Workflow suggestions for integrating galaxy OSC subs and monochrome L and Ha subs

@bobbeanbagsgmail-com Hi Robert, currently it is not possible in APP do process RGB and mono data in a single integration run, especially if in tab 0 Force Debayer is enabled. So processing the RGB data in a single run was a good idea. L and Ha data can be processed together as long as you enabled multi-filter processing in tab 1. Then you can load the L and Ha images as well as L and Ha flats. APP should be able to recognize the filters automatically but it may be possible that you need to indicate those yourself.

Once RGB has been processed in the first run and L and Ha in the second, you can split the RGB channels into separate R, G and B channels. Then you need to register the images of the five channels against each other by loading all of them in tab 1 and then going to tab 4 and execute the regitration process. Then also in tab 4 you can save the registered files.

Then you can go to tab 9 and combine the registered frames. I would do as many steps as possible (light pollution removal for sure) in APP as possible. Not sure about deconvolution though since I don't know what software you will use for that. Color calibration in the RGB image before splitting the channels is a good idea as well.

@wvreeven This seems like a simpler answer to the question I'm asking about on another thread. Post processing the stacked L&Ha pair and OSC data by splitting the latter and then registering 5 single mono files seems natural and easy. I will now erase the 600 files I just made splitting all the subs...

@williamshaw Yes there are more workflows possible to achieve what you want. In the other thread I was merely replying to the questions you asked there. I'm glad that you agree that the workflow outlines in this thread is better.

@williamshaw This is exactly what I'm trying to do right now as well. I've got L and Ha frames from an ASI1600MM-P and colour frames from a borrowed ASI071MC-P. I've just made a large number of separate R, G and B frames and have no idea what to do with them. I'm watching the thread carefully, because I have no experience with colour frames at all!

@skogpingvin  - I decided to follow the pattern in this thread, where (I think!) separate R,G,B files are extracted AFTER the colour data has been integrated. So there are three files rather than hundreds (which is what another thread is about). The trick appears to be to take integrated files for RGB,L,Ha etc and run them through a subset of the steps to first split the RGB into three mono channels using the split channels option, and then just register the resulting five files (in my case). I have literally just rerun the two mono integrations and the colour integration and figuring which subset of steps I need to take those three files (where only the L and Ha are registered) and produce five registered files. I'll try to write down exactly what works, if  I get it sorted....

@skogpingvin @williamshaw

The simplest way to do this, is to load ALL data at once into two sessions. So, for example, session 1 RGB and session 2 L and Ha. You can swap the sessions for RGB and L/Ha if you want. That is not relevant as long as RGB and L/Ha are in separate sessions. Then load the corresponding calibration data (darks, dark flats, flats) for the RGB data into session 1 and for the L and Ha data into session 2. Again, swap the sessions here if you decided to do so when you loaded the RGB and L/Ha data. When loading the data, check and verify that the lights and flats are assigned to the correct channels. Then integrate in tab 6. That should give you 3 files: the integrated RGB, the integrated L and the integrated Ha data. Since these were processed together, they will already be registered w.r.t. each other!

Next you load the integrated RGB data as light in tab 1 and do not load any calibration data. Go to tab 2 and click the calibrate button. Then further down in tab 2 tick the "split channels" option and click "SAVE (calibrated) Light frames". This will save 3 files: one each for the R, G and B channels.

Finally, you can taken the R, G, B, L and Ha fles and load them in "combine RGB" in the Tools tab and combine them as you please.

It becomes more complicated if you need to enable "force CFA" in tab 0. You should only need this in case your imaging software doesn't store the CFA or Bayer pattern in the FITS headers of the RGB files. This is a complication because APP will apply debayering on all files, the mono L and Ha files included, which in this case is not what you want. Let me know if you need help with that.

@wvreeven Ah so the separate session preserves the registration but does not cause the normalisation to fall over. I will have to return that from scratch later on. Right now I am just trying to register the outputs of the Ha,L pair and RGB with the post process split files and calibrate idea. 

I do not need to force CFA any more - there was a time with APP was reading files from my Hypercam 26C grabbed in Altair Capture with the wrong Bayer and I had to force it, but I do not need to do that now. 

Thanks for this. @skogpingvin this is a much better plan, especially if you are starting afresh. Having don the integrations separately I'm going to finish my current plan though.

