NGC1333 LRGB

[vdb]

I use G:75 on my ASI 183 which is 35 below unity gain, I can go longer en capture more photon's so I get a better SNR in the low signal parts of the sky, as always there is no one rule fits all and you need to experiment a little bit, but also and this is important one needs to accept that dusty objects will always be very hard if you do have LP in the mix ... a filter can help, but also kills signal of the object ...

Sometimes it's better to take your car drive 1 hour, get 4 hours of clean data and it will just kill 20+ hours from an LP zone ... Someone once told me that the best telescope is a car (I live in a white zone)

/Yves

 

[wimvb]

3 hours ago, vdb said:

Someone once told me that the best telescope is a car (I live in a white zone)

Couldn't agree more. In my case it was a moving company that owns the car/van. ?

[wimvb]

20 hours ago, ollypenrice said:

I'm also not sure about the magenta zones. I didn't find them in my shoot but I see magenta patches creeping into some posted images. They are not in consistent locations, though, which suggests they are an incipient processing hazard in this target. I really don't know.

Olly

I'm becoming more and more convinced that this is also due to DBE. I've been playing around with DBE on the rgb data, and I keep getting these colour variations. The problem with this target is that you can't put markers anywhere without tripping on the nebula. And where you can put them, weight factors are all over the place. It's a minefield.

Je maintiendrai!

[ollypenrice]

27 minutes ago, wimvb said:

I'm becoming more and more convinced that this is also due to DBE. I've been playing around with DBE on the rgb data, and I keep getting these colour variations. The problem with this target is that you can't put markers anywhere without tripping on the nebula. And where you can put them, weight factors are all over the place. It's a minefield.

Je maintiendrai!

Although I've yet to try it some good imagers I know do DBE separately per channel. Any thoughts?

Olly

[Rodd]

3 hours ago, ollypenrice said:

Although I've yet to try it some good imagers I know do DBE separately per channel. Any thoughts?

Olly

That's how I do it--I create a dbe icon with an image that has the most background--usually lum, but sometimes not.  And I use that icon on all channels.  Seems to work well.  I have noticed that it can remove gradients better than if the 3 channels are combined because the gradients can become complex if they are slightly different--which for me is true because I tend to stick with one filter until its completed before moving to the next filter (stemmed from the jamming filter wheel days of the STT-8300)

Rodd

[Rodd]

8 hours ago, vdb said:

I use G:75 on my ASI 183 which is 35 below unity gain, I can go longer en capture more photon's so I get a better SNR in the low signal parts of the sky, as always there is no one rule fits all and you need to experiment a little bit, but also and this is important one needs to accept that dusty objects will always be very hard if you do have LP in the mix ... a filter can help, but also kills signal of the object ...

Sometimes it's better to take your car drive 1 hour, get 4 hours of clean data and it will just kill 20+ hours from an LP zone ... Someone once told me that the best telescope is a car (I live in a white zone)

Ives 

 

But it would take 4-5 (or more) such moves to complete a decent image.  Now, I focus, then let the rig run for an hour--check focus--then let it run for 2-3 hours--if my temp compensation is working well--I let it run right up to the meridian flip, and if the target is well positioned, I let it run hours past that.  So I am "around"  to eat dinner and do other family related obligations.  If I disappeared completely for several nights--I would be in trouble.  Not to mention it would have to be on weekends--and with that constraint I would never get an image completed due to the weather.  Also, to really make it worth it, I would have to drive 4-5 hours.

Rodd

The bed and breakfast idea is one I am kicking around.  stay for a week or so in the mountains--or where ever, and image.  then it becomes trying to pick a week a month or two in advance that will have good weather.  Tough.

[wimvb]

3 hours ago, ollypenrice said:

Although I've yet to try it some good imagers I know do DBE separately per channel. Any thoughts?

