Author Topic: looking for some pointers  (Read 542 times)

Offline Gavin

  • Newcomer
  • Posts: 10
    • View Profile
looking for some pointers
« on: 2018 October 05 19:55:09 »
Hi, I’m fairly new to this hobby in general but very new to pixinsight. Attached is a copy of horsehead I’m currently working on.

I’ve watched and read a lot of tutorials lately but I was hoping someone could help me out and give me some advice on what do next? Even if it’s just to tell me what processes to look up?

Some of the things I want to do are take off the white sheen/glazed look this image has. Make the background a touch blacker and the colours pop out a bit more. But If I move the left side of my histogram any further in I will be clipping data.

Also I want to figure out how to increase colours of different objects independently. E.g. I want to bring out the blue and purple in the very large stars more without decreasing the orange in the flame nebula.

It’s just a work I progress and I have a lot more data to collect but I’m happy with the way it’s coming along.

 Ive had to crop and degrade it a bit to be able to attach it so here is a link to it on astrobin.
https://astrob.in/full/369833/0/ 


Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 311
    • View Profile
Re: looking for some pointers
« Reply #1 on: 2018 October 05 20:27:12 »
Hi, I’m fairly new to this hobby in general but very new to pixinsight. Attached is a copy of horsehead I’m currently working on.

I’ve watched and read a lot of tutorials lately but I was hoping someone could help me out and give me some advice on what do next? Even if it’s just to tell me what processes to look up?

Some of the things I want to do are take off the white sheen/glazed look this image has. Make the background a touch blacker and the colours pop out a bit more. But If I move the left side of my histogram any further in I will be clipping data.

Also I want to figure out how to increase colours of different objects independently. E.g. I want to bring out the blue and purple in the very large stars more without decreasing the orange in the flame nebula.

It’s just a work I progress and I have a lot more data to collect but I’m happy with the way it’s coming along.

 Ive had to crop and degrade it a bit to be able to attach it so here is a link to it on astrobin.
https://astrob.in/full/369833/0/

Gavin,

So you are interested, I think, in what might be called artistic considerations. I do think there may be a mix of ideas to consider for an image like this....

1. Based on the 11+ hours you spent, I would have expected to see a somewhat less noisy image. I do not know the characteristics of the camera or your processing choices- so I am in no position to say anything more- other than an image that has very good S/N will be much easy to increase the contrast and make a dark sky with real features showing just brighter than that. It is the S/N that allows for a dark background.

2. In terms of color, one of the things that strikes me about this image is the background (low signal) has quite a bit of green. This is muddying up the color contrast. Did you think about applying SCNR? I think it will be an incremental help.

3. In terms of color contrast of the features in general- it appears pretty good to me. You can likely achieve a greater color contrast by doing a number of things without resorting to selectively operating on parts of the image. For example, with ColorSaturation you can selectively adjust the saturation of a certain range of hues in your image. You can also make certain kinds of color adjustments based on size (scale) or brightness with a mask. Another option to consider is adjusting the white reference to slightly favor either towards the red or blue (see Photometric Color Calibration.)

4. The contrast of a luminance image you work with can also greatly affect the amount of color you can blend into your data. Using HDRMT will keep the wildly bright bits of nebulosity under control... which in turn will allow you to color them better and with more ease. Also, adjusting the contrast of different scale sizes using MMT might also be a good choice.

5. Concerning the stars- my sensibility would not be to enhance color halos of them (e.g. the purple of large stars)- as my eye tends to regards this as more of an artifact of the imaging system. The funny thing is that diffraction spikes, although not something technically desirable, for short focal length systems with nearly undersampled stars can help give them color by spreading some of the light into the diffraction. It is the only exception to the rule regarding the aesthetics of spikes really. This is entirely for fun- but you can put some thin thread across your aperture (for a future experiment) just to see...

Anyway, I don't know if this is too vague or not... but perhaps a starting point.
-adam block

Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1296
    • View Profile
Re: looking for some pointers
« Reply #2 on: 2018 October 05 20:48:51 »
Hi Gavin,

Fancy seeing you here  :)

Some good suggestions from Adam.  Here are a couple more thoughts...

After running SCNR you should be able to raise the black point a significant amount without clipping data (and it is OK to clip a small number of pixels.)

It's also OK to apply a little noise reduction to the background - apply a black clipped luminance mask and invert it to protect the bright areas.  After doing this you'll be able to adjust the black point with less danger of clipping.

Another technique you might find helpful at times is to use a subtle "S" curve in RGB/K to darken the background and brighten the highlights.

Cheers,
Rick.

Offline Gavin

  • Newcomer
  • Posts: 10
    • View Profile
Re: looking for some pointers
« Reply #3 on: 2018 October 06 00:05:23 »
Thanks Adam, yeah I guess it is artistic although I still want a reasonably close representation to what’s up there. The attached earlier version shows the star colour I’m looking for, but I’m not happy with the colour of the flame nebula in this version.

With regards to the noise, we’re you looking at the astrobin copy or the 500kb version attached to the post? If it’s the attached small image, Should I expect less noise than what I have on the astrobin version?

I think the excess green is from the type of camera I use which is a RGGB style one shot colour. Thanks for the tip I will run SCNR.

I’ll also look into photometric colour calibration HDRMT and MMT I hadnt heard of those ones before.

Rick, I had a feeling you would be on this one as well 😂.

Will definitely try the black clipped luminance mask. I know how to do a luminance mask but I’ll look up how to do a black clipped one. So with the s curve do I bring the lower/left side down and the right side up a bit? I mean a downwards curve on the left and an upwards one on the right?

Thanks to both of you guys. This is exactly the sort of info I was looking for. It will give my progress a nice kick along.


Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1296
    • View Profile
Re: looking for some pointers
« Reply #4 on: 2018 October 06 00:44:14 »
Will definitely try the black clipped luminance mask. I know how to do a luminance mask but I’ll look up how to do a black clipped one.

Just use HistogramTransformation and move the left (shadows) slider to the right until you're clipping the background.  It helps to display the output histogram while you tweak the slider.

So with the s curve do I bring the lower/left side down and the right side up a bit? I mean a downwards curve on the left and an upwards one on the right?

Yes, just pull it down a little at the left and up a little at the right.  If the middle of the line is still in the original centre position then you won't affect the overall brightness of the image.

Another way to selectively manipulate colour is to build a mask with the ColorMask script, apply it, and then adjust with curves (R, G and B, obviously, but you can also play with CIE a* and b* and the others as well.)

Cheers,
Rick.