OK so if you have already done your colour and mono integrations separately, and wish to merge them after all that, then I got the following to work. I think the plan by @wvreeven is more elegant and if you are starting afresh by far the easiest, but if you've eg captured some colour data way back and haven't kept the lights, and maybe shoot some mono later, this post process will work to line everything up. It might be possible to simplify these steps of course!

This example has inputs data that has already been stacked in APP, and takes the form of a a colour RGB file, and Ha and Lum mono files.  

Tab 1: I load the three as light files, specifying the channels

Tab2: No calibration, just scroll down to the bottom where it says save calibrated frames and tick align and split channels. This takes a little while and I get five files with names beginning “other processed” Blue (Green, Hydro, Lum, Red)

At this stage it feels like you kinda start again and fool APP into thinking those five files are fresh lights, as for some reason it does not recognise them as lights. So I untick all the files and then reload the five I’ve made as lights, assign the channels as I go.

Then I jump to tab3, run analyse stars and then tab 4 register, clicking on start registration and then saving the output. You can then do what you like with the five registered files.

@williamshaw OK, I'll try this soon - I'll be leaving for work soon.

Just to recap though - in step 0 I can leave force Bayer CFA unchecked (I use SGP and the camera is an ASI071 so it'll get the right pattern).

In step 1 I load using multifilter and multisession. Session 1 is the Ha/L for the ASI1600 and session 2 is the "RGB" filter for the ASI071. Load lights dark bias and flats as normal, taking care with filters and sessions.

Then I jump straight to step 6 (because I'll come back and calibrate etc. later, right?). Sorry if I'm a bit behind the learning curve here!

@skogpingvin Yes that is all correct. When you jump to step 6 and click integrate, APP will do all intermediate steps automatically for you. You need to return to step 2 later only to split the R, G and B channels of the RGB integration result. When doing that no actual calibration will take place. 

@wvreeven thanks for that. It's off and running, but it's time for me to run as well. Thanks for your help so far - I'll get back to this later!

Posted by: @wvreeven

@skogpingvin @williamshaw

Next you load the integrated RGB data as light in tab 1 and do not load any calibration data. Go to tab 2 and click the calibrate button. Then further down in tab 2 tick the "split channels" option and click "SAVE (calibrated) Light frames". This will save 3 files: one each for the R, G and B channels.

 

I'm having trouble with this bit.I've got the Ha and the L stacks fine, and the integrated RGB. Now I need to split the integrated RGB stack into R, G and B. If I add this file as a light (with no darks, flats or bias), then go to step 2 and tick "split channels", then click "save calibrated frames" I simply get the single file again, not three. What have I missed?

@skogpingvin You need to click the "create Masters & assign to Lights" button first. Once done, you can tick the "split channels" checkbox and click the "SAVE (calibrated) Light frames" button.

I'm still doing something wrong.

I've cleared all the files out of APP, then loaded one single file as a light frame. It's called St-avg-15300.0s-AR_LN_8.0_4.0_8-x_1.0_LZ3-NS-full-qua-mult-sc_BWMV_nor-AAD-RE-noMBB-RGB-session_2.fits and it's 32 bit grey. When I load it I get a warning that this is a previous integration of light frames (which it is). I can just clear this.

Then I go to tab 2 (calibrate) and click "create masters and assign to lights" nothing appears to happen (apart from the bong sound). Then I tick "split channels" and click "save calibrated frames", and it again creates a single file.

I have tried with and without multi channel/filter processing.

@skogpingvin What version of APP are you using? The name of the RGB integrated image seems to suggest a rather old version of APP. Can you make sure to upgrade to APP 1.082 and see if it works then?

@wvreeven  I'm on 1.077. I'll upgrade and get back to you. Thanks for your help!

@wvreeven OK, I'm on the latest version now (1.082).

There seems to be no difference. The RGB stack is now named NGC1365-RGB-session_2.fits and the single file after the attempt to split it is called NGC1365-RGB-session_2-RGB-session_2-cal-ch1.fits.

