Milky Way images?

[67ssdan]

Okay, I've messed with it some more and got DSS to do its thing correctly, as I eluded to previously. However, when I follow the guy's instructions that you posted previously, it seems to all go sideways in the wipe section. The image is taken of the Milky Way this time of year, which obviously means it's just ahead of sunrise. There was some early pre-dawn light on the horizon already, and the software seems to just go nuts with that.

Here's an example after the auto-dev, but before wipe: https://drive.google.com/file/d/1KzX_Vk ... sp=sharing
Subsequently I tried skipping wipe and just messing with the color and got this: https://drive.google.com/file/d/1jrwf5R ... sp=sharing Better... but still a mess.

Here's the FTS file I'm screwing with: https://drive.google.com/file/d/1VzQnIf ... sp=sharing

After this, I also had DSS spit out the stacked images and I pulled them into Photoshop and screwed around with them as I normally would. Definitely less noise going about it that way, but I can't seem to pull the colors out that Sequator does, and there is a definite orange cast to it all, just like with StarTools.

Is this doable at all?

Thanks in advance,

Dan

[admin]

No worries - here to help.

I just found this great link with best ISOs to use for various models, include the 60D and 6D;

http://dslr-astrophotography.com/iso-va ... n-cameras/

Safe travels!

EDIT: Now for your data set;

it seems to all go sideways in the wipe section. The image is taken of the Milky Way this time of year, which obviously means it's just ahead of sunrise. There was some early pre-dawn light on the horizon already, and the software seems to just go nuts with that.

Wipe will be able to handle it just fine (see further below), however the biggest problem with any non-object-of-interest light is that it adds an overwhelming amount of signal + its noise component. Wipe can subtract the estimated signal, but not its noise component. This is the reason why astrophotographers only tend to image under the darkest skies they can find.

If you could only watch one video ever in your life about atrophotography, let it be this one; https://www.youtube.com/watch?v=EO4QFb3ydNM

In it, Craig Stark (author of Nebulosity and PHD) explains why you'd want dark skies.

Here's an example after the auto-dev, but before wipe: https://drive.google.com/file/d/1KzX_Vk ... sp=sharing

That looks as expected.

Subsequently I tried skipping wipe and just messing with the color and got this: https://drive.google.com/file/d/1jrwf5R ... sp=sharing Better... but still a mess.

As I mentioned earlier, you'll want to use Wipe with a mask that masks out (e.g. make non-green) the terrestrial bit. It's also important to remove (using Crop) any stacking artefacts (here to the left of the image). Anything that is not real sky signal needs to be masked out.

Is this doable at all?

Definitely.

Here goes (using 1.4.330);

--- Auto Develop
To see what we got.
We can see oversampling/star trailing. We can also see a lot of noise. We can see stacking artefacts to the left, top and bottom of the image. There is some stacking mis-alignment in the upper right corner.
--- Bin
Binning trades resolution for noise reduction. I'm binning hard as the image is very noisy and the star trails are "smearing out" detail anyway.
Parameter [Scale] set to [(scale/noise reduction 25.00%)/(1600.00%)/(+4.00 bits)]
--- Crop
Parameter [X1] set to [14 pixels]
Parameter [Y1] set to [11 pixels]
Parameter [X2] set to [1294 pixels (-6)]
Parameter [Y2] set to [863 pixels (-3)]
--- Wipe
As said above, create a suitable mask, should look roughly like so;

Screenshot.jpg (89.94 KiB) Viewed 7579 times

Parameter [Dark Anomaly Filter] set to [5 pixels]
I'm bumping up Dark Anomaly filter to 5 pixels as a precaution (it filters out smaller darker-than-real-celestial-background pixels if we missed them in our mask).

--- Develop
The signal is very weak and the noise is extreme. AutoDev does not tend to deal well with this, so I'm using a manual Develop.
Choose "Redo stretch". Develop has a semi-auto mode though - just press "Home In" a few times until it stabilises.

