DeepSkyStacker with Foreground Issue?

I'm attempting to figure out DeepSkyStacker. I've signed up for the Yahoo Group, but the mods have yet to give me access.

In the event I'm stacking images that have a bit of foreground, will that mess up the resulting stack?

Here is one of the images I'm playing with, which was taken this morning (not a serious attempt, I'm just trying to learn Backyard EOS and DeepSkyStacker):

Heya,

1) Foreground objects on stacks will not work right. Remember, the sky is moving, you and the foreground are not. So when you go to stack, the sky has moved and the foreground image has not. So focus on the sky portion of the stack, not the foreground. The foreground typically is better off left as a single exposure, use masks, and just keep a sharp and well exposed version of an exposure you make that is not worried about the sky and focuses on the foreground portion. Blend the two after you stack so you have a stacked sky and a foreground section that is appropriate. These are always composites. DSS does not do well with objects in the scene.

Follow this tutorial, it works.

Also note, you don't always need to stack. That first image you showed, with a mask/layer, you can just apply levels/curves/contrast to the sky and process the foreground separately and make a good image. Stacking is not always the answer. Poor stacking results in a noisier or weird image often. Stacking requires planning. You don't just take 10 exposures and stack and expect it to be better than a single frame. See the tutorial.

Very best,

Thank you! I'll review the tutorial.

I'm also testing RegiStax, AstroArt, Nebulosity, MaxIm DL, and RegiStar. I've found that RegiStar ignores the foreground completely and only focusses on the stars, but it doesn't read RAW files and I keep crashing it. AstroArt is really user friendly, but it stacks based on the foreground and not the stars: no matter what my setting is.

I would like like to use DeepSkyStacker. The developer is really actively involved with it.

Heya,

Just one other piece of advice: don't focus so hard on this software/stacking. You still need good images to start with. I suggest you start by getting good technique with achieving good single exposures. If you can nail good single exposures that are processable and able to produce good images that you enjoy, then start getting into stacking. Stacking doesn't add information, stacking removes information. It's all about noise reduction. The image is not enhanced, it's just cleaned up. But it takes a lot of frames to reduce that noise systematically without losing detail on things you didn't want removed (hence people stacking 30+ images).

And really, if you truly want to get into more astro, I would keep stacking on the back burner, and focus on a tracker instead. You'll get so much from a tracker that you'll fall over with the results. Stacking comes in handy with that too. But again, stacking is just post processing for the purpose of cleaning up noise. You still need good initial image(s) to work with.

Examples of my own just to kind of push the point:

Single exposure, with an inexpensive tracker, no stacking, no special processing, just contrast/levels were adjusted:

IMAGE LINK: http://www.flickr.com/​photos/mwise1023/13031​601445/

IMAGE LINK: http://www.flickr.com/​photos/mwise1023/11020​375105/

Those are lovely images.

I do have a tracker (an AstroTrac), which I've been playing with a bit. I have yet to successfully align it, but I'm sure I'll get there eventually. I may have to use Backyard EOS to drift align, as I can't see all of the stars I need to get it 100% (or even close to 100%).

It has finally warmed up (for at least a little while) so hopefully I'll get a little more practice over the next couple of weeks. My Astronomik CLS filter is in the mail, too.

I'm basically antsy and engaging in activity to keep myself focussed and entertained, which may ultimately be counter productive: until I get nice clear skies again. It looks like it will be Monday or Tuesday before I have enough time to play with aligning before the moon comes up. Hopefully I'll have clear skies.

Thank you for your intervention!

MalVeauX wrote in post #16755608
If you can nail good single exposures that are processable and able to produce good images that you enjoy, then start getting into stacking. Stacking doesn't add information, stacking removes information. It's all about noise reduction. The image is not enhanced, it's just cleaned up. But it takes a lot of frames to reduce that noise systematically without losing detail on things you didn't want removed (hence people stacking 30+ images).