(The mono Ha and L information from the ASI1600 has gone into session 1)

I notice from the FITS header for NGC1365-RGB-session_2.fits includes this:

SOFTWARE= 'Astro Pixel Processor by Aries Productions' / software
VERSION = '1.082 ' / Astro Pixel Processor version
INTEGRAT= 'Integration' / integration of light frames
CFAIMAGE= 'no ' / Color Filter Array pattern

If there is no CFA pattern perhaps that's what's wrong? It has no way of knowing it was RBBG.

The FITS header for the eventual (single) image NGC1365-RGB-session_2-RGB-session_2-cal-ch1.fits includes something similar:

SOFTWARE= 'Astro Pixel Processor by Aries Productions' / software
VERSION = '1.082 ' / Astro Pixel Processor version
CALLIGHT= 'calibrated-Light' / calibrated light frame
INSTRUME= 'notAvailable' / instrument name
CFAIMAGE= 'no ' / Color Filter Array pattern

It's probable that I've missed a setting somewhere, that tells the system that the stack from session 1 is RGB data with RGGB Bayer pattern.

I'm going to try forcing an RGGB CFA in step 0. See what happens.

@skogpingvin CFA only applies on the RAW data and not on the processed image. Can you upload the image to

https://upload.astropixelprocessor.com/

using upload for both username and password? Please create a directory called skogpingvin_split_channels and put the image in there. Then I will have a look.

@wvreeven Thanks - I've uploaded it.

@skogpingvin I think I didn't make myself clear. I meant for you to upload NGC1365-RGB-session_2.fits. Can you do that please?

@wvreeven OK, sorry! It's there.

@skogpingvin Thanks! It looks like something went wrong while integrating the subs because this image essentially is a B/W image and doesn't contain any color. As soon as step 6 is done, the integration result is displayed in APP. In the upper right you can find the histogram of the image. In case of an RGB image the histogram is supposed to show colors but that is not the case:

Can you upload a raw light as well so I can see what that looks like?

@wvreeven

Posted by: @wvreeven

Can you upload a raw light as well so I can see what that looks like?

I can't get onto the upload site at the moment, but the sub is definitely a colour image - it's a horrible yellowy green colour!

@skogpingvin There was some maintenance at the server today. It is working now. Can you try again please?

@wvreeven OK, uploaded now. There's always a satellite!

@skogpingvin There is no Bayer information in the FITS header of your raw files. So in tab 0 you need to set the pattern to RGGB and tick "force Bayer/X-Trans CFA". However, when you do that and also load the L and Ha files, then those will be debayered as well and they shouldn't.

So your workflow will be

• Start APP
• Go to tab 0, set the pattern and check "force Bayer/X-Trans CFA"
• Go to tab 1 and load the RGB lights and calibration files.
• Go to tab 6 and click integrate
• Go back to tab 0 and uncheck "force Bayer/X-Trans CFA"
• Go to tab 1, unload all files and load the L and Ha lights and calibration files
• Go to tab 6 and click integrate
• Go back to tab 1, unload all files and load the RGB, L and Ha integration result
• Go to tab 4 and click "start registration". When done, click "save registered frames"
• Go back to tab 1, unload all files and load the registered frames
• Go to tab 2 and click "create Masters & assign to Lights". When done check "split channels" and click "SAVE (calibrated) Light frames.

@wvreeven Sorry about the delay, I was on a work trip (servicing astronomical domes - I love my job!)

I haven't gone through your last message, but just to re-cap, the five files I have are:

• NGC1365-RGB-session_2-RGB-session_2-cal-Red.fits
• NGC1365-RGB-session_2-RGB-session_2-cal-Green.fits
• NGC1365-RGB-session_2-RGB-session_2-cal-Blue.fits
• NGC1365-Hydrogen-alpha-session_1.fits
• NGC1365-Luminance-session_1.fits

The first three files are 32b grey, but the last two (Ha and L) are marked as 32b RGB, which seems wrong. This makes the combine RGB tool ask me for lots of extra information about which channel of each file is shot using which filter. Clearly the Ha and L files should be 32b grey.

I assume the workflow you have provided will short this out.

I'll have a go at this soon, but work beckons again. I just wanted to let you know I haven't forgotten you, and I'm very grateful for the help!

@skogpingvin The Ha and L files should be mono as well and not RGB. This is because "force Bayer/X-Trans CFA" was enabled. If you follow my new recipe then they will come out as mono files.

@wvreeven I've done that, and got a result! This is the text I've written for my local Astronomical Society.