Total Noob needs help: Milky Way

[Effix]

HA! It was in my spam folder from the beginning... To think that meanwhile I have processed M42 at least 3 times "to try" without being able to save it...Anyway. Here's the original file of M31: https://www.dropbox.com/s/7l0iuz2e4q2pnrj/M31.fts?dl=0

I haven't tried to work on it because of this blue bias and the noise. The capture parameters are the same as for M45 and M42, who were taken the same night, so that I didn't need to do any more flats or dark frames :p I'll post the orignal file of M42 later if you wish to have a look at it.

[admin]

Great!

Yeah, your M31 could definitely use more signal - the disc is just starting to show.

[Effix]

Wow, nice job Ivo! I gave it a try but was never able to reach such a good result with M31. I see I have a lot to learn about StarTools would you have a workflow to provide me with?

Anyway, after being away for New Year, I am proud to present you my first picture processed through StarTools: M42. The pic was taken the same night and the same way as the two others. Here is the original stacked picture: https://www.dropbox.com/s/2pzbhhrpikxzt ... 2.FTS?dl=0
and here is the result I managed to get: https://www.dropbox.com/s/fx0b7ubb4quqd ... .tiff?dl=0

I'm sure there are many improvements to do, like for instance I have a lot of trouble to deal with the excessive noise. Also, I kinda cheated with the nebulae, by correcting the colours individually (that is, creating a mask on each of them and playing with the colour bias sliders). As a result, M42's little friend (NGC 1973-75-77 I think) appears maybe too blue. But for an official first, I am very glad.

Any comments for improvments welcome

[admin]

Wow, nice job Ivo! I gave it a try but was never able to reach such a good result with M31. I see I have a lot to learn about StarTools would you have a workflow to provide me with?

Absolutely! Processing is hard if your signal is very faint.

I reprocessed your image with 1.4 alpha and came up with this;

M31.jpg (35.58 KiB) Viewed 10643 times

Workflow as follows;

--- Auto Develop
To see what we got. We can see a blue bias, stacking artifacts and heavy noise.
--- Crop
First I crop the galaxy.
Parameter [X1] set to [1588 pixels]
Parameter [Y1] set to [1311 pixels]
Parameter [X2] set to [2875 pixels (-1221)]
Parameter [Y2] set to [2256 pixels (-828)]
--- Bin
Next I bin the data to 50% to reduce noise grain (note I binned to 35% in the previous version).
Parameter [Scale] set to [(scale/noise reduction 50.00%)/(400.00%)/(+2.00 bits)]
--- Wipe
Default parameters, masked out M31, just in case.
Parameter [Dark Anomaly Filter] set to [3 pixels]
--- Auto Develop
Final stretch. Selected Region Of Interest over M31. It's tricky to get right - you don't want too much noise showing up.
Parameter [Ignore Fine Detail <] set to [3.3 pixels]
Parameter [Outside ROI Influence] set to [4 %]
--- Life
Isolate preset with fuill mask set (Clear, Invert, Keep in Mask generator).
--- Color
Final color calibration. Again, a bit tricky. The better the data, the better StarTools default color balance is. I hade to make a good few tweaks. The tweaks I made are by reasoning about the object's and stars' known colors.
M31's core is yellow (old stars dominate), with a blue outer ring (young stars dominate), transitioning through purple in between. The star field around M31 should show a good distribution of all star temperatures from cool red to hot blue and everything in between (orange, yellow, white). The MaxRGB view also helps you to ensure that green channel dominance is countered where needed (green dominant objects in space are very rare).
Parameter [Cap Green] set to [To Yellow]
Parameter [Dark Saturation] set to [8.80]
Parameter [Saturation Amount] set to [355 %]
Parameter [Blue Bias Reduce] set to [7.60]
Parameter [Green Bias Reduce] set to [3.04]
--- Wavelet De-Noise
Parameter [Color Detail Loss] set to [21 %]
Parameter [Brightness Detail Loss] set to [12 %]
Parameter [Grain Size] set to [9.3 pixels]
Parameter [Read Noise Compensation] set to [13.78 %]
Parameter [Smoothness] set to [79 %]

Anyway, after being away for New Year, I am proud to present you my first picture processed through StarTools: M42. The pic was taken the same night and the same way as the two others. Here is the original stacked picture: https://www.dropbox.com/s/2pzbhhrpikxzt ... 2.FTS?dl=0
and here is the result I managed to get: https://www.dropbox.com/s/fx0b7ubb4quqd ... .tiff?dl=0

I'm sure there are many improvements to do, like for instance I have a lot of trouble to deal with the excessive noise. Also, I kinda cheated with the nebulae, by correcting the colours individually (that is, creating a mask on each of them and playing with the colour bias sliders). As a result, M42's little friend (NGC 1973-75-77 I think) appears maybe too blue. But for an official first, I am very glad.

Any comments for improvments welcome

Congrats on getting some imaging time in! First rule of AP; keep at it!

