Page 1 of 3

M27 - need a hand, please!

Posted: Mon Sep 24, 2018 1:01 am
by devonshire
Hi!

I'm still trying to map the ST workflows I read about, to the images I actually have. Doesn't always work...

I wonder if someone better at this than I, can take a look at my M27 stack? I'm finding that if I clean up the noise in the background, it clobbers the faint dumbbell image. Conversely, if I stretch enough to bring out the dumbbell, the image drowns in noise... Argggh! Going round in circles with this one.

The .zip contains both the DSS stack, and enough DSS logs to hopefully answer any questions about stacking options.

Thanks for any help with this, and if you do take a shot, please note the steps you used.

Thx!

- Bob


Here's a link to the stack:

https://www.amazon.ca/clouddrive/share/ ... mkEhQ0ydA5

Re: M27 - need a hand, please!

Posted: Mon Sep 24, 2018 2:00 am
by admin
Hi,

Thanks for uploading!
The main problem is that there is unfortunately not very much signal.

I noticed though looking at the DSS log that you shot at ISO 200. ISO values in the digital age are a bit of a strange beast; your camera has only one native sensitivity ("native ISO") which is the sensitivity of the sensor. Other ISO values are emulated by scaling up or down the signal, usually in the digital realm (which usually involves a simple multiplication or division, inventing or discarding signal). You'll want to avoid this scaling as much as possible by imaging close to your native ISO value (usually around 800 or 1600).

Is there any reason why you think you need to drizzle x2? it appears to me the data is already over sampled at native resolution and drizzle is making this even worse! It doesn't help noise grain either and you end up with a huge dataset for no good reason...

As for processing, if you are left with an object on top of an otherwise "empty" noisy background;
1. Be sure to specify a RoI in AutoDev when using it to stretch your data.
2. Try using the Isolate preset in the Life module (make sure all pixels are selected/green in the mask). It can help push back noisy backgrounds and busy star fields and "Isolate" bigger objects.

Hope this helps!

Re: M27 - need a hand, please!

Posted: Mon Sep 24, 2018 1:53 pm
by devonshire
Thanks, Ivo!

My camera is a Nikon D5300, and its' native ISO is said to be 100. The advice on CloudyNights about shooting in light pollution with that and similar cameras, is to use ISO 200 - so that's how I ended up with that.

I had drizzle on from a previous image where I thought it helped. I'll re-stack later today without it, plus trying the other two points you mention, and report back.

Thanks again!

- Bob

Re: M27 - need a hand, please!

Posted: Mon Sep 24, 2018 2:31 pm
by admin
devonshire wrote:Thanks, Ivo!
My camera is a Nikon D5300, and its' native ISO is said to be 100. The advice on CloudyNights about shooting in light pollution with that and similar cameras, is to use ISO 200 - so that's how I ended up with that.
It seems you got good advice! :thumbsup:
http://dslr-astrophotography.com/iso-ds ... otography/
I had drizzle on from a previous image where I thought it helped. I'll re-stack later today without it, plus trying the other two points you mention, and report back.
Great - let us know how you get on!

Re: M27 - need a hand, please!

Posted: Tue Sep 25, 2018 1:04 am
by devonshire
Ivo,

Nope, no joy. Still stuck with light pollution rising alongside the dumbbell image and the whole thing just ending up murky - or a star-only image with the dumbbell and much of the small-star milky way background suppressed..

What I really wish I could do is lasso a patch of empty-but-light-polluted sky, and then ask ST to subtract everything that looks like that from the image.

A bit frustrating - AutoDev reveals lots of detail, but noise/light pollution is overwhelming it.

https://www.amazon.ca/clouddrive/share/ ... cfJY6jHoW2

- Bob

Re: M27 - need a hand, please!

Posted: Tue Sep 25, 2018 1:57 am
by admin
devonshire wrote:Ivo,
Nope, no joy. Still stuck with light pollution rising alongside the dumbbell image and the whole thing just ending up murky - or a star-only image with the dumbbell and much of the small-star milky way background suppressed..
Welcome to astrophotography, where noise is the bane of our existence. :evil:
What I really wish I could do is lasso a patch of empty-but-light-polluted sky, and then ask ST to subtract everything that looks like that from the image.
That's actually exactly what Wipe does (usually without the need to explicitly specify areas). :)

Going on from here, you might want drastically increase the exposure times. 30s is usually not a long time for signal of faint objects to stand out above your camera's noise floor.

Light pollution will add a lot to the noise (its average light level can be subtracted, as WIpe does, however its noise component cannot and will remain in your data). This is the reason why pristine dark skies are so sought after. What is your light pollution like (on the Bortle scale?)

It's a long shot, but if you leave out the calibration frames do noise levels improve at all?

Re: M27 - need a hand, please!

Posted: Tue Sep 25, 2018 2:38 am
by devonshire
Ivo,

I think I'm in Bortle 6 (Orange).

Whether it's actually working this way or not, Wipe *seems* to be acting uniformly across the image. Making an automated decision, rather than a guided one that targets the pixels I want it to target. Zoomed in, that light pollution is a pretty clear mottle that doesn't look much like stars.

My original plan was to take aboout 90 minutes, but that darn dumbbell ducked behind my roof and I lost sight of it before I was ready to give up for the night. :-)

The calibration frames? Do you mean just the bias frames, or both the bias and flats? There was dew on the lens that night, so I'm thinking that the flats, et least, are earning their keep...

Thx!

- Bob

Re: M27 - need a hand, please!

Posted: Tue Sep 25, 2018 4:35 am
by admin
devonshire wrote:Ivo,

I think I'm in Bortle 6 (Orange).
This will indeed increase the noise component contributed by light pollution a fair bit.
Whether it's actually working this way or not, Wipe *seems* to be acting uniformly across the image. Making an automated decision, rather than a guided one that targets the pixels I want it to target. Zoomed in, that light pollution is a pretty clear mottle that doesn't look much like stars.
Wipe's job is to model any signal bias gradients in your image and subtract them.
The mottle is simply random noise and you can't model noise because it's... well... random. :)
You can say; "well the noise centers around this particular value, sometimes randomly a little more, sometimes randomly a little less, so we'll subtract that one value everywhere".
That's what pretty much Wipe does (though it makes sure that negative values are not clipped to black).
The calibration frames? Do you mean just the bias frames, or both the bias and flats? There was dew on the lens that night, so I'm thinking that the flats, et least, are earning their keep...
Definitely! Glad you're taking flats; they're the #1 thing to improve your images with. I'm only bringing up your calibration frames because if something is wrong with them at all, they can impart noise on your final stack. I don't think that's the case here, but it's just something to check if you feel the noise levels are unreasonably high.

Re: M27 - need a hand, please!

Posted: Tue Sep 25, 2018 11:15 pm
by devonshire
Ivo,

Well... if the user were to select a sample of mottle for StarTools, would it then still be random? :-)

- Bob

Re: M27 - need a hand, please!

Posted: Wed Sep 26, 2018 2:11 am
by admin
devonshire wrote:Well... if the user were to select a sample of mottle for StarTools, would it then still be random? :-)
The mottle pattern? Absolutely! It is in no way, shape or form a pattern match for the mottle elsewhere.