Spiral Galaxy M83 12 hours 20minutes Repro

[jase (Jason)]

Repro looks great Greg. Top stuff.

[gregbradley]

Originally Posted by jase

Repro looks great Greg. Top stuff.

Thanks Jase. I have always found M83 a hard target for some reason.

Greg.

[madbadgalaxyman (Robert)]

Greg,

In the largest (cropped) version of your image, I seem to see an enhanced density of compact blue sources (clusters?) (blue knots?) surrounding the bright portions of M83

What do you think?

M83 is thought to have some very luminous and massive young star clusters in it; more massive than any in the Milky Way.

In addition, in the ultraviolet, it is very obvious that there are many Ultraviolet-emitting knots stretching a long way outwards from the obvious portions of this galaxy.
Here is the GALEX image showing these UV/blue knots, which are very luminous in the UV:

If you go to http://galex.stsci.edu/GalexView and search for ultraviolet images of M83 (making sure to pick the longest exposure image of this galaxy) you can see these UV knots in greater detail. They extend away from M83 for a gigantic distance.

Visually, the UV luminous knots are extremely faint, and I have looked for them in other M83 images without success (detection of these objects at visible wavelengths probably requires excellent depth and also excellent angular resolution). But they "light up like candles" at <2000 Angstroms because one must assume that they contain O and B stars that are so hot that much of their emission of light is in the Far ultraviolet.

At this stage, I can't say for sure if the UV knots correspond to the blue knots that I think I see on your image.

Incidentally, very sorry that I have not answered your email about trying to detect clouds that are high above the plane of the Milky Way.
In my defense, all I can say is that I have been extremely busy with several projects, both astronomical and non-astronomical.

Best Regards,
Robert

Even just by mucking around with your .jpg, I could easily make these knots or clusters look more obvious. Some of them are very obviously resolved and extended objects, so it is highly likely that these extended knots do actually belong to M83.

[gregbradley]

That's very interesting Robert.

I wonder if I can capture those. I'll give it a try next time M83 is up.

Greg.

[SpaceNoob (Chris)]

Now that is a corker!

Well done dude, the lack of donuts makes this image awesome!

[gregbradley]

Originally Posted by SpaceNoob

Now that is a corker!

Well done dude, the lack of donuts makes this image awesome!

Thanks Chris. Startools did a good job of cleaning them up. I can get rid of individual dust donuts but its a difficult processing task and annoying as correct flats should handle that. Lesson learned. The trouble was I was taking many images and tending to reuse the flats as I hadn't changed the camera or orientation or so I thought. Perhaps I did but even so flats clearly have an expiration date. New dust particles can get in.

Greg.

[Paul Haese]

Dust donuts still there Greg. Sorry can still see those, nice colours though and good detail.

[gregbradley]

Originally Posted by Paul Haese

Dust donuts still there Greg. Sorry can still see those, nice colours though and good detail.

You must be boosting the curves or something. You can see a few if you blast the curves but on my monitor they are not visible under usual conditions.

Greg.

[Paul Haese]

Originally Posted by gregbradley

You must be boosting the curves or something. You can see a few if you blast the curves but on my monitor they are not visible under usual conditions.

Greg.

Nope, the screen is showing them there. Not doing anything of the sort Greg.

[gregbradley]

Originally Posted by Paul Haese

Nope, the screen is showing them there. Not doing anything of the sort Greg.

Fair enough. Monitors vary.
Any suggestions on how to tackle them? Startools tutorial implies they can be handled with their tool. PixInsight has dynamic background extraction. There are tricks in Photoshop. Photoshop CS5 and beyond have a content aware healing tool. I guess I try a background sampling approach to remove. It'd be good to get a hold of this as it comes up occassionally. I could post a how to once successful.
I'll do some trial and error and post the results. I am sure others will be interested as well.

Greg.

[alpal]

Hi Greg - it's easy.
In Photoshop you make an inverted layer mask.
It ends up like this:

You then apply it to the image & get the result.
With a bit more fiddling around I could have got it perfect.

[Paul Haese]

Hi Greg,

Martin at the conference said he takes three lots of flats. One at 21,000, one at 26,000 and one at 30,000. Then combines all those to make a master flat. I remember him saying this during his presentation and he has told me this privately too. I suggest you give this a go too. Experimentation of flats will produce the results you are looking for I think.

[gregbradley]

Originally Posted by Paul Haese

Hi Greg,

Martin at the conference said he takes three lots of flats. One at 21,000, one at 26,000 and one at 30,000. Then combines all those to make a master flat. I remember him saying this during his presentation and he has told me this privately too. I suggest you give this a go too. Experimentation of flats will produce the results you are looking for I think.

Yes I remember that. I was the one who asked him how he did his flats as flats are critical on my CDK and 16803. I think in this case its not a matter of an average of flats (although I intend to try that out) its more a matter of the flats I did take were out of date, or I shifted the camera and did not take fresh flats etc. Hence the difficulty. I have done a lot of experimentation with flats and I think I have it down usually but I must admit I want to try what he suggested out. I find with flats on the CDK I get best results from around 25-30ADU and mean combine and no dark flat subtract merely a bias which is applied during callibration. Plus accurate darks. That is with the Proline 16803.

