First processing attempt with PixInsight

[headdown]

Hello everyone,

I am currently processing my first image using PixInsight, and am stumped by a few things and was hoping to find some answers from the more experienced users here. Attached is a dropbox link to a screenshot.  On the left is about 4 hours integration of 12 minute subs, and on the right about 2 hours of 30 second subs.  Both used the same master bias and master flat, and each had their own dark master.  Master bias was 100 frames.

Is the bias visible in the 30 second integration simply because of the aggressive stretch, combined with the lower SNR compared to the 12 minute subs?

And the second concern is the red error message in the screenshot telling me I am trying to use more than one registration image, which I am not.  I intend to use the HDR tool, and understand that I must first register the two images, but can't get past this point.

Amazing software...just scratching the surface so far and I am very impressed.  Thanks for any help!

Dean Fournier
Ottawa, Canada

https://www.dropbox.com/s/c62il3pimt0ulma/Screen%20Shot%202016-02-03%20at%207.12.56%20AM.png?dl=0

[pfile]

oh - i think what's going on is that the xisf in question has 3 embedded images - the high and low rejection map from integration, and the image itself. this is what happens when you use BPP to integrate the final image. i suppose if you opened it and then just saved out the image and used that as the reference, everything would work out OK. but the integration warning in BPP still applies - you'll probably get better results by integrating the registered lights produced by the script, since you can fine-tune the rejection parameters. in that case you'll have an xisf with a single image in it to start with, as the rejection maps come up as separate windows in your workspace.

unfortunately the banding is probably not really the bias, but actual banding, which a lot of canon cameras exhibit. areas of low signal are destroyed by this banding and there's not much that can be done about it. aside from the stars, this field probably does not need HDR, and the stars are probably bright enough that if you did do HDR to recover saturated stars, it would probably work out OK. pay attention to the masks created by HDRComposition and you can see which pixels were taken from which image. obviously you don't want any of the DSO data from the 30s image to creep into the final result.

rob

[headdown]

Hi Rob,

And thanks for your reply.  And I understand about the rejection maps now that you have pointed it out.  But I used BPP to calibrate the images only.  I used Star Alignment to register the images, and Imageintegration to stack them.  And if I recall correctly, the result was the integrated image and the two reject maps.  After looking at them, I closed the reject maps and tried to use the integrated image as the registration image.  I assume from what you are saying that just closing them still leaves them imbedded in the image, and that is causing the problem?

..."you'll probably get better results by integrating the registered lights produced by the script, since you can fine-tune the rejection parameters."

Isn't that using Imageintegration?  Because that is what I did use.

Okay I just saved the 12 minute integrated image, and then reopened it, and I was then able to use it as the registration image. Problem solved.  Thank you!

I mainly wanted to use the HDR tool to try and get rid of the halo around blown out Alnitak.  And since finding out that it also can replace lost star colors, that is another reason I wanted to try it.  And even with the banding in the 30s image, I figured that the data that I may borrow from it will still be the brightest parts of the image, and should have lots of SNR to hide the banding.  And I will watch the masks to keep the dimmer 30s data out of the final image.  Thanks again for your help!

Dean

[pfile]

well since the .xisf seemed to have 3 images embedded, i assumed it had come from BPP, cause that's what it does. not sure how you ended up with a 3-part image from II because if you just save the integration, that's all that should be saved in the file... but i'm not sure of that since generally what i do is just register everything to the same calibrated subexposure; i can't remember the last time i registered to an integrated file written to disk. so it's possible there's something in the xisf of the stack that's messing up ImageIntegration...

rob

[headdown]

Hi Rob,

Being a fresh newbie to PixInsight, I thought I had to use BPP for calibration, but I have just discovered Image Calibration.  Would this be a better way to go then?  Use Image Calibration instead of BPP, Star Alignment, and then Imageintegration?

Dean

[Sean]

Welcome Dean,

My 2 cents worth - when starting with PI, I think it's best to apply the processing steps manually, one at a time, and inspect the results as you go. This gives you a better feel for what BPP is doing. In fact, I still do everything manually, although I've been using PI for years. Just personal preference.

So, in answer to your question, yes - do Image Calibration, then Star Alignment, then Image Integration.

I hope this helps,

Sean

[headdown]

Thank you Sean!

Dean