Mono image processing - simples it is not

[rubecula]

I feel that I must be doing something wrong in my attempts to process LRGB or HaRGB images from data captured in the recent clear nights we had in the UK.  I obtained 12 x 600 secs of Ha and 10 x 600 secs of L,R,G & B with the Atik 383L+.  Focusing was done by hand via SGP Frame and Focus as best I could.  The images below were stacked in DSS and I've only stretched them with Curves and Levels and done a bit of colour balance.  They've not had any gradient removal or noise reduction. 

The first image is the Luminance subs that have undergone 10 small Curves stretches to try to expose the nebulosity without blowing out the stars.  When I combined this with the RGB subs the image was totally dominated by the stars rendering the tadpoles somewhat insignificant.

In the second image I added the Ha as a luminance layer with 80% opacity. This to me appeared much better as it held the stars back a lot and looked more balanced.

The final image is just the Ha on its own converted to RGB. Again to me this looks as good or maybe better than the second so makes me wonder why I bothered with capturing R, G and B.

Do I need different capture times for the various filters?

Is my image stretching faulty - particularly with the Luminance?  I've read several books / web sites seeking info on how to do it.

Is this target inappropriate for LRGB imaging?

1) Luminance

2) HaRGB

3) Ha only

Thanks

[tooth_dr]

Can you upload the unprocessed FITS or TIFF for the luminance please?  I have this camera, and wouldnt mind seeing what data is in that image, it looks dark to me.

[x6gas]

Hi Robin,

Don't lose heart!  IC410 has a strong Ha signal and there is also some [OIII] and [SII] in there as well.  In theory the Ha signal should be in your luminance as well, of course, but in practice it will get washed out by all the other wavelengths so your luminance stack is pretty much what I expect.  For this reason I don't bother collecting luminance on a narrowband target like this...

Your RGB data isn't wasted, though... far from it!  You can make an HaRGB image by blending the Ha and R channels (the 'lighten' blend mode in photoshop works well for this) or you could create a star layer using the RGB data and overlay that on top of the Ha) you need to create a mask by copying the image into a reveal all mask and use levels on the mask to make the mask black and the stars you want to show through white...

[ollypenrice]

Sometimes it can be difficult to hold down the stars in luminance since that's the filter which passes the full light spectrum, so it simply passes the most light. This is especially true when you are trying to image a DSO in a dense, bright starfileld as is the case here. This target is a challenge in LRGB because of the stars which are held down better in NB. NB also works better on the nebulosity. I have the Tadpoles in an HaLRGB widefield image and they are very disappointing and rather swamped. However...

It is absolutely vital to nail focus, which is best done using a programme which measures FWHM or by using a Bahtinov mask if you seeing isn't too stable.

There is a custom curve which can help hold down stars. You need to experiment but it would look something like this:

The idea is to give a steep lift for the darker parts but convert to a flat line earlier than usual. It won't always work.

This set of inexpensive Ps actions has an excellent star reduction tool. There are others but I know this one and use it: https://www.prodigitalsoftware.com/Astronomy_Tools_For_Full_Version.html

I would resist the temptation to put more than a little Ha into an image as luminance. I put the bulk of it into red in blend mode lighten. I may apply it to luminance but only to a 15% opacity.

If you are going to let the Ha layer carry the nebulosity into the image using this method there is no need to push the LRGB with a hard stretch.  Simply settle for a gentle stretch with small stars and let the NB do the heavy lifting - but I would advise you to put Ha into red and OIII into green and blue, all in blend mode lighten. Personally I think it's a good idea to ask yourself what is the purpose of each layer? Why have you shot this layer, what will it contribute, how will it interact with the other layers? Some experience is needed before you can answer these questions confidently but start thinking about this from the outset.

Olly

 

[rubecula]

tooth_dr (sorry don't know your name) here is the stack of Luminance images output from DSS:

The Tadpoles LUM stacked.TIF

Thanks

[FaDG]

2 hours ago, rubecula said:

In the second image I added the Ha as a luminance layer with 80% opacity. This to me appeared much better as it held the stars back a lot and looked more balanced.

The final image is just the Ha on its own converted to RGB. Again to me this looks as good or maybe better than the second so makes me wonder why I bothered with capturing R, G and B.

Hi Robin, personally I don't agree with the statement on Ha: I definitely like the second image best, because of the additional hues generated by the blue and green. So, while the bottom left part of the two images is very similar, I think that the top right part is enhanced in the second one.

True, there seems to be a slight bloating on the stars, but I think it can be managed. So, for me your HaRGB is worthwile.

Fabio

 

[rubecula]

tooth_dr apologies the image I sent you had been stretched. Here's one that hasn't:

The Tadpoles LUM stacked.TIF

Thanks

 

[carastro]

Although there is more data in the luminance, I agree with above, it is better to just do narrowband or HaRGB in these sort of targets.

Carole 

[tooth_dr]

14 minutes ago, rubecula said:

tooth_dr apologies the image I sent you had been stretched. Here's one that hasn't:

The Tadpoles LUM stacked.TIF 16.27 MB · 0 downloads

Thanks

 

Thanks Robin

A stretch revealed lots of nice data

 

All the best

Adam

 

[rubecula]

To process LRGB  I've been using Merge Channels for the RGB in Photoshop to stretch all three at the same time and then add the Luminance as a separate layer above the RGB. 

When adding some Ha as red do you somehow add it to the red channel together with the red subs so that in Merge Channels you have (R+Ha) + G + B?  Or do you stretch it first and add as a seperate layer?

Thanks

[x6gas]

Personally I like to retain as much control as possible so I don't combine the RGB using merge channels.  I generally do some processing to each channel - i.e. several iterations of curves and levels then load them into a PS file, colourising each of the RGB frames and blending them using the "Screen" blending mode.

[rubecula]

22 hours ago, tooth_dr said:

 

A stretch revealed lots of nice data

Thanks for doing that Adam, you got a lot more out of that without blowing the stars. Confirms that I need more practice. It's good to know it's me and not the image ?

[tooth_dr]

18 hours ago, rubecula said:

Thanks for doing that Adam, you got a lot more out of that without blowing the stars. Confirms that I need more practice. It's good to know it's me and not the image ?

No bother Robin.