M31

[admin]

Thank you Ivo for your comments and your time. It is exactly what I needed to improve ! I did not knew if the problem was from my data set or from its development. Now I know Yes my flats were not correct and I will try longer exposures. I understand that the signal/noise ratio is greatly improved with longer exposures. What would you recommend: 2 minutes ? I am still confused about the ISO to use, even after a careful reading of this article http://dslr-astrophotography.com/iso-ds ... otography/ With my camera (nikon D5600) if I understand well it is of no use to increase the ISO above 200 which seems quite low to get a signal with a galaxy.
It might be that the sky will be clear here tonight. If so I will give a new try.
Thanks again

No problem - looking forward to what you come up with! I'm here to help.
There is a lot of misinformation on how ISO settings work on digital cameras. Your sensor only has one sensitivity. It is fixed and cannot change. All the ISO (usually) does, is multiply (or divide) the signal in the digital domain (once it has already been captured by the sensor in the analog domain). Ideally you don't want the signal to be multiplied (because that just overexposes the image quicker), nor do you want the signal to be divided (because then you lose out on photon counts).

E.g. if I captured this 2x2 pixel image (numbers indicate brightness, where 100% is max brightness);

25%, 50%
75%, 100%

And then use an ISO that multiplies these numbers in the digital domain, say by 2.0x, I get

50%, 100%
100%(clip), 100%(clip)

Three pixels are now fully exposed "white" (100%). The first pixel (25%->50%) is now brighter, but any noise will also have been multiplied by 2. Also, because I multiplied everything by 2, there will never be odd numbers like 51% or 63%. The dynamic range that these odd numbers could use is now going to waste. It's just a trick to make it seem like your camera is more sensitive, but it's just that; a trick!

You can see that picking a multiplier other than 1 does nothing and actually makes things worse. If ISO 200 corresponds to a multiplier of 1 (also known as "native ISO") for your camera, then keep using that ISO setting.

Hope this helps!

[Misange]

I understand the example.
Now what if the pixels are
0% 0%
25% 50%
It might be that even multiplying by 2 the 0% are still below the detection limit.
So I guess that the iso must be set to the minimum that allows showing the faintest signal I want to see ? Accumulating pictures will increase the signal (and the noise). On a pratical point of view, how can I choose the number of images to take, the ISO and the exposure ? (other than spending several nights out and having just unuasable results !)

[maxchess]

It might be useful for you to know that I captured this M31 image using similar equipment.
https://astrob.in/378298/0/
I took 60 x 60 secs exposures plus 60 each of Darks, Flats & Bias frames using a Canon 1100d with a 300mm f4.
Your 300mm lens is f6.4 which lets in 2.5 x lets light (log scale) your exposure was 15 secs so combing them that is 10x less light than I used.
Plus the calibration frames remove the various types of noise.
All frames stacked in DSS and processed using Startools (what else?!)

My camera was mounted on an SW AZGT in EQ mode, but unguided. What sort of tracking mount did you use?

So I think Ivo is right not enough data and not enough calibration.

Max

[Misange]

Thank you Max for your help. Nice picture !
I have a star adventurer mini wifi as a mount. Unguided.
The sky seems clear this evening so hopefully, I will try the settings that you recommend. Is it necessary to use as many darks and flats as the images ? I read that a 1/3 ratio was sufficient.

[maxchess]

1/3 should be OK, but I always do more to be sure, don't forget bias frames
best of luck,
Max

[admin]

I understand the example.
Now what if the pixels are
0% 0%
25% 50%
It might be that even multiplying by 2 the 0% are still below the detection limit.

There isn't really something like a detection limit. There is only the probability of a photon being converted into an electron (also known as Quantum Efficiency).
Again, ISO doesn't set a detection limit, threshold or sensitivity factor, there is only one sensitivity for your camera. ISO only scales the actual amount of photons counted and that is it. It doesn't make your sensor more or less sensitive.

So I guess that the iso must be set to the minimum that allows showing the faintest signal I want to see ?

No, you should set the ISO to the native ISO of your camera's sensor and leave it there.

Accumulating pictures will increase the signal (and the noise).

Accumulating multiple pictures and stacking them will increase the signal relative to the the noise (e.g. noise decreases).

On a pratical point of view, how can I choose the number of images to take, the ISO and the exposure ? (other than spending several nights out and having just unuasable results !)

