Problems with PI 1.8.8.3

[hvoneiff]

I am working with PI since several years.
With the latest version 1.8.8.3 I am no more able, to process my images.
When I start the BatchPreprocessing script, I get the message:
Warning: Incremental file reading disabled because of lack of file format support: RAW
I am working with the Canon 5D Mark II since seven years.
When it comes to the calibration of the flats, I receive an Error and the program stops with the message "Wrong Image Geometry" or something like this.
In my lights, flats and darks, I have exactly the same numbers for the image geometry and raw dimensions in all the images.
I tried to do the BatchPreprocessing with older sets of images, which have worked previously and got the same wrong result.
So I switched back to version 1.8.6 and everything is now running smoothly.
I do not need all the fancy achievements, which cause PI to crash.

Kind Regards
Hermann von Eiff

[bulrichl]

Hi Hermann,

please read the following threads:
https://pixinsight.com/forum/index.php?topic=14219
https://pixinsight.com/forum/index.php?topic=14206

Bernd

[hvoneiff]

Dear Bernd,
thanks for your information. It seems to be pretty complicated.
To make it more simple for me: Does that mean, that I cannot use master files (i. e. Masterbias or masterdarks) from earlier PI versions with the new LibRaw
and should calibrate and intergrate them again in the new version? This would be not an issue for me.
Obviously I had the problem with the masterbias from the older LibRaw version. Lights, darks and flats were new.
If you agree with me, I would try the new PI version with a new masterbias.
Thanks for your kind support
Hermann

[bulrichl]

To make it more simple for me: Does that mean, that I cannot use master files (i. e. Masterbias or masterdarks) from earlier PI versions with the new LibRaw
and should calibrate and intergrate them again in the new version? This would be not an issue for me.
Obviously I had the problem with the masterbias from the older LibRaw version. Lights, darks and flats were new.
If you agree with me, I would try the new PI version with a new masterbias.

Yes, the 5D Mark II is one of the affected cameras. The issue is: the old LibRaw version produced a cropped image (the optically black region removed) whose height deviates by 1 pixel from the cropped image of the new LibRaw version. So new images are not geometrically compatible with old Master calibration files.

That means: if you saved the dark frames and flat darks (as CR2 files), you can generate a new MasterDark / MasterFlatDark with the new PixInsight version, and then calibrate the new light and dark frames.

If you did not save the dark frames and flat darks (CR2 files), you would have to capture new dark frames and flat darks.

Probably the CFA pattern that you were used to (RGGB) is also changed thereby. The output in the process console will tell you, which width, height and bayer/mosaic pattern LibRaw detects. The following results with my 600D which is also one of the affected cameras (the height was 3465 and the CFA pattern RGGB before):

Reading 1 image(s):
G:/Astro/171218/20171218_0009.CR2
Reading metadata: done

Camera ........... Canon EOS 600D
Timestamp ........ 2017-12-18T22:07:52Z
Exposure ......... 302s
ISO speed ........ 800
Focal length ..... 50 mm
CFA pattern ...... Bayer GBRG
Raw dimensions ... w=5344 h=3516
Image geometry ... x=142 y=52 w=5202 h=3464

Raw decoding parameters:
Output mode ............... cfa
Auto crop ................. enabled
Wavelet noise threshold ... 0
White balancing ........... disabled
Black point correction .... disabled
Highlights clipping ....... disabled
Auto rotate ............... disabled
Output image .............. w=5202 h=3464 n=1 Gray

Reading RAW data: done


Bernd

[hvoneiff]

Everything works now! Thanks again for your support
Hermann

[bulrichl]

Thank you for the feedback, Hermann. I am glad that it worked for you.

Bernd

[hvoneiff]

I would be happy, if it would really work.
From now on my images are green and I don't get the right Colours with Colour Calibration.
Switched back again to 1.8.6 and hope, this problem will be solved at anytime.
RGGB is RGGB for Canon and not GBRG.
Hermann

[hvoneiff]

Sorry, this was my mistake.
I had to enter the right Bayer Matrix into the according field of the BatchPreprocessing script.
Kind Regards
Hermann

[bulrichl]

Yes, the changed bayer/mosaic pattern (GBRG instead of RGGB) is a consequence of the changes made in LibRaw. However, it should have worked also with bayer/mosaic pattern set to 'Auto'!

Bernd