Ngc3324

[Paul Haese]

I took this data a month ago now, but have only just to do some processing on it. I am planning on doing a narrow band and broad band rendition on it but for now I have only got Ha RGB data.

This is a crop of the actual data as I was not 100% happy with the Eta Carina wing being in the image. I thought it dominated too much for my liking.

With the crop in mind I composed the image to show off the star cluster in the top left and of course NGC3324.

As to image I am torn between a slight darker image that I have on hand and this one. Several renderings of star reduction have made for small stars generally but I wonder if the image is a little flat?? In the end I think it is a matter of taste but I welcome other peoples thoughts.

Click here (http://paulhaese.net/NGC3324.html)for larger image. Look under image for image details.

Feel free to provide critique.

[LewisM]

Interesting that your 3324 is predominately red, whilst mine is predominately blue. Probably because I only have 55 minutes data on it :) Our crops and composition are almost identical!

[Paul Haese]

Am I right in thinking your image is DSLR image Lewis? If so that will be the main reason why our images look so different. A DSLR cannot reproduce the exact same results as a filtered shot on a monochrome camera. The region has lots of Ha nebulosity in it and will show the response I have in my image with filter images. I have also blended the Ha into the image anyway which makes it more red. If you are using a mono camera then maybe it is processing.;)

[LewisM]

DSLR, no filters. Might add some Ha and see what happens :)

[gregbradley]

I quite like it Paul. The stars look fab, the neb perhaps a tad washed and orangy pink. It'd be nice if you could get some more solidity into the neb colours if possible.

But its a nice spread of colours as-is overall. I like the star cluster.

Greg.

[Paul Haese]

Might I suggest you take longer subs to get the detail you want and take shorter subs for your star colours. You will note star colours blow out very quickly with DSLR imaging.

[LewisM]

I agree with you Greg - I honestly blinked several times thinking I had goop in my eyes - it is a little soft/washy.

Otherwise, love it

[Paul Haese]

Yes agree about the washed out. I have now fixed that and placed the adjusted image on my website via the link. I have left the first image which is posted here as a comparison. What do you think?

[LewisM]

Already started that - there's a couple subs in mine of about 2 min duration for the stars, and 5 to 10 min for the nebulosity. But thanks to QLD's remarkable weather, I only have 55 minutes in it, which DSS throws out 10 mins of so far :)

The redux is better, but it still looks like a bit of High Pass filter, especially near the cluster. Almost like a grey difference layer or something?

[DJT]

Huge fan of star colour and these look great. Will bear in mind comments on sub lengths to retain it.

Interested in how you deal with star reduction. Still learning as always.

Surprised at how much detail you are getting though as there isn't a huge amount of data

[Paul Haese]

Just watch out for using curves function before you apply and histogram stretching. Take a look at Ken Crawfords tutorials on processing. His tutorials are very thorough and will improve your processing abilities by a quantum leaps.



For star reduction I use select/colour range with about around 50-120 (it depends on the star field, so just try a bunch of different settings). Then I use select/ modify/ expand (1-2 pixels). Then modify again and then select feather 2 pixels. Now finally I use the minimum filter at 1 pixel. If necessary do this via a mask and use the opacity function to alter the effect of the filter.

The data is primarilty the Ha data which I blended with the R filter data and then used that same Ha data again as a luminance layer. For blending some where between 35% and 55% is fine. For Luminance I use somewhere around 55%-85%, just depending on how the blend looks. Everything is trial and error.

Hope this has been helpful.

[marc4darkskies]

I like the image Paul - as is. Colour & composition look spot on.

To be picky, the stars are a bit soft and a bit too large for the subject matter - and I mean a bit. Are your Ha stars smaller? If so strengthen the contribution of the Ha stars. Some minor star reduction may make things pop more. Do you do any decon?

Cheers, Marcus

PS: just saw your last post - mostly Ha Lum - just ignore my picky comment :)

[Paul Haese]

No problem Marcus.

Out of interest I don't usually use Decon on my DSO imaging (planetary always). I find I don't like the results of doing this and have tried many different iterations. I suspect it will work better on the SBIG images than on the QSI images. Star sizes are usually reasonable full when I first start but I do several iterations of star reduction method and yet some of the larger stars are quite large. Perhaps a product of the sensor and pixels??

I see your point about the stars appearing a bit soft. That is brought on my the noise reduction I have employed with an inverted mask. I will take a look at that and see if I can improve the sharpness of the stars. :)

[marc4darkskies]

I do my decon in CCDStack after combining (but before processing in PS). Stars need to be well sampled though. I sometimes scale up by a factor of 2 before decon and then scale back down to tigthen up the stars and remove some of the fuzziness. In your case the FWHM of 2.7 would become over 5 and once decon'ed and scaled back down you may knock off a few tenths of a pixel, reducing the impact of the stars (potentially).

Greatly oversampled stars (very bright ones) don't benefit much from decon anyway and tend to contribute to the appeal of an image IMO. It's the multitude of smaller softer stars that can be distracting.

Like I said though, I am being picky! But I find sometimes it's the subtle adjustments that make the difference.

PS: I also occasionally use curves in PS on a stars only layer to reduce fuzziness (halos). This doesn't always work but a delicate touch can, if your lucky, make that subtle adjustment.

