NGC1333 LRGB

[MartinFransson]

That last version looks good!

I´ve seen some of your images so I know you are a skilled imager. This is just a very hard part of the sky to capture and process and I think that if you´re in lightpolluted environment you could shoot 10000 subs and still not get it the way you want. I think this target needs the darkest of skies possible. I tried it once and got nothing! Sure, I didn´t catch a whole lot of subs but it was so bad I decided not to try it again. And I live in a green zone (Bortle 4 or so).

Bottom line - it´s not you, it´s the LP

[Rodd]

9 minutes ago, MartinFransson said:

That last version looks good!

I´ve seen some of your images so I know you are a skilled imager. This is just a very hard part of the sky to capture and process and I think that if you´re in lightpolluted environment you could shoot 10000 subs and still not get it the way you want. I think this target needs the darkest of skies possible. I tried it once and got nothing! Sure, I didn´t catch a whole lot of subs but it was so bad I decided not to try it again. And I live in a green zone (Bortle 4 or so).

Bottom line - it´s not you, it´s the LP

Thanks Martin.   I am so looking forward to imaging from a truly dark site--from a place where there are so many stars you have trouble recognizing the constellations.  that would be awesome.  BTW--a final tweak--lifted brightness a tad (this image has a tendency to darken as you progress.  Not sure it will become a permanent change--need time to assess.

 

[carastro]

Quote

 it was so bad I decided not to try it again. And I live in a green zone (Bortle 4 or so).

That's reassuring to know as my dark site is Bortle 4 (my home site is Bortle 8!!!) which might explained why I have been struggling at my dark site, not even bothered from home. 

Carole 

[Rodd]

On 15/12/2018 at 17:47, wimvb said:

hat's a huge improvement over v 1.

I spoke too soon.  I tried ACDNR for the first time.  There were still some pixel scale dark speckles (pixels) that TGV denoise could not remove no matter how much I used.  necessity really IS the mother of invention.  I decided to try ADDNR for the heck of it.  I guessed at a very small radius and 3 iterations.  Different settings would probably do a better job--but I was amazed at the effectiveness of this one attempt without even knowing what settings to use.  Let me know if I am chasing phantom successes hear!  I think it was you that suggested ACDNR--thanks Wim!

[wimvb]

On 14/12/2018 at 16:58, Rodd said:

But M8 and M20 came out hjust fine, as did the cacoon--even the tadpoles LRGB.  And galaxies (struggle with background)--but M31 was fine.

Rodd

All of the above are emission nebulae or stars. In this target you're trying to capture dark nebulae, or reflection at best. It's like trying to image a black cat in a coal storage, during the night.

So, it's not that strange that you struggle with a galaxy background.

[wimvb]

53 minutes ago, Rodd said:

  I think it was you that suggested ACDNR--thanks Wim!

I'm not sure it was me, to be honest. I don't use acdnr much. Only very occasionaly to even out mottle. Then only on chrominance.

But it seems to have worked. ?

[Rodd]

11 minutes ago, wimvb said:

I'm not sure it was me, to be honest. I don't use acdnr much. Only very occasionaly to even out mottle. Then only on chrominance.

But it seems to have worked. ?

Remarkable.  I think I need to reprocess this from the start as there is no doubt a more appropriate place in the workflow for ACDNR--and probably more suitable settings as well.  I have never used it, other than when I first got PI I experimented with many tools without having a clue how to use them.  I think I used ACDNR for some lunar images I took.  I didn't know about TGV denoise back then so I can't say which was more effective.

Rodd

[wimvb]

I use tgv denoise on small scale "salt and peppar" noise in luminance, where it seems most effective. But I allways use it with a luminance mask where the histogram peak has been lifted to 50%, by lowering the black point. This works on linear as well as nonlinear images. The only parameter you need to adjust is the threshold.

I use mmt on chrominance, with the same or a stronger mask. On some occasions I've used acdnr in one of the last stages of processing. I used it more when I had my dslr. Not at all often nowadays with rgb imaging.

[wimvb]

A new version, completely reworked. This time I took a reference image for DBE and applied it to the masters before RGB combination. The blue channel has a completely different gradient than the other channels. Normally DBE can handle such a situation, but this image is a tough nut to crack.

