it was obvious those hot pixel areas were NOT supposed to be there! AND... it was my fault. I will re-run WBPP, correctly this time.
    And this is why your training is so helpful, because who knows how long it would have taken me to figure out what I did wrong.
    Thanks for the quick response!
  • Cindy Yepez
    Well darn it, the issue is still present after I re-ran the data. I watched your YouTube video about this issue three times but can't figure out what I might be doing wrong. 
    Link to a screenshot of my workspace and copy of an example file: 

    Based on your explanations, these are hot pixel areas. But shouldn't I also see the same hot pixel on the unprocessed file? I tried your technique of doing a debayer on a raw file to look for the hot pixel but the ones that show up on the calibrated file after wbpp do not show up in the same place in the raw file. Is there something I could be doing that is adding these?
  • BlockHeadBlockHead
    You did find a key insight by noting the different hot pixel population. At the moment I can’t help. I’m unavailable for the next 4 to 5 days. If you can wait that out I’ll be happy to look at the files. What you really need to do is make three raw files in all of the light frames dark frames Bias frames available as well as the calibrated frames. That should be enough to answer your question and see what’s going on.
  • Cindy Yepez
    I did some tests with smaller subsets of data, just 12 raw files. The issue appears to be with the setting used with the cosmetic correction template. 

    The first time I simply used the default CC template settings, hot pixel sigma value of 3. This is when I see the apparently random new hot pixel artifacts appear in the debayered files. I compared the same three files with each other at each stage of the process and the flaws only show up after they are debayered. I checked the X and Y coordinates of the flaw areas and compared them with the same coordinates in all the previous stages of processing. As far as I could tell, none of them are in the same place as hot pixels that were showing up in either the raw files or the darks.

    I did two other tests, first using the built-in cosmetic correction in WBPP and left that at the default sigma of 10. The other test I used a CC template again, but this time raised the hot pixel sigma threshold to 5. In both instances the original problem disappeared. So doing CC with a sigma of 3 was too aggressive?

    I did upload some zipped files of the information you requested but please don't feel obligated to spend a bunch of time looking at these. The issue was solved by changing a setting. 
    I'm mainly interested to know if having CC set too aggressively can actually cause this with OSC files.

    Link to Google drive files: 

    Thanks!
    Cindy
  • BlockHeadBlockHead
    Hi Cindy,

    Sorry. I do not think we should stop here. There is likely a misunderstanding of something- but I cannot put my finger on it. You need to upload some screenshots of what you think is the problem. You indicate a before and after effect that where artifacts are appearing. 

    Here is the thing. Cosmetic Correction is more aggressive the LOWER the value. Raising the value would not somehow make things better. Also, we are both working with the same data correct?

    I think what you are seeing is that you can see a lot more information after frames have been calibrated and debayered. You would not have been able to see some of the stuff you are talking about- unless you manipulated the screen stretch just so.

    Please do upload some JPEGS to your folder that explains what you think is going on. I am reluctantly to let this go since your solution shouldn't really be one. (Cargo science as Feynman would say.)

    -the Blockhead
  • Cindy Yepez
    Hello, I am quite sure I am misunderstanding, and probably more than just one thing! I am a rank beginner and often I'm not sure how explain what I think could be wrong. WBPP feels to me like a magic black box you throw a bunch of stuff into and get an image out the other side. That is why I wanted to take your courses which break everything down into separate components.

    *we are both working with the same data correct?*  Well I hope so...

    I am following along in the basic FastTrack training, using the data set you provided of the Flaming Star Nebula. Per your instructions, I am just using the lights files from the first three nights of imaging.

    When I ran WBPP, calibration only, I just set mine up like you had in the video with a CC template using the hot sigma value of 3. (and yes, I understand that the lower the value, the more aggressive it is)

    I just used Blink to examine the images afterward. I did not do anything unusual. When I zoomed in, I found lots of those spots that had a bright white pixel surrounded by multiple bright blue ones. 

    I uploaded some screen shots to the Google Drive folder.

    Richard Feynman was quite a guy :-)

    Cindy
  • BlockHeadBlockHead
    Thank you for creating the screenshot. That is very helpful to understand the issue.
    You are demonstrating something that in practice does not make sense. If CC with 3 doesn't remove the hot pixels and 5 does..that is somewhat non-sensical. 