Olly

I usually manage quite well with dbe on rgb, since it calculates the gradient for each colour separately, as far as I know. But you can't control the weight factors for each channel that way. So, if you want to go into the nitty gritty of dbe, and follow the math, then yes, it would be better to do it on each master. I tried with this image, and it didn't make much difference. I think that the correct placement of markers is the critical step here. I got best results with putting PixInsight best practice overboard. 150 pixels sized samples put right over nebulosity and (not the brightest) stars. I only considered the colour. This steamroller approach flattened the colour gradients best, as well it should. For lum I used smaller samples with critical placement. Afterall, rgb isn't about detail, it's about colour. Lum is about detail.

[Rodd]

Here my best effort to date.  Magenta areas not as bad--pretty good detail in the dust, and what I like most is the background space is not flat black.   But to make this colossal effort worth it....it needs to improve!

[wimvb]

54 minutes ago, Rodd said:

The bed and breakfast idea is one I am kicking around.  stay for a week or so in the mountains--or where ever, and image.  then it becomes trying to pick a week a month or two in advance that will have good weather.  Tough.

A trip for two, perhaps. I don't know about your neck of the woods, but over here, late winter and spring can be pretty decent. And if it isn't, it can still be a nice getaway.

Some astrophotographers invest in a trip to really dark skies in Africa. That would be a once-in-a-lifetime for most, I guess.

[Rodd]

13 minutes ago, wimvb said:

ome astrophotographers invest in a trip to really dark skies in Africa. That would be a once-in-a-lifetime for most, I guess.

Yogi Bear-catch a big fish (scope/mount), drop it to catch 6 little ones (airline ticket and motel reservation).  If only Boo-Boo would help!

Rodd

[wimvb]

Daytime photography is sooo much easier.

On another note, I don't think that the edge artefacts that you and Olly discussed are due to DBE, but are in fact from stacking.

I did a generous edge trim and processed the RGB data with DBE and a lot of stretching/colour saturation. This is the result (never mind the noise and stars, I wanted to see if DBE + stretching affected the edges)

[Rodd]

5 minutes ago, wimvb said:

This is the result (never mind the noise and stars, I wanted to see if DBE + stretching affected the edges)

I think I came to the same conclusion but didn't want to give up the real estate.  Drop the histogram on this and it will be nice.  I am going to reprocess this one more time and see if I can tame the beast.  The Balrog is a mighty foe (and I am no Gandalf)

Rodd

[Rodd]

Using color mask script I reduced the magenta in several patches, and I reduced the green in the dust.  This is it until I redo the data.  Still a bit of magenta in dust arc on left--but I can't isolate that with the colormask.

 

[Rodd]

OK-last tweak--star control--brings out the dust (no pixel peeping!--for your own protection) I am planning on the reprocess---but not sure how to proceed that's different

 

[wimvb]

That's a huge improvement over v 1.

Btw, it's quite easy to clonestamp those coloured streaks out.

I used a small circle and 50 % opacity.

[Rodd]

43 minutes ago, wimvb said:

That's a huge improvement over v 1.

Btw, it's quite easy to clonestamp those coloured streaks out.

I used a small circle and 50 % opacity.

Thanks wim. It took some doing. Not a complete reprocess though.  That is what I will now do. It might take a few days. I need a break

[vdb]

17 hours ago, Rodd said:

But it would take 4-5 (or more) such moves to complete a decent image.  Now, I focus, then let the rig run for an hour--check focus--then let it run for 2-3 hours--if my temp compensation is working well--I let it run right up to the meridian flip, and if the target is well positioned, I let it run hours past that.  So I am "around"  to eat dinner and do other family related obligations.  If I disappeared completely for several nights--I would be in trouble.  Not to mention it would have to be on weekends--and with that constraint I would never get an image completed due to the weather.  Also, to really make it worth it, I would have to drive 4-5 hours.

Rodd

The bed and breakfast idea is one I am kicking around.  stay for a week or so in the mountains--or where ever, and image.  then it becomes trying to pick a week a month or two in advance that will have good weather.  Tough.