Now that we have finalised the luminance/brightness/detail, it's time to do final color calibration;
--- Color
The Color module tends to come up with a pretty decent color balance all by itself. But sometimes it needs some help (especially images like these unfortunately where things near the horizon appear redder due to the atmosphere).
Things to look out for are a good random distribution of star colors (equal amounts red->orange->yellow->white->blue).
Another thing to look out for is purple/pink HII areas.
The Max RGB view should not show large contiguous areas of dominant green (speckles and noise is ok though). If it does, dial back the green parameter (as I did here). Very few things in outer space are green-dominant (there a few exceptions).
However, this is also where things get exciting and where all using an AP-specific program starts to pay off - you've picked up some purple HII areas. If I zoom in there is a purplish blob with a cyanish core - this is M8 if I'm not mistaken;

Autosave002(4)_purple.png (42.76 KiB) Viewed 7579 times

Parameter [Cap Green] set to [To Yellow]
Parameter [Dark Saturation] set to [3.50] to introduce more color in the dark regions.
Parameter [Blue Bias Reduce] set to [1.17]
Parameter [Green Bias Reduce] set to [1.70]
Parameter [Red Bias Reduce] set to [2.63]

No noise reduction applied when switching off Tracking (do this to taste).

What you end up with is something like this;

Autosave002(4).jpg (228.18 KiB) Viewed 7579 times

You just need many, many more exposures under the darkest skies you can find. Tracking will help, but you may be able to get away with no tracking if you keep the exposures short and reduce rsolution.

Hope this helps!

[67ssdan]

WOAH! Yeah, that's pretty awesome!

Okay, so you mentioned getting many many more exposures, how many would you aim for in a situation like this? I don't have a tracker, so keep that in mind. Also, is the noise a function of not having enough exposures, or my camera... or all of the above?

Thank you so much for the links, commentary, and help in general. This is extremely cool!

Dan

[67ssdan]

One more thing, I've noticed misalignment in that exact are of my images before, stacked with other programs. Any ideas on what part of my kit would be causing this? I'm assuming it's a function of something weird about the lens (Rokinon 14mm, f2.8), and if so, what can be done?

[67ssdan]

No worries - here to help.

I just found this great link with best ISOs to use for various models, include the 60D and 6D;

http://dslr-astrophotography.com/iso-va ... n-cameras/

Okay, one more question on this... so it seems logical to me that, given what these guys are saying about ISO and dynamic range, you should really just use the suggested ISO all the time, and not just for astrophotography... right? Am I still thinking about that wrong?

Thanks,

Dan

[admin]

Okay, so you mentioned getting many many more exposures, how many would you aim for in a situation like this? I don't have a tracker, so keep that in mind. Also, is the noise a function of not having enough exposures, or my camera... or all of the above?

Keep shooting sub frames. That's all there is to it. The longer the exposure you can get away with (minding start trailing), the better. Try turning off "long exposure mode" (setting/name varies per model) where your camera takes another exposure after your "real" exposure of the same duration with the sensor covered. It's to counter hot/dead pixels, but it's a bit of a waste of time, since you will be stacking multiple images. In the meantime the sky is drifting as well.
E.g. if you find your 30 second exposure seems to take exactly 60 seconds, that mode is still on.

Try to minimize any terrestrial light sources. Try to minimize any solar system light sources; wait for "astronomical" sunset and don't image when the moon is out. Find the darkest site possible. Avoid the light domes of towns and cities in the distance. Aim above the horizon if you can (due to blue extinction I mentioned before)

Thank you so much for the links, commentary, and help in general. This is extremely cool!

It is!
Just wait until you start researching/interpreting the colors you captured. You can tell a crazy amount of things (chemical composition, temperature, processes going on, past, future, current neighborhood characteristics, and more) by just looking an object's color.

One more thing, I've noticed misalignment in that exact are of my images before, stacked with other programs. Any ideas on what part of my kit would be causing this? I'm assuming it's a function of something weird about the lens (Rokinon 14mm, f2.8), and if so, what can be done?

Not entirely sure - it's clearly DSS, messing up alignment. You might be able to remedy it with one of its settings - use more or fewer stars to calibrate against?

Okay, one more question on this... so it seems logical to me that, given what these guys are saying about ISO and dynamic range, you should really just use the suggested ISO all the time, and not just for astrophotography... right? Am I still thinking about that wrong?

Well, a big difference is that in deep space, things are just ridiculously faint - you're almost never at risk over exposing (there are one or two exceptions, for example M42) during long exposures. You'll want to fill up your buckets (electron wells) with photons as quickly as possible (but without meddling with their water levels). Also, you don't care about fast moving objects, nor care about depth-of-field aesthetics. Your aperture is typically wide open - not letting any photon go to waste.

All that is different with terrestrial/daylight photography, over-exposing happens quickly, so you do want some way of making sure your electron wells don't fill up too quickly. You can use a different means, aperture, exposure time, and - yes - ISO (if the former two are not applicable, suitable or enough), depending on what effect you are after, focus, speed of the object, etc.