first go combining Ha and Siii using pixel maths

[brrttpaul]

It was supposed to be all three but for some reason sgpro stopped and aborted , anyway managed to get 4x20mins each of Ha and Siii so thought lets try combining them, I know I have gone wrong somewhere as there are little red dots next to each star and my processing skills are not up to scratch but dont care got a colour image and that was my goal

[wimvb]

Nice!

Are the red dots 'double' stars? That would suggest tracking errors somewhere. If the stars have an ok shape, you can try registering the two master images against each other (use staralignment in PI; which I assume you have, due to the pixel math reference) before combining.

[brrttpaul]

i stacked the Ha and then the Sii in maxim using sigma clipping, then I opened them up in PI and just done the pixel maths, pretty sure I had something similar happen before when I done sigma clipping, im guessing its an ialignment thing as the seperate images dont have the double blips only when i combine them

[wimvb]

You need to use the same sub as reference for both stacks. Easy remedy: load all subs at the same time into the staralignment tool of PI. Then use image integration to do the stacking of each filter separately.

For the masters of this image, just load the Sii in staralignment, and Ha as reference. This will create a new, aligned Sii master.

[brrttpaul]

By the sounds of it then I done it all totally wrong, heres what I done, calibrated the Ha images in maxim then calibrated the Siii in maxim also, next stacked the Ha in maxim and saved, then same again to Siii, then in PI I used pixelmath and put red as Ha  and Sii as blue and green ( i think that was the sequence i was reading it in a tutorial) i also ticked create new image, this give me the image you see now done a little messing about with it in Pi and thats it

[wimvb]

That is basically correct. The only thing you missed is that the Ha and Sii are not aligned to each other. You can either align all subs with reference to one sub, or after the stacking of Ha and Sii individually, use the Ha (or Sii) as a reference for aligning the other.

When I process data from the Liverpool telescope, I use the former method: I align all subs in one go, with a certain sub as a reference. After that I combine (integrate or stack) the subs from each filter separately. Then I use channelcombination in PI to create a colour image from the different filters.

Btw, here's your image (jpeg) restacked. notice the edge on the left and top. I separated the channels (r, g, b), used the red channel as a reference. Then registered the green and blue channels, and recombined them.

Because in your original image the red channel was a little missaligned towards the lower right, the blue and green where shifted in this direction during my restacking, causing the red edge left and top.

[brrttpaul]

Ah right didnt even occur to seperate the channels, so have I got this right? seperate the RGB, then using star alignment use the red channel as a reference, then register the G and B (in star alignment) click combine and then im guessing crop it a little? Edit just opened the image in PI and seperated the RGb then using star alignment I used R as a reference and put GB in  target images clicked apply and I have a new combined image except its in greyscale lol how do i get it back to colour now?

[wimvb]

For this image, you probably have saved the master Ha and master Sii images. I only use PixInsight, so here's how it's done there:

1. load the Ha master and the Sii master into PixInsight

2. open the StarAlignment tool in PixInsight (under processes > imageregistration)

3. Load the Ha master as reference view, in the top section of StarAlignment

4. Select the Sii master by clicking on it

5. Apply StarAlignment by clicking on the square icon in the lower left corner of the StarAlignment tool

When the Sii is aligned, a new image is created called '..._registered' (where '...' is whatever the Sii master image's name is)

6. You can now close the original Sii master image, and work with the registered version as you did before; i.e. combine with the original Ha image

Here's what the tool looked like when I realigned the channels in your image

Btw, I would avoid working on astroimages when they are in jpeg format. This format is just too crude for the kind of processing you would normally do. Of course, when you only crop and image or add a signature, jpeg is ok. But for all other work, I would use fits, tiff, xisf, or a similar file format.

[brrttpaul]

ah right done that wrong as well, I still come up with a grey master though?