In the L image, I separated the small scale and the large scale, then pushed the large scale a little more.

After LRGB combination I was left with magenta in the dust, so I inverted the image, and applied SCNR Green with a mask to protect the red details.

I'm not happy with the left hand side, where the dust cloud is too heavy.

Btw, there are some issues with the brightest stars. I think that this is a result from sigma clipping during stacking. Here's a closeup of the red channel.

[Rodd]

45 minutes ago, wimvb said:

Btw, there are some issues with the brightest stars. I think that this is a result from sigma clipping during stacking. Here's a closeup of the red channel.

That looks good.  I have tried again and again and can't better my last attempt.  I am pleased with my last one though--it's respectable-not great, but its getting close to a 90 (since you are familiar with grades).  Maybe a B+.  OK...a B.    I can live with it (until I can't!!!).  ButI need some distance--maybe in a few weeks I will be able to look at this data again.  I DBEd each channel separately--but then the last time I did it all together and it seemed to yield less speckles.  Who knows.   Yeah the stars are not great--but then they never are with this setup.  I hope its because the distance thing--which I will change when I am dome with M78.  I am 1mm short (supposedly).  Re stacking the data will take days (literally) , so it is not high on my list.  I have to align one channel at a time, then delete the aligned subs before moving on to the next stack.  Normally I Align everything at once--but I don't have the space for it because I have to use SFS to weed out the bad subs--the SFS subs are saved to their own folder--so I need more room.  Maybe I will re-align them....but I need to take a break from NGC 1333.  Your last image shows that the noise can be removed to good effect, so there is hope.  BTW, I have made inverting images part of my work flow.  Sometimes you can use SCNR straight away--other times you need a loosely fitting star mask.  The colors are opposite.  Thanks for showing me there is a way through the brier patch.  You get a 92.  Hey--tough image, so that's good

Rodd

[wimvb]

I don't think you need to redo the star alignment. Just the stacking needs a revisit. You can then also do a large scale pixel rejection to remove the artefacts in the colour channels which cause the blue and red streaks. If you restack, you can do so for only one channel. Tweak the rejection settings until you get good results. Stacking uses the file cache, which means that it's a lot faster 2nd time around. Then, when you have the best rejection parameters, apply these to the other channels as well.

Btw, I based my process on this method for enhancing luminance:

http://www.deepskycolors.com/archivo/2010/05/07/multi-scale-Processing--Revealing-very.html

( which is in a hidden section of Rogelio Bernal Andreo's website: http://www.deepskycolors.com/tutorials.html )

It has the advantage of not pushing the noise (which is mainly a small scale effect) and keeping the stars under control.

[Rodd]

6 hours ago, wimvb said:

don't think you need to redo the star alignment.

I deleted all the aligned subs after integrating.  Not having enough space is a big issue.  This is one major drawback of CMOS.

Rodd

[wimvb]

A tip to recover disk space: find the pixinsight temporary directory (something like c:/users/yourname/AppData/local/temp/) and delete all the swap files (.swp) that are older than about a week or so. These swap files can be very large. Usually, PixInsight should delete them, I believe, but it doesn't always.

[Vixen4eva]

So much going on in this image...........fantastic!!! Wow!!

Alan

[Rodd]

6 hours ago, wimvb said:

A tip to recover disk space: find the pixinsight temporary directory (something like c:/users/yourname/AppData/local/temp/) and delete all the swap files (.swp) that are older than about a week or so. These swap files can be very large. Usually, PixInsight should delete them, I believe, but it doesn't always.

Didn't know that.  I will look and see.  Hard to imagine they are bigger than a couple GB.  But I have often wondered what is taking up so much space.  Out of a 500BG hard drive I have about 150GB without any images on the computer--almost 300 GB worth of software (a big reason I can't install PS).  Windows takes up a lot I know--all the features I do not want.  I would love to get a computer with nothing but AP related software.   I am hoping to get a 1-3TB drive for Christmas.  That will solve my problems.

Rodd 

[Rodd]

2 hours ago, Vixen4eva said:

So much going on in this image...........fantastic!!! Wow!!

Alan

Thanks Alan--great region, bad for your sanity though

Rodd