SubframeSelector Measure Subrames error

txastro

Member
Jun 18, 2017
13
0
When running the SubframeSelector Measure Subframes Routine I get this error message.

Image Calculations: 11.728 s
Star Detector: 18.406 s
4741 Star(s) detected
4716 PSF(s) fitted

[006] Measuring subframe 128 of 151
Loading image: w=6024 h=4022 n=3 RGB Float32
*** Invalid block position: attachment:8192:290742336 (line=4 offset=208)
*** Error: The XSIF file contains no readable images.
Abort on error.

This is the first time I've encountered this error. If I run the routine on only this subframe, it completes without any error. Also, I have set On error: to "Continue", yet the routine aborts. I am not getting any other error messages.

Does anyone have any ideas what might be happenning?

Thanks,
Xavier
 

pfile

PTeam Member
Nov 23, 2009
5,631
106
if you can upload the smallest subset of files that causes the problem (to google drive or similar), then perhaps Juan can look at it. otherwise it is anyone's guess.

rob
 

txastro

Member
Jun 18, 2017
13
0
I've spent a little time tonight to determine the smallest subset. Not there yet, but on two machines accessing the same files (shared folder) the error appeared on subframe 128. Later, using the laptop, I picked another group of images (CR2). It also aborted on error, although the error was different. I don't remember the number of subrames. I'm wondering if there is a memory issue or an upper limit to the number of subframes that can be measured. I will get a subset together for testing. Thanks.
 

pfile

PTeam Member
Nov 23, 2009
5,631
106
yes, i wonder if this is just an out-of-memory error. especially with CR2 files which can not be incrementally read and need to be read into memory in their entirety...
 

txastro

Member
Jun 18, 2017
13
0
Updating the problem I encountered yestereday. It turns out there were 24 corrupt images. I ran BatchFormatConversion to compress the files prior to upload. That had an error on frame 128 of 151, just like SubframeSelector. Last night I misread the actual frame number to which 128 corresponded as #100. It was really #108. I tried to open frame #108, but it failed. So, those frames on which BatchFormatConversion failed I reran ImageCalibration, CosmeticCorrection, and Debayer. After that SubframeSelection had no problems. I don't know what happened. Each process succeeded on the correct number of frames, but somewhere outside of PixInsight's control the files were corrupted. Things are running now.

Thanks.
Xavier
 

pfile

PTeam Member
Nov 23, 2009
5,631
106
ok - thanks for the update. that actually does correlate well with the error you reported first - it indicates corrupt XISF files. however that is a pretty rare occurrence so it didn't cross my mind as a possibility. perhaps BatchFormatConversion should delete files for which it encountered an error during conversion but on the other hand it might be good to keep them around in case somehow they are salvageable.