Filters - I really need some help/advice with this please

[Carole]

I have a manual filter wheel and narrowband filters Baader 7nm set (and another set with LRGB filters).
I have taken a set of Oiii and Sii images to go with the Ha (for the Horsehead), but my problem is I am left with a crap image when I try to combine them.

I stack each filter image individually and then I have been using either Maxim or nebulosity (or even both) to align the frames.  This is a manual procedure which requires me to click on a couple of stars.

Now one of the problems is that Oiii seems to pass more light than the Ha and Sii so it is letting more light through and thus the stars appear to be bigger (at least that's what it looks like to me).  Additionally, no matter how accurately I try to align the stars (zooming in etc) and the individual resulting images look to be aligned when I click them on and off, they do not seem to be perfectly aligned when I try to combine them, plus I have what I can only assume are thousands of hot pixels (also slightly out of alignment) from all 3 filters, resulting in this crap image attached.

Is there some better way of doing this?
Am I missing something?
Also, why am I left with so many hot pixels.  Not dithering yet, but the manual change of filter results in a minimal "shift" which I would have thought should have eradicated the Ha since I did this in two different "goes". 

Crop of resulting combination of filters:

[RobertM]

Carole,

The three filtered NB images should all overlay and be exactly the same image scale.  Are all your filters of the same set and by the same manufacturer?

Could also be that something has 'slipped' in the focuser.

Robert

[Carole]

Hi Robert, Yes they came as a set from Bern, all Baader 2" 7nm, Ha 7nm, Oiii 8.5nm (doesn't come in 7nm), Sii 8nm (doesn't come in 7nm).

I did bin the Oiii and Sii and had to double the size when I aligned them. 

I don't understand what is happening here and why all those thousands of tiny spots (? hot pixels).

Carole

[MarkS]

In my opinion:
1) The Red channel looks well focused and stacked nicely
2) The Blue channel looks less well focused abut still stacked nicely
3) The Green channel looks badly focused, so much so that the stacking didn't work - leaving smeared defocused stars.
4) It also looks like your dark subtraction has not worked very well because so many hot pixels remain.
5) It would appear that the hot pixels have been used to determine the RGB channel alignment because they align precisely whereas the stars don't.   Or, more likely, no RGB channel alignment has been achieved at all.

Mark

[Carole]

The Green channel in this case is the Oiii.  These filters are supposed to be parfocal, I did check that this was the case the first time I used them, but did not re-check them on this occasion since I had already confirmed they were parfocal.

4) It also looks like your dark subtraction has not worked very well because so many hot pixels remain.

I totally agree, and I found problems with hot pixel subtraction when I previously used DSS (This time I used nebulosity).

It would appear that the hot pixels have been used to determine the RGB channel alignment because they align precisely whereas the stars don't.

I don't see that, as everywhere there are little groups of 3 hot pixels. 

I have definitely aligned the channels in nebulosity and also did it with Maxim.  You can see the shift in alignment, and when I lay them on top of each other and click them on and off, they definitely look as though they are aligned properly, this is why i am so perplexed.

Any further suggestions for aligning the filters that I could try.

I don't know what to do about software not removing the hot pixels.

So currently a bit stuck until I can find out what is going on and how to fix it.

Carole

[JonH]

This looks almost exactly like the problem i was having with DSS!
If it were me i would register them all as one set in Iris then split them up and stack them wherever.

[MarkS]

Carole,

I don't doubt that the filters are parfocal.  But the focus probably slipped during the course of imaging.

Those groups of 3 hot pixels across the image are quite strange.  Do you have an explanation for them?  Do they appear as groups on each sub or is this a dithering pattern?

Mark

[Carole]

I have spent a good part of the day on this, and despite moving over to Nebulosity for stacking, I decided to have a try with DSS and then re-align them in Nebulosity.  I set the cosmetic preferences to remove hot pixels, and this seems to have worked.  So my conclusion is that Nebulosity seems to work better than DSS with DSLR images, but DSS seems to work better for removing hot pixels.

I still haven't managed to overcome the filter misalignment entirely, but it is much closer now.

I am adding a couple of DSS stacked and colour aligned images in narrowband to the Horsehead thread.

Many thanks for your help guys, I eradicated one problem, but still need to find out why the filter alignment procedure is not working efficiently.

Carole

[Mac]

This may sound a silly idea, but if you are getting hot pixels in the same place in each image,
and the image has not moved, why not use the hot pixels to align the frames,
Looking at your image you can see that the three layers are not aligned, by the multicoloured stars,
but you also have three hot pixels not aligned, in the shape of the three dotted triangle,

Might be worth manually aligning on the hotpixels to see if it will make a change.

Also in photo shop, if you use the ruler tool and measure between two points (stars),
It will asume thast the line you have measured is supposed to be horizontal.
If you goto rotate arbitry (cant remember which menu its in), then the angle it give is the amount it will rotate the image by
to get your line horizontal,

So if you select the three individual images and select the same two stars on each, then each RGB image can be rotated corectly
and then combined.

Mac.

[Carole]

Hi Mac, I seem to have got rid of the hot pixel problem by stacking in DSS with cosmetic hot pixel remover.  The method you suggest to align the filters with hot pixels, a) Doubt if the hot pixel would show up before the images are stretched, but if I've got rid of them I can't align with them.

b) The method of alignment you have suggested in Photoshop is surely doing the same thing as clicking on two stars in either Nebulosity or Maxim alignment routines, and they look to be perfectly aligned, but hey presto as soon as I start stretching it shows a very small discrepancy.  Alnitak is aligned perfectly, but some smaller stars are not, maybe I shouldn't be choosing a bright star to align, perhaps that is the problem.   

Carole

[MarkS]

Alnitak is aligned perfectly, but some smaller stars are not, maybe I shouldn't be choosing a bright star to align, perhaps that is the problem.    

Don't use saturated stars for alignment because the software finds it impossible to accurately calculate where the exact centre of the star is.  As an aside, for the same reason you should never use a saturated star for guiding.

[Carole]

Thanks Mark, yes I gave it a try not using Alnitak and the alignment was much better, not 100% but not far off.  So another lesson learnt.  Worked better in Maxim than in nebulosity. 

Hopefully all problems now ironed out.

Carole

[RobertM]

Carole,

Just a quick note.  Sometimes MaximDL gets really confused when aligning so if it does then try one of the other methods: Auto star matching/Auto correlation or Astrometric.  You'll have to download the Hubble Guide Star Catalogue to use the last one but if you do then I can help there.

Hope you're all sorted now.

Robert

[Carole]

Thanks Robert, I'll keep a note of that.

Carole