[Solved] Bug Report APP 2.0.0-beta22-windows-64bit: Encountered error in module: CombineThreeChannelFloatBufferWorker

is there any workaround or bug fix available? Channel Combination isn't any longer usable in APP... 😥

I think Mabula addressed this for beta 23, which is very close to release

Hi,

thanks.

I have found APP 2.0.0-beta23-test3 in the download area and with this it works again without crash.

Best regards,

Markus

Hi markusd112,

THank you very much for reporting this.

Yes, the issue is known and is still on our list. Another bug fix that I fixed for the beta23 version might have caused this bug as well. Can you please try with the official 2.0.0-beta23 release that we released today and let us know if the bug is completely gone?

From another APP user we got the report that this bug happened a bit randomly, sometimes the calculate would work and sometimes not. Was this also the case for you?

Mabula

Hi Mabula,

thanks. I have tried it with offical beta23 and it works now without crash here.

With the previous beta22 it crashes each time here and I tried it several times...

Best regards

Markus

Hi Markus, still no crash with beta23, we have a report from another user, that it still crashes with beta23 when he tries RGB Combine... on his data.

Mabula

hi @markusd112 Markus,

Still no crash with beta23? We have a report from another user, that it still crashes with beta23 when he tries RGB Combine... on his data.

Mabula

Just noticed this same issue in beta 23. I just tried upgrading to beta 24 and same result.

I'm trying to combine Ha and O3 , 2 FITS images. I've tried cropping the stacked images to be data only (no black borders) but no change in behavior.

Full error below:

Encountered error in module:
CombineTwoChannelFloatBufferWorker

Error message:
java.lang.ArrayIndexOutOfBoundsException: Index 73105011 out of bounds for length 73105011

Cause:
Index 73105011 out of bounds for length 73105011
eN.a(Unknown Source)
eN.doInBackground(Unknown Source)
java.desktop/javax.swing.SwingWorker$1.call(Unknown Source)
java.base/java.util.concurrent.FutureTask.run(Unknown Source)

-edit1 Same occurs also in 3 channel combine. 

-edit2 Tried same images in beta19, works fine.

Hi Zak, @zakky2k

Thank you very much for reporting this. This bug is at the top of our list and we need to fix it a.s.a.p. !

But I can not reproduce it. Can you share your data with me by uploading it here:

https://upload.astropixelprocessor.com/

username: uploadData

password: uploadTestData

Please make a folder like ZakForeman-RGBCOmbine-ARRAYINDEX-BUG and let me know once uploaded.

Please also let me know the settings that you are having in menu 0) RAW/FITS, and what formula you choose in the RGB COmbine Tool to try to make the composite. Thanks.

As soon as this bug is fixed, I will release a fixed version for sure.

Mabula

Hi,

"sorry", but with my data here it works with beta23 without crash. Maybe it depends on some settings or on the way you choose your data and which parameters you set...

Markus

@mabula-admin Thanks Mabula! I've uploaded the problematic images into the suggested folder  ZakForeman-RGBCOmbine-ARRAYINDEX-BUG. 

Regarding your questions, the raw sub-images where loaded using the following 0) RAW/Fits settings then processed individually to create the uploaded sample integrations:

Pattern: RGGB

Algorithm:

1) Ha-OIII Extract OIII for the OIII integration

2) Ha-OIII Extract Ha for the Ha Integration

3) Ha-OIII  Mono for the mono integration

All of them used the below options selected / checked:

Force Bayer/X-Trans CFA

Use camera white balance

Use camera color matrix

Then once completed, the output images were cropped (using the batch crop tool) and light pollution removed before loaded into the combine RGB tool.

Initially I tried with uncropped images but maybe the black areas in the integration output frames might be causing a problem. 

I tried different formulae in the combine tool but all give the same error, either:

CombineTwoChannelFloatBufferWorker (when only 2 channels loaded like HOO 1 or 2)

or

CombineThreeChannelFloatBufferWorker (when 3 channels loaded like SHO, RGB etc) when trying out using the mono frame as a luminance layer. 

As I mentioned, I re-installed an earlier version (beta 19) and was able to process successfully with the same images.

Hope this helps!

Zak

Hi Zak @zakky2k,

Thank you very much for all the details, that should really help to solve this. I will test your data now and will let you know if I can reproduce the error now and fix it 😉

Mabula

@zakky2k Ah finally, I can now reproduce the error! I should be able to fix this now a.s.a.p. 🙂 Thanks a lot Zak

Hi all , this is issue will be solved in the next release 😉

https://www.astropixelprocessor.com/community/release-information/astro-pixel-processor-2-0-0-beta25-release-notes-work-in-progress/

It was caused by several things :

1) Some tools in APP would save FITS files with CFA tag = "yes" which was a bug, this is solved in all tools now

2) When running the RGB Combine tool, it should disregard the 0) RAW/FITS settings which it wasn't. So setting 0) RAW/FITS to force CFA and Ha-O3 mono for instance, would still treat the data in the RGB COmbine tool as CFA allthough it was already debayered.

Thank you @zakky2k, your detailed response really helped here.

I will try to release a fixed version as soon as possible.

Mabula

Hi all,

Please try 2.0.0-beta25 which has been released with the fix for this issue.

https://www.astropixelprocessor.com/downloads/

Mabula

@mabula-admin Perfect! Issue fixed in Beta25 as far as I can tell 🙂

Excellent @zakky2k !