[wimvb]

You probably already know this, but here goes anyway:

Masters are allways gray-scale images. If you shoot mono with filters, your images will show how much light intensity is collected on the camera sensor at a certain wavelength. The wavelength (= colour) information is only provided by the filter, not by the mono camera. That's why it is called mono-chrome: one colour. Your Ha master, Oiii master, Sii master are all monochrome, or black and white images. The same goes if you use RGB filters, because you collect the intensity from only one colour (R, or G or B ) at a time.

If you take a colour image, you can extract the different channels that make up the image, and end up with three slightly different gray-scale images. It is when you pop these three images back into the channelcombination tool, that it will convert the monochrome images into rgb images, by using one image for the Red information, one for the Green information, and one for the Blue information.

If you want to do more elaborate combinations of mono images, like the Hubble palette, or HSO, or HOO, or whatever pleases you, you can do so by using PixelMath. BUT, you need to use a combination expression for EACH COLOUR CHANNEL, otherwise the final image will end up the same sort (gray-scale) as the source images. You need to uncheck the tick box 'Use a single RGB/K expression'. And to get a colour image out of it, you need to specify the Color space as 'RGB color'.

Here's a screenshot showing the settings in PixelMath:

The three images going into the expression (here called Ha, Sii, Oiii) can be combined in anyway you like, but for a colour image, there has to be an expression in at least two of the text boxes. Otherwise your image will be a monochrome image in one colour (all red, most likely)

In this example, red will be Ha, green will be the average of Ha and Sii, and blue will be the median of Sii and Oiii. This is probably going to be a funny looking image, but it shows what freedom you have in combining.

Also note that the Color space is RGB, which means that your resulting image will be a colour image, and not a mono image.

[brrttpaul]

I must admit I did not know it, only now just reading this and will have a look at it see if I can do it, thx very much EDIT (didnt have the RGB ticked )  prob a stupid question but the whiteish cloud ? what is the correct colour of that , what I mean is would doing a colour balance give the right colour?

[wimvb]

In narrowband imaging, you can't maintain colour balance. Colour balance means that you define a neutral background (no particular colour, or the same amount of red, green and blue), and a white point (ideally a star that is exactly white (red = green = blue = 1). But in narrowband, you don't have red, green and blue. Instead you only have the colour(s) from the narrowband filter(s). For Ha this is deep red. So, the 'natural' colour of an Ha image would be deep red. On the other hand, hydrogen (the 'H' in 'Ha'), also emits weakly in the blue/cyan, which is why with a dslr, most Ha targets can look purple. This blue/cyan wavelength is named Hb. Some astrophotographers add Ha to red and add a little of the Ha signal to the blue channel, just to mimic Hb.

The origin of narrow band imaging is scientific, and different colours are used to indicate the presence of Hydrogen (Ha), Oxygen (Oiii) and Sulfur (Sii). One such combination is the Hubble palette. But neither of these combinations even resembles a 'natural' colour. That's why they are called 'false colour images'. And that's also why colour balance isn't possible in narrowband imaging.

In your image, you can make the colours pretty much the way you like them to be. But to get a conventional colour scheme, or palette, you should google 'Hubble palette'

If natural colour is what you're after, you need to mix the narrow band images with RGB images. These will have a natural colour, and you use the narrow band images to enhance the 'natural' colours, to indicate the presence of H, O and S in a 'natural' looking image.

My advice: google for narrow band images of the nebula. Find out which colour scheme you like best, and create an image in this colour scheme. This will get you started, and if you want another colour scheme, you can allways create a new image from the masters.

BTW, if you plan to do most of your processing in PixInsight, you should definitely get the book 'Inside PixInsight' from Warren Keller, and also have a look at the (video) tutorials from Kayron Mercieca (lightvortexastronomy.com) and Gerald Wechselberger (http://www.werbeagentur.org/oldwexi/PixInsight/PixInsight.html)

[brrttpaul]

Thx for taking the time to respond, I will definatly get the book as for the colours it was just something I was thinking as part of it was deep red but other parts looked white so I wondered was it a differentelement that was causing this , thx once again