I think you will find the same on your RC12. Its the long focal length where these donuts show up way more than on shorter focal lengths. The best thing obviously is to keep everything regularly cleaned. These donuts are on the sensor CCD chamber window not so much the filters.

I thought it was a clever idea to average several as you definitely do get different results with different ADU.

Originally Posted by alpal

Hi Greg - it's easy.
In Photoshop you make an inverted layer mask.
It ends up like this:

You then apply it to the image & get the result.
With a bit more fiddling around I could have got it perfect.

Hey that's good. How did you make that inverted mask?

I used a similar background sampling using the ellipse lasso tool, move it over the dust area and blend it in with a reveal all mask to let the original stars through.

I've posted the result of doing that and whilst there may be some minor residual dust it only shows up on huge curves boost now. See what you think - same links.

Thanks for the suggestions.

Greg.

[alpal]

Hi Greg,
learn inverted layer masks from Louie here:

http://www.youtube.com/watch?v=32Mj2Ia-tC0

(in this case the filter is brightness adjusted with curves)
In the case I showed I just adjusted levels & curves to
bring out the dust doughnuts for an inverted layer mask.

This was another layer which was used to adjust a copy
of the image for curves.
The area that is white was therefore adjusted to increase
it's brightness with curves without affecting the black areas.

Masks are essential for processing - especially inverted ones
which when white adjust only in that area.

cheers
Allan

[gregbradley]

Thanks Allan. I use inverted masks so no problem there.

I have made an inverted mask, then used levels to bring the whiteness down so I get a layer that looks like yours with mostly the dust donuts now slightly white and most everything else black.

Now how do you apply that to the image to subtract out the background dust donuts from the original image? Difference layer? or Apply image/subtract?

Greg.

[alpal]

Originally Posted by gregbradley

Thanks Allan. I use inverted masks so no problem there.

I have made an inverted mask, then used levels to bring the whiteness down so I get a layer that looks like yours with mostly the dust donuts now slightly white and most everything else black.

Now how do you apply that to the image to subtract out the background dust donuts from the original image? Difference layer? or Apply image/subtract?

Greg.

No -
you work on a copy (just above the original like Louie does)
If you have the inverted layer mask showing the doughnuts on the right hand side -
if you adjust the curves by selecting that copy layer you will be able to lighten the darker areas only
which will be exactly where the dust doughnuts are.

cheers
Allan

[gregbradley]

Originally Posted by alpal

No -
you work on a copy (just above the original like Louie does)
If you have the inverted layer mask showing the doughnuts on the right hand side -
if you adjust the curves by selecting that copy layer you will be able to lighten the darker areas only
which will be exactly where the dust doughnuts are.

cheers
Allan

Thanks Allan - you're a genius! If I meet you along the way I owe you a dinner or a beer!

I did the inverted mask, lowered the brightness with levels to the point where the dust donuts and some other gradient junk showed up. I lightened it slightly so they were slightly white - not too white just a bit.

I then switched to the image the mask was connected to and used selective colour/black/black level and sdjusted a little to brighten the background only where the dust donuts were allowed to brighten because of the mask (white reveals in masks).

I then similarly adjusted the black slider in neutrals in selective colour and that finished it off. It seemed to remove them quite thoroughly without affecting anything else about the image.

Wow, that is a valuable technique and some minor variations of that could do a good job on gradient removal.

Resulting images posted to the same links as before.

Cheers,

Greg.

[alpal]

That's great Greg - you can fiddle around forever with these pics.

You should thank Louie Atalas - all his videos are here:

http://www.youtube.com/channel/UCJ5b6pFHBGe66vsuSaXb-0A

I consider all his videos essential & many people here
would benefit greatly from using the techniques shown.
Of course it does require Photoshop.

cheers
Allan

[gregbradley]

Originally Posted by alpal

No -
you work on a copy (just above the original like Louie does)
If you have the inverted layer mask showing the doughnuts on the right hand side -
if you adjust the curves by selecting that copy layer you will be able to lighten the darker areas only
which will be exactly where the dust doughnuts are.

cheers
Allan

Louie has definitely done us a great service there.
Also its how you apply these basics that makes a difference. This is a clever application. One for the astro toolbox for sure. Its fast, easy and very effective where other approaches are tedious, hit and miss or alter the image too much.

Greg.

[alpal]

Originally Posted by gregbradley

Louie has definitely done us a great service there.
Also its how you apply these basics that makes a difference. This is a clever application. One for the astro toolbox for sure. Its fast, easy and very effective where other approaches are tedious, hit and miss or alter the image too much.

Greg.

Yes - just remember that Louie was showing an inverted layer mask for a FILTER.

that filter can be anything you want:
brightness, contrast, sharpening, blurring, adding noise, levels, curves etc etc.

You can select the part of the image you want to change & adjust only that.
( also - remember to blur your layer mask & sometimes feather it )

Your image could actually do with some more contrast in the brighter areas
at the center of the galaxy - but only a touch
which could be adjusted with opacity.

However it's never ending - you just have to know when to stop.

cheers
Allan