PixInsight Forum (historical)
PixInsight => Bug Reports => Topic started by: rtemen on 2019 December 17 06:12:17
-
When trying to calibrate my darks from last night, I get the error message about image geometry.
I have been through many example solutions on the forum to no avail.
Then I tried to calibrate a dark that I already did back in October, and it will not work either.
Is there a problem with the latest version that I recently upgraded to?
Or, is there some sort of new setup that I have neglected to do properly?
Thanks,
Rich
-
are you using a DSLR? there have been some problems caused by the libRaw folks that messed up how DSLR images are interpreted by PI...
rob
-
Yes, I am using my Canon T3i.
Same camera I have been using the last 5 years.
-
the easiest thing to do is make a new master bias from the same files, then calibrate your dark subs and remake the master dark.
the problem is that the CR2 format is proprietary and all 3rd party decoders are reverse-engineering the file format. the libRaw people recently made a decision about how to treat the sensor overscan area that changes the file dimensions. they did this somewhat unilaterally and since most people doing terrestrial photography don't care about calibration or indeed what their file dimensions even are, no one really noticed the change except for people like us doing AP.
having said all this, i think juan made a change to PI which undoes the changes that the libRaw guys made. have you checked to see if you have any PI updates available and/or have you upgraded to the latest version of PI?
rob
-
I currently have 1.8.8-3
There are no available updates at this moment.
I always keep up to date on updates, including the one from yesterday.
-
Hi Rich,
Rob has described the situation perfectly. Here is the announcement post with complete information about the causes of this unfortunate problem and the solution we have implemented:
https://pixinsight.com/forum/index.php?topic=14219.0
-
if i remember correctly, there was always a difference between DCRAW's decoding of CR2 files and libRAW from the beginning, right?
meaning if the master bias was made long ago against a pixinsight using DCRAW instead of libRAW, this could also account for the geometry difference.
rob
-
if the master bias was made long ago against a pixinsight using DCRAW instead of libRAW, this could also account for the geometry difference.
Unlikely, but that may happen too. In such case we have no solution...
-
Thank you, Juan.
This is now working again.
Rich