[Solved] Java error during normalization

Hi,

there are many reports of Java errors in this latest version of APP if you look through the forum. I assume Mabula is aware & working on a fix, I hope so. 😉 

I would imagine that you have just found this error too & it is not something that you have done incorrectly.

Regards..,

If you could provide us with the exact error (maybe a screenshot), that would be handy. Also what kind of data and steps were involved. Thanks! 

ps. Mabula is abroad at the moment, APP is getting so successful he's getting busier. I did notify him about these type of errors, if he has access to his laptop he will probably have a look.

Think I just found my error.

In reviewing file list I notice the Reference frame was deselected. I selected it and the normalization proceeded.

In the end it was my error but there is an opportunity for error checking at this point.

Thanks for your help.

That's very interesting, is this repeatable? If so, it should be a relative easy bug to squash.

Yes, this is repeatable. If you attempt normalization with the reference frame unchecked, it provokes and exception error.

Excellent! Let me tag Mabula (@mabula-admin) so he can have a closer look. Might take a while as he's abroad right now.

I just got this error when I'm combining two different fields of view where previously I've used the technique of running through calibrating all my images from one scope, but including a frame from another scope - the idea that the end result is registered according to the same orientation to allow combining. 

No matter the combination of selecting, selecting ref frame, unselecting things - I get Java errors on Normalisation.

Previously I made notes on this process and worked great. I assume something changed with Normalisation between 1.073 and 1.074. I tried 1.074 and 1.074.1.

Ok, check (@Mabula-admin). This suddenly seems to be a very specific error that crops up with very specific data, Mabula already told me this should be possible to solve as it points to the data processing part of the program. I hope he gets to it fast.

Hi @dominique-mackenzie, @1cm69, @vincent-mod, @itarchitectkev,

DMacK, Thank you for reporting the issue, I will test today the issue with the unselected reference frame in normalization. Do you know by any chance if this only happens in multi-session mode ?

And I will do additional tests to make sure there are no more such bugs in the normalization engine.

If any of you have a workflow that always gives a bug, please describe it as precise as possible, I will then be able to sort this out very quickly 😉

Apologies for the long delay in my response 😉

Kind regards,

Mabula

Sorry. We did not do any other testing.

Hi @dominique-mackenzie, @1cm69, @vincent-mod, @itarchitectkev,

I have located the source of this bug and it is now fixed. APP 1.075 will contain this fix 😉

Previously I made notes on this process and worked great. I assume something changed with Normalisation between 1.073 and 1.074. I tried 1.074 and 1.074.1.

Indeed, this actually is a regression bug caused by optimizing star analysis in 1.074.1,  the star analysis processing in 1.074.1 is done now in floats instead of doubles, and the normalization engine wasn't updated properly for the case of an unselected reference frame. APP will automatically select another reference frame for normalization if the reference for registration is deselected before starting normalization and to be able to do that it needs to read the analytical results from star analysis...

Kind regards,

Mabula