Checkerboard in images after Batch Preprocessing.

scadvice

Member
since the last two upgrades I had problems getting a checkerboard pattern when batch preprocessing.
Its happened for the last four images I've tried to do and I've run out of ideas and stopped processing any images. If I stretch a single image it is clean and without the checkerboard.
The Registered files all have _a_c_d_r

The Calibrated flat files have their_a _c

The Calibrated light files have their _c
and the debayered _a_c_d

Pulling up and looking at one of each look right.

Current version 1.8.8-5/

Camera DSLR Canon T3i Modified.

Is there a process log after shutting down and starting back up again? I wanted to look back through the last image Process Console.



CB sample.jpg
 
Last edited:
what kind of files are coming out of your camera? CR2? fits?

if fits, you may need to change the state of the "up-bottom FITS" checkbox and rerun everything.

rob
 
The files are CR2. As I said earlier this showed up the last two rev's ago. Prior to that everything worked great. Also if pull up an indivual image prior to processing there is no pattern it any of the lights, darks, bias, or flats.
 
Last edited:
well the checkerboard pattern would indicate to me that somewhere along the line the files were not debayered, or debayered with the wrong bayer matrix pattern. one recent problem with the RAW file handler is that the guys who write it (the libRAW project) changed how they parse the CR2 files and this completely messed up the normal debayer pattern for certain canon cameras.

if your masters were created with an older version of PI then this could be the cause of the problem. one way to overcome this is to just load all your CR2 files (bias, darks, flats, lights) into BPP and re-run, thus recreating the master files with the new bayer pattern. i'm not fully versed in this problem - it's still possible that the bayer pattern will be incorrectly automatically selected even if you do this, but i think it should work. @bulrichl knows a lot about this problem and can correct me.

if your RAW file handler is set to debayer CR2 files upon opening onto the desktop, that's why you don't see the checkerboard pattern. however BPP and WBPP override the settings of the RAW file handler so that initially the files are opened without any debayering.

rob
 
The masters are newly generated new each time I run through BPP.

The only thing that has changed is I set up a new faster computer just for PI. I believe this new computer was installed with the version prior to the latest version at first then to 1.8.8-5. In the morning I'll run one of the images on the old computer as I believe I still have it using version 1.8.8-2
 
Last edited:
The Canon Rebel T3i (= 600D) is one of the cameras that is affected by the change in the LibRaw software that is used by PixInsight as raw image decoding software. Due to this change, the images are cropped differently (formerly the output image had h=3465, now it has h=3464) and so the bayer/mosaic pattern changed from RGGB to GBRG.

This is not a bug, but intended behavior. If you don't use old master calibration files that were produced with PixInsight version 1.8.7 and older versions, you only have to keep in mind that the bayer/mosaic pattern is changed. However, now the correct bayer/mosaic pattern will be used as well when you set it to 'Auto' when the Debayer process is applied, you don't need to set it explicitely any more.

Bernd
 
Hi Bernd,

To be sure I understand what your telling me. When I pulled up the BPP screen and the bayer/mosaic pattern selection under Debayer it is defaulted to RGGB. Are you suggesting to change that to Auto selection and that this should or may correct my problem?
 
Suggestion: open a light frame from the camera (in CR2 format) and watch the output of LibRaw in the Process Console. It will tell you (among other things) the correct bayer pattern: GBRG.
 
Selecting Auto in BPP screen and the bayer/mosaic pattern selection under Debayer seems to have fixed it.

bulrichl I will check that before I do the next set..

I'll post again the after I reprocess another image set.
 
Back
Top