Image integration showing old results

[Theresamarie1]

Multiple issues/bugs.

My first issue is with the integration process on a new set of data based on fixed flats to remove an un-fixable gradient (un-fixable in pixinsight DBE) in any reasonable amount of time or effort.    So after reprocessing all of my data, and running integration, I find the same integration result file based on old data..  It took the normal amount of time, so implied to me that it was running correctly.  I thought I must have picked up the wrong (old) files or something.  I checked all the files I used and even ran it twice.. same bad result file comes out with bad gradient, with input files of all correct gradients (or rather no gradient).   So how to heck do you fix this?    This is crazy if some internal files, or file pointers, stick around for days inside of your project file (I save no data files in my project file, only iconized processes) and show up as results to new input data.    I still have no idea how this is fixed?  I keep on getting the same result based on data from a few days ago, even though inputting different data!! 
 
Issue two, while pre-processing and using subframe selector, I couldn't get the graph to show... it was just a blank panel.  Nothing I did would fix it except logging out, and then not using my prior iconized process set up at all.  So I  had to start from scratch.     After getting it to work correctly I also noticed scoring on 2 items I would have thought only effected by stars, which didn't change in my subframes, have changed pretty significantly.    Would a gradient change in flats calibrating my lights effect eccentricity and Fwhm?   So multiple issues here (graph in subframe selector blank and changing results).

After playing with pixinsight, or struggling I should say, for a month or so while taking advantage of excellent tutorials (A. Block and Light Vortex), I'm beginning to feel like this is someone's partially finished science project and not a software package ready for primetime.    Documentation seems sparse and a bit cryptic (as is the interface in most cases).   I'm an engineer who likes astronomy and to take photos.  Software is a necessary evil, not something I want to be immersed in, nor do I want to be an expert in the science behind pixel processing.   Many I know feel the same.    I have been a long time user of photoshop and other software packages that I've used for my images but thought I'd give this a try since I'm going into mono imaging.    I don't know how long I'll stick it out with your software.    It's possibly too much of a chore to use.

[pfile]

i don't know how you actually re-ran the calibration, but by default most processes don't overwrite existing files, so you may have ended up with a bunch of _1.xisf files (or even _2.xisf) with the latest calibrations and didn't point to those when re-integrating.

nothing is sticking around in the project file or the internal state of PI. in 10+ years as a user i've never once seen anything like that.

one possibility is that in fact you are integrating the new files; sometimes you can't see gradients in individual subs because the gradients are very weak... but once you integrate they show up clear as day due to the higher SNR in the integration.

you happened to come along inbetween major versions of PI. it's clear that there are some issues with windows + the latest version, but apparently juan has not been able to reproduce these problems. i think the prior version of PI is available on the distribution server and if you are using windows it might make sense to revert back to that version. i have no doubt that in time juan will resolve the existing issues with the latest version.

while it's somewhat cathartic to vent as you have, a better strategy in this forum would be to post the image and ask for help. for instance, i have solved some ridiculous gradients with DBE (including those caused by bad flats and reflections inside the OTA) and so could probably at least tell you whether or not your original problem could have been solved with DBE if i were able to see your original integration.

rob

[Theresamarie1]

You are very condescending and not very aligned on how to accomplish customer service, if indeed you are someone from the company.   There were no gradients in my new data, and I'm not a newbie  in digital imaging.    As I said,  running the program with new data, created the same old result. I did this several times.  What I had to do to fix it, with the same old data, is not recreate or change my data, but to stop using my saved iconized process, and restart the process from scratch (just opening a new process window).    And without any more issues, it created the correct result and did it several times in a row with all my different color files.   I have no idea what's going on and honestly only want things to work.    I've experienced this kind of behavior in at least one other program as well, but don't recall the one and I'm not likely to take my limited free time to put in constant bug reports.. this stuff should be tested better.

I'm not sure what to provide you other than ALL Of the data that went into the program, which is Gigabytes.. just showing the result, which I did save, would be meaningless.     

Cathartic ,maybe.. but I spent hours and hours on this stuff is disappointing given I had high expectations.     Especially with the people saying this is a real professional program for professionals.. I guess I expected more.    My take is you've created a monster that may be too unwieldy to manage.     I have a high tolerence, but it's getting to it's limit.