Here is what I got from your data;

M42_2.jpg (52.8 KiB) Viewed 10643 times

Same thing applies though; obtain more and better data and you'll find processing it will become much easier!

Workflow as follows;

--- Auto Develop
To see what we got. Lots of noise, and a yellow bias.
--- Crop
Frame M42 and Running Man better (though you may also have the beginnings of the Flame nebula in your frame outside this crop).
Parameter [X1] set to [1029 pixels]
Parameter [Y1] set to [1081 pixels]
Parameter [X2] set to [2849 pixels (-1247)]
Parameter [Y2] set to [2338 pixels (-746)]
--- Bin
Parameter [Scale] set to [(scale/noise reduction 50.00%)/(400.00%)/(+2.00 bits)]
--- Wipe
Parameter [Dark Anomaly Filter] set to [7 pixels]; always good to bump this up in the face of severe noise.
--- Auto Develop
ROI over M42
Parameter [Ignore Fine Detail <] set to [3.1 pixels]
Parameter [Outside ROI Influence] set to [8 %]
--- Deconvolution
Worth a try... I had to touch up the mask that the Auto mask generator came up with to include the core of M42 (but not any of the stars around it)
Default values. Small improvement.
--- Wavelet Sharpen
Using same mask as we created for Decon.
Parameter [Amount] set to [243 %]
Parameter [Small Detail Bias] set to [97 %]
--- Life
Like in M31, using Isolate preset with full mask set to push back noise.
--- Color
Tricky again. Had to go by the stars and the knowledge that HII areas like M42 show mostly blue (H-beta), red (H-alpha) or purple/pink (mixture of Ha and Hb).
Parameter [Cap Green] set to [To Yellow]
Parameter [Dark Saturation] set to [4.90]
Parameter [Blue Bias Reduce] set to [1.56]
Parameter [Green Bias Reduce] set to [1.73]
--- Wavelet De-Noise
Parameter [Color Detail Loss] set to [21 %]
Parameter [Brightness Detail Loss] set to [11 %]
Parameter [Grain Size] set to [20.9 pixels]
Parameter [Read Noise Compensation] set to [30.62 %]
Parameter [Smoothness] set to [83 %]

Hope this helps!

[Effix]

Hi Ivo and the community!

Sorry for digging this up after over a year, but as I started this topic and I'm now going on M42 again, I thought it'd be better than creating a new one. I didn't work on any new picture in over a year simply because I don't have much time to take my gears out, and last time I tried, in August, my pictures were a disaster: a few stars lost in a cloud of noise. Anyway, I am shortly back into business, and as mentionned I have decided to go on M42 again, trying to follow the many advice you gave me, as much for picture acquisition as for StarTools processing.

First, data: Olympus E600 with a 70-300 zoom lens set to 300mm, 1:5.6 aperture, 800ISO, 15 shots of 2 minutes each +/- 10s as I was controlling the shutter manually. about 12 darks, 10 to 15 flats and dark flats, and 15 to 20 offsets (exact value I don't remember, but it's more or less that). I could have taken more shots but several events prevented me from going further: my camera battery died, my equatorial mount batteries were dying too (for those wondering it is a Vixen Polarie mount, set on a good but not professional tripod. It's a nice little box ), clouds were coming (Orion wasn't as visible as when I started), and my feet were getting cold too. Stacking with DSS of course, using the parameters I found on another thread about which Ivo said they were good. Here are the results:

First, I followed the tutorial video which explains how to deal with "poor" images, and it gave me this:

https://www.dropbox.com/s/ynr27yakxxvz9 ... .tiff?dl=0

A lot of noise, though the extensions of the nebula are visible a bit. Yet I was not satisfied with it, because of this noise, which is the problem I'm always dealing with. So I decided to give it another try, this time following almost exactly the process I was given last year in this very thread. New result:

https://www.dropbox.com/s/cpbsovyxyiw4x ... .tiff?dl=0

The noise is totally gone, which is great! However I have lost a lot of colours in the stars, or even lost stars themeselves. The core of the nebula is looking okay, but I am now missing the outer edge. Also I've applied the "repair" module to fix the stars, and they look really fat, especially in the nebula core where they really seem to have been put on top, whereas in the first pic they are more blended in. Anyway I am rather satisfied with this second result, as I now know how to reduce the noise. So thanks Ivo for the good advice Now I need to know how to keep the noise low without losing too much data in the signal.

I know some things I made which may have reduced the quality of the picture. First I set the zoom to maximum to get a closer view of the nebula, but I must have lost some light there (though at minimum zoom the aperture is 1:4, so not that much higher). Also having a city south, I used a light pollution filter (namely the Hoya red intensifier) to reduce it, but it may have taken some colours off the image.

For those who want to play around, this is the original file that DSS spat me out.

https://www.dropbox.com/s/2pzbhhrpikxzt ... 2.FTS?dl=0

Thanks on advance for any advice!