Choose the native ISO of your camera (which seems to be 200). The more images you take the better.
Exposure should be set as long as you can get away with. This depends on light pollution and skyglow, it also depends on how good your tracking is.
As a rule of thumb, the sky background should occupy no more than 1/3 of the histogram. The sky background correlates with the peak of the histogram. Most (all) cameras have a histogram function.

Hope this helps!

[Misange]

Thanks again Ivo for your precise answers. I will have to wait a few days before the conditions are set for a new try. I will let you know.

[Misange]

HI
I made a new set of images and the corresponding FTS is there
https://drive.google.com/file/d/1Hnx5YF ... sp=sharing
I fell its better but I still struggle to get something nice with this. Can you help me showing the steps that you follow ?
Thanks a lot

[admin]

Wow, that's a huge step in the right direction!
This dataset can now be processed with a relly standard workflow, without hacks or problems to work around. Excellent!

Processed as follows;
--- Auto Develop
To see what we are working with.
We can see great improvements! Flat fielding has worked well.
We can see oversampling, a red bias, some stacking artefacts. The only issue remaining is star elongation - a field flattener may help here. You may also wish to make sure your camera is sitting pefectly flush with the focuser (e.g no tilt or play).
--- Crop
Cropping away stacking artifacts.
Parameter [X1] set to [800 pixels]
Parameter [Y1] set to [847 pixels]
Parameter [X2] set to [5227 pixels (-789)]
Parameter [Y2] set to [3714 pixels (-302)]
Image size is 4427 x 2867
--- Bin
To reduce noise and make use of oversampling.
Parameter [Scale] set to [(scale/noise reduction 35.38%)/(798.89%)/(+3.00 bits)]
Image size is 1566 x 1014
--- Wipe
I used the Vignetting preset, as there was still some sort of gradient or light fall-off visible. It's much improved over the previous dataset though.
Parameter [Dark Anomaly Filter] set to [5 pixels]
--- Auto Develop
I click & dragged a Region of Interest over part of M31; the goal here is to give AutoDev a sample of what the stretch should be optimized for (e.g. the interesting stuff and not the empty background for example).
Parameter [Ignore Fine Detail <] set to [6.0 pixels]
Parameter [RoI X1] set to [617 pixels]
Parameter [RoI Y1] set to [406 pixels]
Parameter [RoI X2] set to [891 pixels (-675)]
Parameter [RoI Y2] set to [632 pixels (-382)]
--- Deconvolution
Thanks to binning, there is even enough signal now to slightly enhance fine detail in the core.
Auto mask. Default settings.
--- HDR
M31 has a bright core and whenever I process M31, I tend to run the HDR module with default settings to tone down the core a little, as there are some small details hidden in the highlights.
Experiment to taste obviously!
--- Color
There are some problems with the star colors (stacker unable to align properly?), which causes StarTools to get the initial color balance wrong.
I used the MaxRGB mode combine with some basic knowledge about the object to color balance instead. You can read more about this here. I used a combination of White balancing in MaxRGB mode and White balancing by known features and processes.
Specifically, you'd be looking for a "milky" yellowish core (older stars, less star formation), bluer outer rim (younger blue stars, more star formation), red/brown dust lanes. With deeper data you'll star seeing purplish HII areas and knots as well (e.g. the same sort of nebulas, emissions and colors likethe HII areas like M8, M16, M17, etc. in our own galaxy).
Parameter [Dark Saturation] set to [8.10]
Parameter [Bright Saturation] set to [3.30]
Parameter [Saturation Amount] set to [230 %]
Parameter [Blue Bias Reduce] set to [1.21]
Parameter [Green Bias Reduce] set to [1.14]
Parameter [Red Bias Reduce] set to [1.36]
--- Wavelet De-Noise
Default parameters.
Parameter [Grain Size] set to [6.0 pixels]

As you can see, the above workflow is pretty simple and follows the flow of the workflows used in most videos. I hope this gets you on your way!

190905-andromede.jpg (323.37 KiB) Viewed 7786 times

[Misange]

Whaou ! thanks a lot Ivo. I have now a goal : repeat your treatment first to understand it then try to do it by myself. I am just leaving for holidays, I hope that I will have nice opportunities to do nice images. Tnaks again