Problem with Alignment

Hi Uri,

First I process the Ha and get a nice stacked image. Secondly I load all the OIII data and also add the Ha image which was the reference for the Ha frames, I do this to ensure proper alignment of all stacked images. I make this Ha frame as the reference frame but unchceck it before the final Integration. All goes well except that the stacked OIII is not aligned to the Ha, seemingly ignoring the Ha reference frame. Also, the stacked images are not same size.

Excellent, the solution is to set the integration composition mode to "reference" instead of "full" in 6).

The "full" composition mode will include all pixels of all frames that are loaded and that are registered to each other. The resulting field of view will always be larger than the field of view of the reference frame.

The "reference" composition mode will only use the field of view of your reference frame. This will ensure that the Ha, OIII and SII integrations will have the same image dimensions and that they are registered to each other using the same Ha reference frame.

Another solution can be accomplished with the "full" composition mode: load all your calibrated Ha, OIII & SII frames and register them all at once. Save the calibrated and registered frames. Then proceed as normal per channel (load registered frames, analyse stars, register, normalize, integrate) , but use in 4) register, the "no registration" mode. Then no pixels are lost outside the initial reference frame, and you will have a larger field of view to work with, at least larger than the H-alpha reference frame. Sometimes this is usefull, especially with the LNC and MBB techniques enabled, because they will remove stack artefacts leaving you with more data to play with.

For stacking the 3 NB sets, I want to reuse the Master dark and Master bias created from the first process, can I simply load the master dark/bias for APP to recognize they are Masters ?

Yes, you can load APP master calibration frames with any of the load buttons, APP will recognize that they are master frames from the FITS metadata 😉 

Are you using both a Master Bias and a Master Dark for calibration of your lights? If so, be aware that that usually is a suboptimal calibration path. You will subtract 2 different masters from you lights, giving 2 times noise injection. And realise that the bias signal and pedestal are always  included in any frame, so also in your darks. If you need dark frame subtraction due to dark signal patterns then it's better to just use a Masterdark without a masterbias.

Let me know if this helps and if all of this is clear 😉

Kind regards,

Mabula

Hi Mabula

I am getting a little confused about the optimum way to calibrate.  In previous videos (for example your video of my Pelican data) you load in all the frames - bias, darks and flats.   You first create the BPM.  Then you calibrated using the BPM, the masterbias, the darks and flats.  The master dark that is created is calibrated with the master bias.

Are you suggesting that we should now do away with all bias frames, since the bias signal will be subtracted anyway if the darks are not bias subtracted?

Steve    

Hi Steve,

There are a lot of ways to calibrate data and depending on the sensor characteristics there will be a more optimal way of calibrating the data. It also depends on whether the photographer has a setup that can fully dither.

Maybe it's best to open a separate topic for this ?

In this particular case:

If you need to use dark frame subtraction for light calibration (possible reasons are amp glow, or fixed pattern noise in the dark current), why use a masterbias as well?

I mentioned this, because using both a masterdark and a masterbias in light frame calibration is complicating things, won't give you better results and possibly even introduces problems with accurate dark frame and flat-field calibration. I am not saying it can't or shouldn't be done, but I am saying it does complicates things which isn't needed I think.

Furthermore, each time that you subtract a calibration frame from data, each time noise is injected. If you subtract the masterbias from a masterdark then the masterbias noise is injected into the masterdark. And then you subtract the masterdark and masterbias from the light, the masterbias noise is now injected into the light, was already injected in the masterdark, and thus is again injected into the light by the masterdark... so it does complicate things, especially since the bias signal and pedestal is simply in each dark. 

Kind regards,

Mabula

Hi Mabula,

Thank you for clarifying the matter of using the Reference frame, I will practice more with your solutions above regarding alignment and let you know if I encounter any problems.

 I think I understand why the master bias is unnecessary when using master dark, this is how i always done it with DSS. 

Again, I want to praise your fantastic support and very clear responses

Regards

Uri

It sounds as if the program should assess what calibration files have been loaded and are available to it (check/uncheck) and offer a suggestion on how to best proceed?

Posted by: Uri Abraham

Hi Mabula,

Thank you for clarifying the matter of using the Reference frame, I will practice more with your solutions above regarding alignment and let you know if I encounter any problems.

 I think I understand why the master bias is unnecessary when using master dark, this is how i always done it with DSS. 

Again, I want to praise your fantastic support and very clear responses

Regards

Uri

Thank you very much Uri, and you're most welcome 😉

Do let me know if you succeed with registering your Multi-channel data. (And be aware that I intend to implement more automatic support for Multi-channel processing..)

Kind regards,

Mabula

Posted by: xsnrg

It sounds as if the program should assess what calibration files have been loaded and are available to it (check/uncheck) and offer a suggestion on how to best proceed?

Hi xsnrg,

In my priorities list i have added a smarter calibration engine, so this is scheduled to be improved just like you suggest. Based on available calibration data, the application should present the user with possible calibration paths together with it's PROS and CONS, so the user will be able to make a well-informed decision and learn as well from it 😉

Mabula

I have a batch of 1547 calibrated images which I want to align for star photometry. I can do this in PixInsight, but the result isn't very good. Can I do this in AstroPixelProcessor and, if so, how would I go about it?

Hi John, just go to tab 4 "Register" and go way down. There you can click on "save registered frames" i think thats what you want to have. Maybe you have first to "start registration"... ( I never used that before)

Hi,

Here are my notes on registering integrated files together. I assume the workflow will work for files that have only been calibrated. Some of the settings may not be needed for star photometry purposes.

In APP: Close and restart.

Tab 0 as usual (set pattern to RGGB if lights were from OSC camera)

Tab 1 Load

Auto-detect Masters & Integrations box must be checked.

Load the integrated files/channels of the image as light frames.

            Click yes in the pop up.

            Don’t load calibration files (they are already calibrated).

Tab 2 Calibrate

Don’t need calibration because all files are already calibrated

Tab 3 Analyse Stars

May need to increase enabled automatic #stars target to 2500 or higher.

                        Especially with short focal length and a lot of stars.

Click analyse stars button.

Tab 4 Register:

            Option to manually set a different reference frame than one automatically selected

                        Inspect the files and select the one to use.

            Enable use dynamic distortion correction

            Disable same camera and optics

Best to go on to Tab 5 and normalize the registered files before saving.

If instead you want the option to save the registered files:

Scroll down and click start registration button (can check results graph).

They are in the file window as other/processed with -reg postfix.

Scroll down, set the Save Directory and click Save button.

Tab 5 Normalize is recommended before saving the registered files

            Defaults are ok:

(Other than maybe disable Neutralize background as recommended by StarTools).

Click normalize lights button.

They are in the file window as other/processed with -reg ___??? postfix.

            Save those saved normalized (and registered) files.

Scroll down to Save normalized lights section and leave settings at defaults:

Click to set the Save Directory

Click Save normalized frames button.

These saved files are now aligned for use in APP Combine RGB Tool or in StarTools.