NGC 891 help

[Ibbo!]

Hi

Rob suggested I post this question here

Last year just before Kielder I started to image NGC 891 from home

From Kielder I took some more data with the FS102 while aquiring the RGB with the 80 mm ED and guiding with another scope

( so 3 up on the NJP FS102/H9 L 80 Ed/716 RGB guiding FS60/QHY5)

L data was 600 s from home x31 and 10 1200s from Kielder unbined

Colour was 300s x10 each unbinned

darks and flats applied

After seeing Robs great image of NGC891 and playing with my data for nearly a year I had another go and it came out as normal

ie

Not quite aligned

Bloated stars

and washed out colour in the stars

lots of data but can't get it right

what am I doing wrong how can I improve this image

Cheers

Steve

[RobH]

Hello Steve,

This isn't an easy galaxy to pull out, as it sits in a quite busy starfield.

It looks like what has happened here is that you've ended up burning out the stars when you've stretched the image to bring out the galaxy.

This has made the stars bloat, and clipped the white point on all colour channels, so the star colour has gone.

Did you use levels, curves, or both when you stretched your data?

A trick you could try is to add your 80mm RGB data again as another layer, and set the blend mode to 'colour'.

Then go to 'Image-adjustments-match colour' and bring up the colour intensity. This is more effective for boosting colour than 'Hue/Saturation' as it won't add colour noise.

How did you register your 80mm data to your luminance data?

Cheers

Rob

[roundycat]

you didn't mention how you are doing the alignment although at this image scale there is little direct evidence of misalignment. You don't have many subs, can you combine them manually in Maxim or similar.

If you attend to the stretching routine more carefully you will hardly move the white point. That is really how it should be, ie, don't move the white point. Star bloat and, to some extent, lack of colour will be a thing of the past.

Dennis

[Ibbo!]

Thanks I will have a look at my processing again as I have not done colour for quite a while and will try to blend in P/S

I did use levels and curves in p/s with a mild unsharp mask selection around the galaxy and selective bits in the arms but hopefully not too much

They were aligned and stacked in Maxim auto star matching for the individual LRGB then manual 2 star alignment to align the LRGB stacks

I tried DSS and was not happy with the result at the time

I also did have a try with registar but the trial ran out before i could work out how to use it properly

Dennis

Just to ask how many subs do you think I need to take I tend to go down the route of one more nights L and then suddenly so it seems the object has set

31 x 600 L plus 10 x1200s L (thats 41 total L subs a time of 510 mins 8.5 hours)

or are you saying 50 mins each colour is not enough and thats my problem

thanks for the suggestions

Steve

[RobH]

Hello Steve.

You certainly have enough luminance data.

I would double the amount of colour, but even with what you have, you should have decent colour, but maybe a bit noisy...extra subs will cure that.

Dennis and I both do all of our stacking etc in Maxim.

I use auto star matching for everything except the most tricky alignments.

Cheers

Rob

[roundycat]

Steve, first, if you have licked two star manual alignment in Maxim don't use any other method, especially anything that says auto. When it comes to combining them use an outlier rejection method such as SD Mask. Median or Sigma clip will both work but they generally need more subs to drive the noise down. If your computer cannot cope align them first and save. Close all and then use the File Combine dialogue to combine, ie, don't open all the sub frames on screen.

For a clean rejection of the noise when combining with SD Mask you need at least 8-10 subs. Really bright sat trails may need more subs or more than two passes to cleanly reject the noise.

A good rule with LRGB is to have about equal amounts of all four channels. So, 50m of L or 8hours of R, G and B. The choice is yours! This is more noticeable in nebula where the red colour tends to be diluted with too much L.

As regards processing, try the careful Levels and then Curves stretching but before you have it finished duplicate the layer, select the stars and put them onto their own (top) layer and set the middle layer to overlay and apply High Pass filter to the middle layer. Radius between 3-10 pixels until you get the feel of it. This will accentuate the galaxy detail and not sharpen the stars as they will be on their own layer.

If you feel the need to finish off with some USM do it to the flattened image but be careful of the settings. Start with 80, 1.5, 15 as Amount, Radius and Threshold. The values can be critical, the idea is to slightly sharpen the stars and finer detail without going over the top. Remember that High Pass, USM and any other sharpening filter work by enhancing contrast and that means the bright bits will tend to saturate. Avoid saturation at all costs.

Dennis

[RobH]

Dennis.

Why not use 'auto'?

I've bben using it for 2 years and have no trouble at all with alignment.

Maual works fine of course, but is a lot more work, and if auto works, why make life hard for yourself?

I always use auto first, and then, if that doesn't work (very rare) I use manual 2 star.

Cheers

Rob

[Ibbo!]

I find that in most cases the auto works very well and I do go this way to start

but it does get confused and I end up with a sort multi stamp collage of galaxies (or whatever)

I will have another go from the start when Helen goes out (much easier when quiet ducks in anticipation of blow to head)

Steve

[roundycat]

Rob, I had to laugh when my previous post obviously 'crossed in the night' with yours. There's you saying use auto and me saying don't touch it with a barge pole.

I have had trouble with auto in the past and had to use the less accurate version of the two different auto methods to get it to work. Then I read one of the US gurus saying always use 2 star manual and tried it and never had a problem. The thing is with manual, you get to choose your own alignment stars so anything that looks remotely like a bloom or microlensing spike, is mishapen or too bright can be ignored. I'm not sure whether auto does that.

I'm also not sure whether the same centroiding algorithm is used but two star manual reads out to three decimal places of a pixel.

When using 2 star remember to turn on Auto Next.

Dennis

[RobH]

Hello Dennis.

It's that classic case of 'if it works, keep doing it'

Works for us!!

Cheers

Rob