[pfile]

i am not part of this company; i am simply a longtime user. i have no way of knowing how much experience you have with imaging, but i had to venture a guess. since the problem you had is extremely unusual (as mentioned i've never seen or heard of anyone with your 'bug' before), i had to assume you were either a newbie at computers or at astroimaging, or both.  hence my suggestion about the gradients; not sure why you would try to solve a calibration problem with DBE instead of fixing the underlying problem, so i assumed you had some kind of gradient in the data.

anyway all i can tell you is what i've now said twice; you have made some kind of mistake in understanding how PI works with respect to saved process icons. multiple times a day i clear out the file list from a saved process and add different files, and multiple times a day i overwrite files on disk and re-run imageintegration with no problems.

anyway *i* haven't taken PI anywhere. you need to take up your complaint with Juan.

rob

[Andres.Pozo]

You are very condescending[...]I have a high tolerence, but it's getting to it's limit.

Wow, that's how you respond to someone who tries to help... How did he dare? We should keep the people trying to help outside this forum!!!

[pfile]

i guess i should not have replied; i've seen this story too many times. there's a certain type of person out there - when they can't intuitively understand PI, they immediately blame the tool and cast aspersions at juan. it's just my opinion, but the last paragraph of the original post is an indicator that nothing good can come of replying; probably everyone else here with experience is smarter than me and would not take the bait in the first place.

rob

[Niall Saunders]

And,

naturally enough, there will be some individuals who will claim the *we* are just banding together, supporting PixInsight no matter what the issue might be.

So be it - there as some PI users who have actually invested years of effort trying to get to know PI better, always trying to learn, and always trying to improve their image processing abilities.

And - isn't it funny how it is nearly always those same people who then go out of their way to try and help others, trying to break down the questions into manageable chunks for which meaningful explanations can be given.

You are very condescending[...]I have a high tolerence, but it's getting to it's limit.

No, this is not an approach that will win friends and influence people!

However, it's a New Year and the problem could still be resolved amicably all round. Why doesn't the OP reconstitute the issue and try and ask the question in another way?

[Theresamarie1]

Guys, I've gotten off on the wrong foot obviously, or two wrong feet.  My apologies to all.     I've obviously become a bit frustrated and wanted to share that frustration a bit but not take it out on any particular person, and maybe i've inadvertently done that, and it wasn't my intention.   Again, my apologies.   I appreciate the community support greatly, even though in my heart of hearts I feel if someone/company sells a product, they should be supporting it.  But maybe that would mean a license is $500 instead of $100.   

The issue I brought up, the best that I can tell, was real and it was solved as I wrote.     

I am not a newbie at image processing, or computers (I have been designing, and then managing the design of semiconductors for almost 40 years) nor am I an expert in image processing and was an early adopter of digital (canon d30).  I've been able to do things I'm pleased with in nebulosity and photoshop, using imperfect and limited data from a terrible environment for astrophotography (northeastern united states).    But I am a new comer to pixinsight.  Although I have invested time in both free and paid for tutorials I don't feel I am making progress consistent with my time investments.   Hiccups like the one I mentioned are short term irritations, but the general usage of some items has been a tough experience (for me, anything to do with background neutralization).   I've never spent so much time fiddling with measurements points to get results that are for the most part no better than what I started with.  In any case, I will post some problems that I can't solve and see what I can learn from the experts.   

So sorry to those I may have offended due to my frustration coming through.

[pfile]

 i don’t see how anything could be as frustrating as allegro, or magma, or DC so if you have fought with those tools then you should be more than well positioned to master PI.

rob

[Theresamarie1]

Thank you for any understanding.

So you know a little about our business.     I have to admit in recent history, I've been more management than execution and design of circuits and/or more often chips.    So Design Compiler is the tool of choice for most of our customers, although our own shop has used Magma extensively.
Last i had to actually touch tools, was when I was managing development of place and route software..   

Terri

[Niall Saunders]

Terri,

Contrary to perceptions, most Forum members here are extremely thick-skinned, and therefore nearly impervious to name-calling and accusations (except me - I am too thick-headed to be thick-skinned  ). That said, we will (nearly) always defend the product - after all, we have grown with the product, and many can class Juan Conejero as a 'good friend'.

Rob is certainly giving you good advice, and knows how difficult it can sometimes be to get your head around the UI and methods of operation of PI.

Just persevere - that's what we programmers are good at!!