I would do 4 hours L remote ... it beats 20 hours L in LP by a wide margin ... RGB you can be more aggressive in noise reduction and gradient removal becomes easier as you know what true signal is and not ...

/Yves

 

 

[carastro]

OK, I had another go at it.  I haven't cropped any edges off it.

Processed in Photoshop.

 

 

[Rodd]

3 hours ago, vdb said:

I would do 4 hours L remote ... it beats 20 hours L in LP by a wide margin ... RGB you can be more aggressive in noise reduction and gradient removal becomes easier as you know what true signal is and not ...

/Yves

 

 

Oh--and I forgot a rather big issue for me.  I do not like to take the camera off, or the whole rig down for that matter, because even with a CAA I can never find the right angle.  Vlaiv has taught me a way to do it, but I have yet to try.  I would have to shoot all new flats.  Don't get me wrong--I do want to try imaging from a dark site (true dark site 4-5 hours away min).  But I need power too.  I would have to buy another battery and the fixings to equip it with connections for a multiport hook-up.

In short--not an easy proposition for me.  Now--a week stay so I can shoot a complete data set....now that is another story

Rodd

[Rodd]

3 hours ago, carastro said:

OK, I had another go at it.  I haven't cropped any edges off it.

Looks great Carole.  Is the histogram balanced?  I see a large circular area that is quite blue on my screen.  I would be tempted to run DBE (or ABE in non linear mode)--or scnr blue.  PS probably has ways to address this as well.

Rodd

[carastro]

Yes it is balanced Rodd (as near as damn it), I re-checked it at the end as the sky background came out a little blue, but that may have been because I used Lab colour to enhance the blues. 

I don't use Pixinsight but you are welcome to tweak my process.

Carole 

[Rodd]

1 hour ago, carastro said:

Yes it is balanced Rodd (as near as damn it), I re-checked it at the end as the sky background came out a little blue, but that may have been because I used Lab colour to enhance the blues. 

I don't use Pixinsight but you are welcome to tweak my process.

Carole 

Just a quick palette adjustment.  Not perfect. I think you have done well with the data--the desired end result may not be achievable with this data set. (I am NOT going to add to it--I'm moving on.  I am sure I will do it again).

PS...PNG is new for me--I see it is much larger.  is it better than JPEG...or TIFF?  Those are the only 2 I know will display on SGL.

[carastro]

Quote

PS...PNG is new for me--I see it is much larger.  is it better than JPEG...or TIFF?  Those are the only 2 I know will display on SGL.

I always use PNG (portable network graphic) for uploading to Astrobin as it is smaller than a TIFF which AB won't accept, but much better than a Jpeg.  TBH I think it works as well as Tiff.  I process in Tiff but convert to PNG for uploading.  Having said that it was still 82mb and SGL took a while to upload it.

Generally I link to my Astrobin image (already uploaded as a PNG) but this morning AB wasn't playing ball which is why I uploaded to SGL.  (of course I was trying to upload it to my staging area as it is not my image). 

HTH

Yes colour balance does look better now.

Carole 

[Rodd]

On ‎15‎/‎12‎/‎2018 at 04:56, wimvb said:

Couldn't agree more.

see below

2 hours ago, carastro said:

Yes colour balance does look better now.

see below

On ‎15‎/‎12‎/‎2018 at 07:06, ollypenrice said:

Any thoughts

Well--here it is--I know the color palette is a bit garish, but the noise is better.  This is a REALLY HARD image to get right in all respects.  Too bright, too dim, too much noise, too smooth, to much red here, not enough red there, and the blue...ah!!!!  There is still a fine scale dark peppery effect at the pixel scale.  Noise control does not help.  And, as if I need more ands, the bright patch above the main target is not cooperating.  But Better?

EDIT--I keep waffling between these 2 so I decided to include them both

[Rodd]

OK--This is IT!!!  Last change.  Reduction of blue regions, cropped the edges.  More balanced palette.  Lower left region is a bit flat still.