NGC 4945, dust deserts. Now with 46 hours of exposure

[DJT (David)]

Beautiful image, MnT. The colour is great and the stars are a-popping.

Nicely done.

[Placidus (Mike and Trish)]

Originally Posted by Peter Ward

I am in awe of "M&T imaging Inc."

Your own code to create this image is no mean feat. (my last piece of code was in the Jurassic era of Fortran IV...)

What an impressive result.

P.S Dare I say....it also rather looks like a fish swimming towards me.

Thanks very muchly Peter. This one doesn't remind me of any particular life form, though I expect there is life there somewhere.

Originally Posted by DJT

Beautiful image, MnT. The colour is great and the stars are a-popping.

Nicely done.

Cheers, David!

[Placidus (Mike and Trish)]

We now have a total of 13 hrs Lum (in 1hr subs) plus 11 hrs each RGB (in 30 min subs).

The new beastie is here.

We offset the new data by a couple minutes of arc so that artifacts would not line up. Seems to have helped.

We hoped that the extra colour data might make the colour of the main galaxy more definite. Almost nothing happened. Main thing that happened with the extra data is the background is less gritty, with less artifact, and there are more faint fuzzies visible.

In processing, we fixed the green gradient that Greg noticed. That removal of the green gradient produced the expected very slight change in the colour of the galaxy.

Cheated a tiny bit and did some local brightening round the little face-on spiral that's at about 2 o'clock from the main one.

Overall, we're pleased to have put in the extra work. Thanks to everyone for the suggestions.

Best,
Mike and Trish

[Atmos (Colin)]

The galaxy appears a bit more defined now, the extra data has certainly helped, especially with some of the stars along the right side of the image. The background is a lot nicer, especially around the brighter areas and the background galaxies are showing by better and pop out more.

What I do prefer from the original is the stars. They are a lot softer, have a far gentler fall off. They are more distracting in the newer one, sharpening? I have given up sharpening my images because it tends to leave the stars with a bright core and then a hard fall off.

The extra data has helped

Edit: On review, not sharpening, they're just a fair bit brighter over the original

[Placidus (Mike and Trish)]

Originally Posted by Atmos

The galaxy appears a bit more defined now, the extra data has certainly helped, especially with some of the stars along the right side of the image. The background is a lot nicer, especially around the brighter areas and the background galaxies are showing by better and pop out more.

What I do prefer from the original is the stars. They are a lot softer, have a far gentler fall off. They are more distracting in the newer one, sharpening? I have given up sharpening my images because it tends to leave the stars with a bright core and then a hard fall off.

The extra data has helped

Edit: On review, not sharpening, they're just a fair bit brighter over the original

Thanks for your close and accurate inspection, Colin. I agree with your sentiments. I'm finding a global sharpening can produce nasty sequin-like stars, and am trying to avoid that. My goal was to just brighten the numerous extremely distant galaxies which were too gritty to see before, but are now plausibly there.

I produced two versions of the image. The first was optimized for the main galaxy details, strongly sharpened, which showed the details of the main galaxy nicely but had very ugly stars. The second version was optimized for the background and extremely distant galaxies: wavelet filtered, brightened, but not sharpened.

I then combined the two images according to the 150 pixel low pass filtered regional brightness, thus getting the sharp version in the galaxy core, the bright smooth version in the background, and pro rata in between. (This is analogous to using two layers and a mask in PhotoShop).

The vast majority of stars got brightened a lot and sharpened a tiny bit.

Agreed that the stars do look a tad intrusive now, but attempts to use a more conventional star mask produced nasty transition zones, and I was reasonably happy with the current version, which I think achieved the goal of showing the distant fuzzies better than before.

Perhaps one could mount an argument that distant galaxies are not the point of the image, and should be sacrificed, but my motivation was that I was really rather surprised to see any at all, given that we're in full-on Milky Way here, and so I wanted to show them.

Thanks again for looking so closely, and again, you are right, and I broadly agree with you.

Very best,
Mike

[Shiraz (Ray)]

that really is an exceptional image, with lots of interesting detail glowing through the dust and stars - and very careful management of the colour. This has to be one of your best, bravo .

thanks for the insight into your processing technique.

[Atmos (Colin)]

That is the beauty of image processing, every time we do it it is different! Taking the new one on its own, I wouldn't fault it, I do have your previous one to compare against though

As a whole it is a better image and the stars are not bad, just preferred the softness they once had I do like that you have brought out the fainter stuff, I wasn't sure if that was on purpose or if it was a consequence of extra luminance (better signal to noise).

What I do like about PixInsights PixelMath is that the expression:
iif($T<0.1, sh, $T)

What this would do is any pixel on the target ($T) image that has a value less than 10% will be replaced by the sharpened (sh) image. This will allow you to increase the background without effecting the stars.

I've never actually tried it but I have tried something similar to increase the background... Just hasn't worked very well. Increases the background but doesn't graduate as it goes.

[Placidus (Mike and Trish)]

Originally Posted by Shiraz

that really is an exceptional image, with lots of interesting detail glowing through the dust and stars - and very careful management of the colour. This has to be one of your best, bravo .

thanks for the insight into your processing technique.

Originally Posted by Atmos

That is the beauty of image processing, every time we do it it is different! Taking the new one on its own, I wouldn't fault it, I do have your previous one to compare against though

As a whole it is a better image and the stars are not bad, just preferred the softness they once had I do like that you have brought out the fainter stuff, I wasn't sure if that was on purpose or if it was a consequence of extra luminance (better signal to noise).

What I do like about PixInsights PixelMath is that the expression:
iif($T<0.1, sh, $T)

What this would do is any pixel on the target ($T) image that has a value less than 10% will be replaced by the sharpened (sh) image. This will allow you to increase the background without effecting the stars.

I've never actually tried it but I have tried something similar to increase the background... Just hasn't worked very well. Increases the background but doesn't graduate as it goes.

Thanks again, Ray and Colin.

In a fit of madness we did yet another 8 hours of luminance, and tried hard to do something along the lines discussed above. The large faint spiral came out even better, but the stars remain big and fat. After our success with H-alpha on NGC 300 and Barnard's, Trish is keen for us to add some H-alpha in to 4945 as red, but we've headed off to some galaxies in Hydra for the new moon. No more fiddling with this one for a while.

Very best,
M & T