[strongmanmike]

Overall it looks pretty nice to me Paul :thumbsup:, the composition looks good and the colours are quite pleasing. As you would expect from me :whistle:, I think the smoothing is a bit too obvious :scared3: also, to me, the stars look worked on and a little unnatural, I would like to see a version with the original stars. Processing the stars artificially to be smaller is always obvious to me, particularly if one uses the minimum filter in the sequence somewhere... otherwise, these minor mostly personal preferences aside...a really nice looking image :thumbsup:

Mike

[Rigel003]

A really dramatic image Paul with great colours and composition. I always think that star reduction gives a very distinctive look with somewhat soft stars. The nebulosity is great.How did you manage to find the clear nights for this data? It's been cloud city around here.

[TheDecepticon]

Hi Paul,

Nice image, but sorry, not a big fan of it.

Stars are way to big and bold while competing with the nebulosity buried in them. Also not keen on the milky look of the image as a whole and it looks a bit too smooth in some spots.

It may be better to choose what you are trying to do with this image, the stars or the nebulosity.

Cheers,

Gray. :)

[Paul Haese]

Yes I see your point with regard to the stars. I will go back over previous versions and take a look. The smoothing is minimal though. On my new screen (3 weeks from new) there is still plenty of noise and I had hoped you would see that and I have come to agree with you that noise should be a little present to add definition.:)



This data was taken the night you killed your filter wheel Graeme. I had just not got around to processing until now.



Unfortunately the big stars are what I have been left with from the imaging train. I get a faint reflection back from the reducer on the TSA and that means I get these sort of stars. Even isolating these stars which I have for each part of the processing means the stars look big. on a true HD monitor the stars do not look that big. I am running a new 1080 monitor and the image at 1900 pixels just fits on a 17" screen. That said I take your point and would like smaller brighter stars but am left with what I have.

The milkiness in the image is naturally there. I did one stretch and this appeared. Imaging from dark skies tends to produce these sort of results. I am certain this is a faint dust component present in the region.

All this said I will go over the data again and ascertain how I get some of the elements that are pleasing and reduce the distracting ones. Thanks for the feeback.:thumbsup:

[gregbradley]

"there is still plenty of noise and I had hoped you would see that and I have come to agree with you that noise should be a little present to add definition.:)"


Say what dude? In terrestial images where one wants to create a film like effect I can agree with that. But in digital astro images I think noise is almost always a negative. Of course we have various noise tolerance and a small amount is minimally detractive but there is no substute for long enough total exposure. The battle is always to get enough data despite work, clouds, mechanical and software failures, tracking errors etc etc. Very often images therefore are underexposed and then excessive noise reduction then results in a plastic look. But look at any of Rob Gendlers images. There is no noise and the signal is totally solid and defined and he never underexposes. Virtually none of his images are under 8 -12 hours. A longer exposed image almost always looks way better than an underexposed image. Virtually any image under 6 hours is underexposed. Its difficult to get long exposure times but that is the incovenient truth. It also adds a lot of time to processing the image as well. But the final processing is much more responsive when you have lots of signal. In fact you could almost make a principle out of it. The amount of post processing is inversely proportional to the total exposure time. The longer the total the lesser the post processing required.



"Unfortunately the big stars are what I have been left with from the imaging train. I get a faint reflection back from the reducer on the TSA and that means I get these sort of stars. "

Not sure about QSI CCD window if its antireflection coated which it should be but halos like this are usually from the filters and their lack of appropriate antireflection coatings. If they are coated then next culprit is the CCD window of the camera. Its not the reducer as such, but the fact its a fast system and its picking up a slight error in design in the camera or filters, probably camera. Roland Christen commented on his QSI583 showing halo effects on RHA300 due to the CCD window. The cover slip on the chip also can add to minor haloing. But its usually filters and then CCD window. FLI went through a change on its Microline CCD window due to this problem when they changed suppliers and started having issues. They corrected this.

As I recall you are using Astrodon Gen 11? They seem good in this regard so perhaps CCD window needs an upgrade.

I agree with Mike. The minimum filter is a junk processing tool. Light decon works in moderation. Curves on stars as per Louie Atalas's tutorial works fine too. Otherwise you are better off living with larger stars than shrunk artificial ones.

Greg.

[Paul Haese]

I have done a reprocess of the data and it can be seen in the link at the start of the thread. Less smoothing, only minimal star reduction and less of a milky look to the image.

Greg I guess we will have to differ on opinion about noise levels and ways to reduce star sizes. True; time is difficult in the equation to capture enough data but not all of us have robotic scopes where you can program in each night and just gather data. I do have a robotic scope but broad band does not exist at Clayton so I must go there each time to gather data. Recently I have not had much time or inclination to go and gather data. I should setup my autofocus so I can sleep while data is being gathered.

I don't use Astrondons but I use Astronomiks which have been replaced in the last year so it might well be the CCD window that is at fault here. I personally don't find the reflections it too concerning but some of you seem to do.

So its all good and I appreciate the feed back.:)

[tilbrook@rbe.ne]

Hi Paul!

I've been reading this thread with interest, I have nothing to add as I don't have the experience of processing mono images.
But I will say I like your second version better, looks a little more detailed and more depth, the bit of noise kind of adds to the image.

Cheers,

Justin.

[DJT]

Very helpful indeed. Thanks for that. It's been an interesting thread.

[Ross G]

A very good photo Paul.

I love the colour and contrasts.

Nice composition also.

Ross.