[Effix]

Update: I gave it another try tonight, here's the result as a PNG, so that it's not as heavy as the other files.

https://www.dropbox.com/s/sdw635r95es24 ... 2.png?dl=0

I also made a bit of editing in Gimp to increase the saturation a bit more once I was done with StarTools. I am more pleased with the results: I got better colours this time and the stars don't look strange; I didn't try to fix them this time though. Only because of the heavy nooise removing I'm losing a lot of details, and as a result the nebula looks a bit like a water painting, but I can also pretend it's an artistic interpretation of the nebula

Question: What is the Bin module for? I understand it reduces the size of the image, but not only. This time I didn't use it too see what it would do to the image, if I was still able to remove the noise with the full size picture. Can someone explain me please?

[admin]

Hi,

This rendition definitely looks the best, with the Running Man nebula starting to come through as well. You're just a little hamstrung by the amount of data you collected by the looks of it (although 30 minutes would typically definitely be enough to show both nebulae in greater detail), as well as your camera's tendency to cause 'negative blooming' (there is a dark horizontal band where the trapezium is overexposing). The latter is very hard to process out...

For more info on using the Bin module, have a look here.

You seem to have done reasonably well with processing the data, it just appears your camera and/or lens is the limiting factor here...

[Effix]

Hi all!

Over a year later I am back, necroing my own thread. I hope it's ok.
Since the last time I was here I'e got another camera, a Canon EOS 450D. I think it's about as old as my previous Olympus, yet I gave it a try on M31 last night. 17 pictures of 3 minutes each with a 250mm lens opened at 5.6, 400 ISO. 10 darks, 20 flats, dark flats and offsets. I also used a Hoya Red Intesifier filter as I live in a region with a good amount of light pollution. I know that normally the data should be taken without filters to ensure they are white balanced, but I thought I'd give it a try anyway.
Stacking with DSS as usual, and then I followed more or less the tutorial for the processing. Here is the result:

Original data:
https://www.dropbox.com/s/9djx8v9fvtrcfxn/M31.FTS?dl=0

After Star Tools:
https://www.dropbox.com/s/01ix5jben21rh ... .tiff?dl=0

After a last processing in Photoshop:
https://www.dropbox.com/s/8qhifzdknxwe1 ... 0.png?dl=0

I have to say that first this camera generates much less noise than my Olympus, which makes processing easier, as StarTools won't mistake noise for stars. Also increasing the exposure time definitely brought up the extensions of the galaxy, though they are maybe still a bit faint in my opinion. A massive improvement nevertheless. Yet I am still not totally happy with the stars, they look blury. And M33 seems to have suffered a lot from the process.
I think next time I'll try M45 and go to 5 minutes exposures.

Anyway, thoughts?

[Effix]

So I gave it another try today on the exact same starting picture, and I've got a totally different result:
https://www.dropbox.com/s/l4n7qjowcv68i ... 2.jpg?dl=0
I can't decide which one is the best. The first one is more subtle, but on the second one the galaxy's limits are clearer, and we can see further on the left and right but the colors are too agressive. I think it needs a bit longer exposure to have more signal in the arms.
In either case it looks better when it's small rather than full size

[admin]

Hi,

Congrats on the camera upgrade! The 450D is a fine camera.

More data is always better, but detail is definitely coming through.

This is how I processed your data;

--- Auto Develop
To see what we got. We can see a bias, noise, slightly soft, deformed stars.
--- Crop
Framing M3 and companion better.
Parameter [X1] set to [893 pixels]
Parameter [Y1] set to [1126 pixels]
Parameter [X2] set to [2847 pixels (-1443)]
Parameter [Y2] set to [2391 pixels (-465)]
--- Bin
Reduce noise.
Parameter [Scale] set to [(scale/noise reduction 50.00%)/(400.00%)/(+2.00 bits)]
--- Wipe
Parameter [Dark Anomaly Filter] set to [5 pixels] - always a good idea when data is noisy.
--- Auto Develop
RoI over part of M31.
Parameter [Ignore Fine Detail <] set to [3.1 pixels] to make AutoDev ignore fine noise
--- Life
Isolate preset (full mask set), to push back noise and isolate the galaxies from the background.
--- Color
Final color calbration. Default color balance was too green. How do I know? I used this technique.
We also know that spiral galaxy cores tend to have older stars (and are thus yellower), while the outer regions have younger stars (and are thus bluer).
Parameter [Dark Saturation] set to [5.20] to increase color in the darker parts.
Parameter [Saturation Amount] set to [165 %]
Parameter [Blue Bias Reduce] set to [1.17]
Parameter [Green Bias Reduce] set to [1.12]
Parameter [Red Bias Reduce] set to [1.13]
--- Wavelet De-Noise
Default settings.
Parameter [Grain Size] set to [8.0 pixels]

M31(1).jpg (110.34 KiB) Viewed 9922 times

Hope this helps!