Strange stars cross mark on Luminance channel appearing after integration

adding correct image...sorry

Very interesting, not seen that before i think. Can you post a screenshot of your masterdark, masterbias and masterflat as well? Just to have a quick look. I'll probably ask for some of the data as well later.

OK Thanks.

Attached as requested.

Also note that I use a refractor so those artifacts are generated by the software and not by the lens. I use FLAT DARK instead of BIAS frames. Never had this issue.

Thanks, seems ok, so I'd have to see the data to check further.

Please upload some of the relevant subs to our server;

Go to https://upload.astropixelprocessor.com and use upload1 as username and upload1 as password.

Create a directory named “sebaastro-crosses” and upload in there. Thank you!

I have uploaded 7 original light frames (from different sessions).

One note that could help to narrow down the root cause: the integration is a 5 sessions (nights) but for the first session i didn't have any luminance but only RGBHa frames. This is the only thing different I did this time from previous integrations i have made with APP.

Would you mind also uploading your master calibration files?  Forgot to mention that sorry.

No problem. Just finished to upload those as well.

Hi Sebastiano,

I know this thread is from a few years ago, but I hope you see this ....

Did you ever get to the bottom of your 'star cross' problem?  I've suddenly also got the same issue, been using APP successfully for a number of years, not seen this before!  Took R,G,B,Ha & OIII subs and it seems to only be the NB subs affected with exactly the same 'crosses' as yours.

Once I realised it was the NB subs causing the issue, I tried to reprocess them again separately (numerous times, trying  different settings each time) but unable to eradicate the problem no matter what I tried.  Re-did my calibration masters to make sure ... still the same.

Any pointers you could give me as to where to look to resolve?  Presuming you did manage to resolve this of course!

Many thanks,

Richard

Hi Richard,

Unfortunately I never got APP to follow up for a fix. At that time I was also using Pixinsight and trying out APP. Because of this issue I temporary stopped using APP and continued my journey with PI without looking back.

Perhaps after this message APP will address it and see what happens when issues are not looked ... customers migrates to different applications.

Best of luck.

Seba 

Hi Seba,

Thank you very much for responding so promptly.

So - are you saying that you processed exactly the same data with PI and that worked OK ....?

Many thanks,

Richard

That's correct. No issues with PI using the same data. The issue is caused by APP.

Seba

That's a blow, because up to now I've found APP excellent in all other respects (for me).  I may need to consider my options as well in that case!

Thanks again,

Richard

Update!!

After some further checks, I've actually now discovered that my latest darks are no good - exposure was correct but it seems I took them with incorrect camera temperature!

I've now processed the same data again using some older darks and the strange 'cross' anomalies are no longer present.  The 'faulty' darks initially looked good when just viewing them, it was only looking at the metadata that revealed my mistake!

I'd rotated my camera slightly a while ago, which is why I needed to do new cal subs at the time ....... doh!

So, not sure if this may have been your original problem Seba, but looks like for me this was down to my error in producing the darks.

You may wish to go back and try APP again yourself with some new (correct) cal files?  May or may not be your issue of course, but perhaps worth a try?

All the best,

Richard

Hi Richard @richbandit and @sebaastro,

We never have this issue occur in the test data that we have so something odd is causing those artefacts. And I really would like to know that is causing it so I can prevent it from happening to other APP users.

Richard, can you share the data on which this occured for you ?

Sebastiano, please accept my apologies for not fixing this issue back then... It looks bad and I fully agree, this should have been dealt with back then.

Richard, please upload the data here on which it happens:

https://upload.astropixelprocessor.com/

username: uploadData

password: uploadTestData

And make a folder like: RichardLewis-star-artefacts

where you can upload the data and let me know once done, thanks !

Mabula

Hi Mabula,

Thanks you so much for responding!

I'll be happy to share my data with you if that may help - I'll upload my files shortly.

In the meantime, since my last post a closer inspection of my OIII stacks still revealed some odd artefacts around stars, similar (but not so bad) as before.  Having checked everything once again I've now discovered another error on my part which I believe now definitely confirms what happened!  What I've discovered in total is this:

1.  As previously mentioned, my OIII darks were acquired using 0 degrees C camera temperature instead of my normal -10 degrees (which was correctly set for the actual subs)

2.  All of the OIII subs were inadvertently captured at 0 gain, instead of my normal 'unity gain' of 111 (ASI183mm Pro).  I think I simply modified a previous imaging sequence but I must have not paid enough attention and forgot to check/adjust the gain setting!!