This really should be part of a stacking sticky.

hollis_f wrote in post #16757647
This really should be part of a stacking sticky.

That is exactly what struck me the most.

I have been reading about exposure stacking to avoid the blowout of highlights in starclusters and the retention of less bright star data, which is essentially what the dataset I've been playing with was shot to get me thinking about for when I head out into the field.

Ultimately, that's pretty advanced stuff: much more advanced than where I'm at. MalVeauX's statement was a good reality check for me.

I'm an astro and stacking novice, so please educate me. I don't follow the logic behind the statement that stacking does nothing more than eliminate noise.

It seems to me that by superimposing images (i.e. stacking), that while any noise present will be reduced as it will vary from frame to frame (light and dark frames anyway), any signal present will be amplified since it is the same in each frame. To put it another way, the number of signal photons captured per pixel should be the same regardless of whether it is from one 60 minute exposure, or 60 one minute exposures. Thus, I'd expect to see in the final image much fainter objects in a stack than I would in an individual frame, contrary to the statement that stacking only removes information.

Is that right or am I missing something?

xarqi wrote in post #16760043
I'm an astro and stacking novice, so please educate me. I don't follow the logic behind the statement that stacking does nothing more than eliminate noise.

It seems to me that by superimposing images (i.e. stacking), that while any noise present will be reduced as it will vary from frame to frame (light and dark frames anyway), any signal present will be amplified since it is the same in each frame. To put it another way, the number of signal photons captured per pixel should be the same regardless of whether it is from one 60 minute exposure, or 60 one minute exposures. Thus, I'd expect to see in the final image much fainter objects in a stack than I would in an individual frame, contrary to the statement that stacking only removes information.

Is that right or am I missing something?

Heya,

The signal is not amplified, it's simply kept while the variable pixels (considered noise or hot pixels) are then removed. If the pixels you are referring to that were static, were amplified, they would all result in a white pixel and there would be zero detail. And you do get different amounts of photons, they do not just pass through solid objects and they bounce off things on the way, and the atmosphere changes how many you capture every nanosecond that your'e exposed. You do longer exposures to get photons from the same source to guarantee that it's treated like a static source and not just noise on super feint objects. It's why you have to expose longer for some thing, than others. There is a huge difference in photon gathering from 60 minutes to 60 x 1 seconds. Huge. Go test that, you'll result in a nearly white image possibly. You can even over-expose in space.

Very best,

I had much better luck last night with DeepSkyStacker:

Heya,

Can you post a single sub/light frame, not stacked, that was used?

Very best,

I just ran it again:

Heya,

So I tested it just off your JPEG, and honestly, I really recommend you spend less time trying to do stacks, especially with only a few frames. When you stack with less frames, you actually introduce noise and problems, not just retract them. I compared your stacked image to an edit of your single sub that you just posted, and with a simple curve adjustment and contrast adjustment, they looked essentially the same. That tells me that you need to work more on simply doing a single exposure with more information in it.

When you do an exposure, try adjusting it's color temperature in your RAW software. Then export and adjust contrast. You'll see the same image basically.

Again, stacking is useful when you have a good single image, but it's noisy, and then take like 20+ working light frames, some flats and darks and, again 20+ each, and generate a good master file that it then works from. But a single frame should still be a workable image. Stacking doesn't generate information--it takes information out.

Also, without EXIF, wasn't sure how long exposure time was. On that 6D, with a 24mm lens, you should be able to easily do 30 second exposures at 6400 and get a clean single image.

Very best,

MalVeauX wrote in post #16772811
So I tested it just off your JPEG, and honestly, I really recommend you spend less time trying to do stacks, especially with only a few frames.

I appreciate the time and effort you've invested in this. I really do. But, can you please stop trying to convince me to not learn something that I would like to learn? I'm not trying to be rude. I've a dozen, or more, things to learn about this as time and conditions permit. I'll learn it all, in time.