The Astro Pixel Processor Group. The group is open for everybody, downloads are available: Download APP.
Java error with registration
May 25, 2017 at 22:33 #15483
van OverzeeParticipantposts: 219
By registration of the lights of IC 1396 this Java error appeared.
This happened before.
I just click it away and the process continues.
It looks like if it has no influence on the stacking process.
Any idea why this error sometimes appears?
HansMay 25, 2017 at 22:42 #15485
Moved from the user wall to the APP Beta Group.May 26, 2017 at 12:09 #15492
Thank you for reporting this, thanks to @keesscherer, I just became aware of where this error originates from. I’ll fix it today. The error is annoying , but it will not affect your end result like you have noticed.
I’ll make sure this doesn’t happen again in the next version.
June 3, 2017 at 11:19 #15639
- 2 people like this.
With the latest Beta I’m getting frequent Java errors during registration. I’m using APP to calibrate and correct my comet frames, use PI to do a comet-alignment and then integrate again using APP. Same for the frames I corrected in PI (stars only, substracting the comet) and then integrating in APP. Here’s a screenshot, the error pops up just after the memory to file mapper closed message. I clicked the error away and no integration was done.June 3, 2017 at 12:07 #15642
Here’s the error, now in the beginning, seems a bit random. APP stops after clicking OK.June 3, 2017 at 13:18 #15644
Hi Vincent, this is not during registration, but integration ?
And it’s not a null pointer, so it’s a completely different bug.
Can you describe what you were doing and the settings that you used in 3) to 6) in a seperate thread?
Or send me those frames that you are trying to integrate?
From the console, we can see that integration is aborted at the first pixel block.
What bit depth has the data?
MabulaJune 3, 2017 at 13:27 #15647
If you can send me some of those frames and if you can describe the integration settings that you are using we should be able to solve this quickly.
MabulaJune 3, 2017 at 13:37 #15654
Bit-depth is 32-bit. I was trying an integration on a set of photo’s, in which PI substracted the comet. These photo’s were already registered and normalized in APP prior to PI processing. In App I then switched off registration and tried integrating with LNC 1, Windsor Sigma Clipping at 1.0 and 3 iterations. I was just trying some rather extreme settings as I want to get rid of as much background as possible to try and combine the comet and stars later.June 3, 2017 at 13:41 #15656
Okay, I’ll check it, thanks !June 3, 2017 at 14:41 #15657
“Windsor Sigma Clipping at 1.0 and 3 iterations. I was just trying some rather extreme settings as I want to get rid of as much background as possible to try and combine the comet and stars later.”
Winsow clipping 3 iterations kappa 1? Why a kappa of 1? This will reject 1/3 of all data, so good data as well?
Outlier rejection filters will not get rid of background, it will get rid of outliers. Depending on the specific dataset this can have all sorts of unwanted effects if you set the filters too aggresively.
Remove backgroung should be done with the remove light pollution tool I would suggest 😉
(In the case of more than 30-40 frames, you probably should never set the kappa below 2 I think for outlier rejection. Above 3 is also pretty useless if you have plenty of frames)June 3, 2017 at 14:58 #15659
I used this manual which uses extreme values as well, maybe cause his data was just bad to begin with. 🙂 But great case for that final bug! I’ll reprocess again.June 3, 2017 at 15:30 #15667
Ok, so the core problem is found, I went through the APP process, calibrating, registering and normalizing the frames. Then pressed “Save normalized frames” and went to PI. However, you need to select “register frames” when saving the normlized ones. It was a bit confusing, I didn’t select that as I thought I already did that. ?June 3, 2017 at 20:55 #15669
Luckily solved 😉