It therefore looks like these 2 combined errors on my part were no doubt the cause of my 'problem'. Unfortunately, due the weather here being rubbish I've not had the opportunity to re-take the OIII data yet (and I need more overall data anyway - imaging PuWE1 nebula, which is pretty faint).

Anyway, I'll confirm in a few minutes that upload done.

Best regards,

Richard

Hi Mabula,

My OIII subs plus config files, uploaded as requested ..... have fun!

Thanks,

Richard

Hi Richard,

Thank you very much for uploading your data.

I have analysed the data and I see 2 big issues with the dataset in general:

1) the lights have zero offset. This is a major problem. The lights should never have a 0 offset. This will cause important data to be lost about the dark current of your sensor, so dark subtraction will never work properly... thus flats can never be applied properly:

The fits metadata shows a sensor offset of 0, and we can also see it from the histogram. Even with 180sec exposure, the O3 data is still clipping on the Black = 0 ADU level. This is something that must be avoided at all cost to be able to produce good data.

2) The MasterDark is not compatible with your lights like you indicated already. But not only the gain is a mismatch, also the sensor offset is not compatible. The MasterDark clearly has a non-zero offset:

Both issue 1 and 2 will have the effect that you will get very bad data calibration and thus unwanted results. If you can fix both issues, the results will be much better for sure 😉 !

Now, to address the star artefact issue of this topic, I have created an integration with the masters supplied and I do not see these artefacts. Yes, the overal integration is not good, because of the 2 mentioned issues, but those weird star artefacts are not in this result:

Zoomed in, the stars look fine without those artefacts:

So maybe something went wrong? Maybe you saw the artefacts in another dataset?

Mabula

Hi Mabula,

Thank you so much for looking at my data.

So, not only did I leave the subs at 0 gain, but I also left them at 0 offset as well.  And, of course, I also messed up my darks.  Oh, dear!

I think it's fair to say that I well and truly screwed things up on this occasion!!  I've not had this issue previously, so my only excuse is that it was so long since my previous session (due to awful weather for long period) that I simply dived into this one at the last minute without checking and/or setting the appropriate correct parameters!

Interesting that, even with these mis-match errors, you didn't see those artefacts.  It was definitely this dataset, because it's the only session I've been able to do for months, and was my first attempt at PuWE1.  This perhaps suggests that I messed something else up as well?  Maybe I inadvertently used some of my old calibration files or something, which created another separate issue?

To satisfy my mind I think the best thing I can do now is to write this batch off as a 'bad job' and produce some new darks, flats and dark flats for my current set-up and start my imaging session from scratch, this time being very careful not to introduce any errors along the way!

Note to self - pay attention!!!

Thanks again for your response, appreciate your time and advice.

Best regards,

Richard

Hi Richard @richbandit,

You are most welcome, I guess the issue is all down to how you shot the lights, right? Definitely also happened to me in the past 🙂 The lights were gain 0 and offset 0, while the darks, flatdarks and flats were all shot with gain 112 and a certain offset. So if you would have shot the lights with the same gain and offset as all the calibration data, all would have been fine. These things can an will happen, I know for a fact... so many things need to be right in an imaging session, and sometimes the camera controls are somehow messed up, maybe due to a driver update, or capture software update...

Sorry for pointing out, that the data was a mess, but I wanted to let you know what exactly was wrong 😉 so you can have better data next time !

The star artefacts still needs to be sorted out, so if you do encounter it, please let me know and share the data, I will be most grateful.

Oh, and I do know how such artefacts can occur, it has to do with the data resampling/interpolation settings. In 6) Integrate, you have the setting of pixel interpolation with default filter lanczos-3. Depending on the data, sometimes these data interpolation filters can produce such artefacts, especially with up or down scaling the data (setting the scale to other value then 1.0). And if you disable no under/overshoot then similar artefacts can occur 😉 It might explain the issue of the topic starter @sebaastro. Maybe the no under/overshoot setting was turned off?

Mabula

Hi Mabula,

Your theory regards the artefacts is a possibility I guess, although I generally tend to leave things at default settings, which normally work out OK for me - but, who knows what I did on this occasion!

I think I know what may have happened regards the lights - I use Voyager for acquisition, and I used a previously saved LRGB sequence template which I then modified to also include Ha and OIII as well for this session.  Such new additional rows in the sequence default to '0' for gain, offset, etc. so it would look like I forgot to amend these settings correctly for the OIII subs.

Anyway, no problem, the only way to learn is by making mistakes so you can hopefully avoid making the same mistakes next time around.

Regards,

Richard