    So...I am going to propose something that would explain the behavior. (Because one of the things about being an instructor is that I have made most possible mistakes...including this one...) It is possible t create a CC template and forget to actually check the Hot Sigma box. If you did this on the 3 sigma template and not the 5 sigma template..it would explain everything.

    I am sorry to ask you to do this..but I would try to run a CC template again with 3 and make absolutely certain you check BOTH the Use AutoDetect AND the Hot Sigma checkboxes.

    -the Blockhead
  • Cindy Yepez
    I did another run, using the CC template, hot sigma 3. Everything looks totally fine this time. I did run a small batch of files with a deliberately misconfigured CC template where I left the hot sigma box unchecked, but it does not reproduce the problem. It also throws a red flag that says cosmetic correction failed and I would have noticed that.

    I thought there was a possibility with the original bad set that I thought I had checked the CFA box in the CC template but then did not save it correctly. Remember, this all started when I forgot to check the CFA box. However, I checked the log file for that run and the template was correct. I pasted an excerpt below. 

    So basically, I cannot reproduce the problem!  Thank you very much for your time and patience. I am going to chalk this up to an unidentifiable transient issue and move on.

    [2025-04-05 23:19:35] * Begin cosmetic correction of light frames
    [2025-04-05 23:19:35] ************************************************************
    [2025-04-05 23:19:35] Group of 182 Light frames (182 active)
    [2025-04-05 23:19:35] SIZE  : 6248x4176
    [2025-04-05 23:19:35] BINNING  : 1
    [2025-04-05 23:19:35] Filter   : NoFilter
    [2025-04-05 23:19:35] Exposure : 120.00s
    [2025-04-05 23:19:35] Keywords : []
    [2025-04-05 23:19:35] Mode     : calibration
    [2025-04-05 23:19:35] Color   : CFA
    [2025-04-05 23:19:35] Cosmetic Correction: applying CC process icon.
    [2025-04-05 23:19:35] Sanitized group folder name: [Light_BIN-1_6248x4176_EXPOSURE-120.00s_FILTER-NoFilter_CFA] -> [Light_BIN-1_6248x4176_EXPOSURE-120.00s_FILTER-NoFilter_CFA]
    [2025-04-05 23:19:35] ------------------------------------------------------------
    [2025-04-05 23:19:35] var CC = new CosmeticCorrection;
    [2025-04-05 23:19:35] CC.targetFrames = [ // enabled, path
    [2025-04-05 23:19:35] ];
    [2025-04-05 23:19:35] CC.masterDarkPath = "";
    [2025-04-05 23:19:35] CC.outputDir = "D:/Astrophotography/Adam Block Flaming Star Nebula/WBPP Output Calibration only/cosmetized/Light_BIN-1_6248x4176_EXPOSURE-120.00s_FILTER-NoFilter_CFA";
    [2025-04-05 23:19:35] CC.outputExtension = ".xisf";
    [2025-04-05 23:19:35] CC.prefix = "";
    [2025-04-05 23:19:35] CC.postfix = "_cc";
    [2025-04-05 23:19:35] CC.generateHistoryProperties = false;
    [2025-04-05 23:19:35] CC.overwrite = true;
    [2025-04-05 23:19:35] CC.amount = 1.00;
    [2025-04-05 23:19:35] CC.cfa = true;
    [2025-04-05 23:19:35] CC.useMasterDark = false;
    [2025-04-05 23:19:35] CC.hotDarkCheck = false;
    [2025-04-05 23:19:35] CC.hotDarkLevel = 1.0000000;
    [2025-04-05 23:19:35] CC.coldDarkCheck = false;
    [2025-04-05 23:19:35] CC.coldDarkLevel = 0.0000000;
    [2025-04-05 23:19:35] CC.useAutoDetect = true;
    [2025-04-05 23:19:35] CC.hotAutoCheck = true;
    [2025-04-05 23:19:35] CC.hotAutoValue = 3.0;
    [2025-04-05 23:19:35] CC.coldAutoCheck = false;
    [2025-04-05 23:19:35] CC.coldAutoValue = 3.0;
    [2025-04-05 23:19:35] CC.useDefectList = false;
    [2025-04-05 23:19:35] CC.defects = [ // defectEnabled, defectIsRow, defectAddress, defectIsRange, defectBegin, defectEnd
    [2025-04-05 23:19:35] ];
    [2025-04-05 23:19:35] ------------------------------------------------------------

Sign In or Register to comment.