Hi Guys,
This is my very first ever mosaic so I chose a nice bright target to practice on. This is made up of two panels and was taken over four nights. I will probably add to this image as I collect more data. I have some funky stars in there I know. I have some focuser tilt which I am currently working through. Constructive criticism greatly appreciated. I have uploaded a lerger version to my private area in astrobin, until I am totally happy with it. the link to the larger version is here.
Technical data:
Imaging telescope: Skywatcher CFP2008 Carbon Fibre 8" f4
Imaging cameras: Canon 1100D Astro modded
Mounts: Skywatcher HEQ5 Pro
Guiding telescopes: 4" f5 Newtonian
Guiding cameras: QHY5
Resolution: 6209x2583
Dates: Aug. 21, 2014
Locations: Townsville NQld Australia
Frames: 134x300" 67 for each panel.
Integration: 11.2 hours
Avg. Moon age: 25.40 days
Avg. Moon phase: 18.06%
Software:Astro Photography Tool, Pixinsight 1.8, PHD guiding
Filters: Astronomik CLS-CCD Clip
Accessories: Baader MPCC coma corrector
Edit: I have attached a second image for comparison which has had no noise reduction routine applied (see Richards Post below). Am interested to know what the common opinion is about the two of them.
Edit2: I did a masked stretch instead of histogram stretch as Richard suggested below. Lots more detail has come out, may need to adjust the settings a bit though. Let me know what you think. I placed it in this post so it is easier to compare with the others. Thanks for looking.
Edit3: Did a LHE on the image to restor some detail back into M8, as suggested by Richard below. Image attached for comparison. Thanks again for the help richard
One strategy to improve the appearance of the stars is to use a Masked Stretch in Pixinsight instead of stretching the image with HistogramTransformation. Find the darkest part of your image and select that as a preview (I would often use a preview box about 20x20 pixels - it doesn't need to be large). Then use this preview as the background sample in the MaskedStretch process.
The effect of the MaskedStretch is to stretch the nebulosity without overly stretching the stars. This reduces the star size in the final image which lessens the impact (to a degree) of funky star shapes caused by optical issues such as tilt.
You can also use a star mask (masking everything except the stars), and then use MorphologicalTransformation in 'erosion' mode to reduce the size of the stars. Dont over do it here. try increments of say 25%.
I would be tempted to probably back off your noise reduction as well. I reckon there is some extra detail in there that you're missing due to the noise reduction step.
These are just small things, as that's a good image. In fact i'd call that a graduation from "Beginners Astrophotography" into the main game
One strategy to improve the appearance of the stars is to use a Masked Stretch in Pixinsight instead of stretching the image with HistogramTransformation. Find the darkest part of your image and select that as a preview (I would often use a preview box about 20x20 pixels - it doesn't need to be large). Then use this preview as the background sample in the MaskedStretch process.
The effect of the MaskedStretch is to stretch the nebulosity without overly stretching the stars. This reduces the star size in the final image which lessens the impact (to a degree) of funky star shapes caused by optical issues such as tilt.
You can also use a star mask (masking everything except the stars), and then use MorphologicalTransformation in 'erosion' mode to reduce the size of the stars. Dont over do it here. try increments of say 25%.
I would be tempted to probably back off your noise reduction as well. I reckon there is some extra detail in there that you're missing due to the noise reduction step.
These are just small things, as that's a good image. In fact i'd call that a graduation from "Beginners Astrophotography" into the main game
Thanks Richard for the compliments and the tips, I am still on that slippery learning curve wit PI but I am absolutely loving it. As for the noise reduction, I only did 1 round of ATWT, but the image was very clean to start with so maybe I need to tamper with some of the other settings before running it. As I'm processing my images I always save each step as I go. I went back and got the image from before the noise reduction and have attached it to my first post for comparison. the only processing that has been applied to the second image is DBE, BG neutralisation, Colour Calibration, and of course a Histogram stretch. The rest is as it came out of the integration process. Thank you for the graduation certificate, lol, maybe once I have sorted out the issues with this image, I will post it in the Deep sky section. Thanks again for your help.
Quote:
Originally Posted by raymo
That is also a ripper, Ripper Rex.
raymo
lol, Thanks Raymo.
Quote:
Originally Posted by cometcatcher
That's a top shelf "bigfoot" Rex.
Thanks Kevin. I actually feel like I am improving with each image. Can't wait till I sort out the focuser issue.
Quote:
Originally Posted by LightningNZ
That's a really nice capture Rex. The data's starting to look really natural and smooth.
For noise reduction I now use the TGVDenoise process in Pixinsight, on the unstretched image. Its not as heavy-handed as some of the other noise reduction options. Experiment with the settings, using (as a starting point):
Strength: 0.5
Edge Protection: 0.0002
Smoothness: 2.0
If you still need some noise reduction after stretching, then increase the strength and edge protection by approx 5-10x
ie
Strength: 2.5 to 5.0
Edge Protection: 0.001-0.002
Smoothness: 2.0
For noise reduction I now use the TGVDenoise process in Pixinsight, on the unstretched image. Its not as heavy-handed as some of the other noise reduction options. Experiment with the settings, using (as a starting point):
Strength: 0.5
Edge Protection: 0.0002
Smoothness: 2.0
If you still need some noise reduction after stretching, then increase the strength and edge protection by approx 5-10x
ie
Strength: 2.5 to 5.0
Edge Protection: 0.001-0.002
Smoothness: 2.0
Thanks again Richard, I went back and did a masked stretch instead of histogram stretch and placedit in the first post. Brought out heaps more detail. Thanks for the heads up.
You're gonna hate me but I kinda like the first one best. In the last version the Lagoon region lacks detail, looks smokey and is washed out compared to the first.
You're gonna hate me but I kinda like the first one best. In the last version the Lagoon region lacks detail, looks smokey and is washed out compared to the first.
this is where the LocalHistogramEqualization process is your friend (try contrast Limit of 2-5, Amount of 0.1-0.2). Then Curve adjustment to restore some colour saturation.
Here's my quick edit using Rex's third version.
[edit: looking at this again on a different screen and I've probably overcooked this a little for some tastes. working with just the lowres jpeg doesn't help. still, it gives an example as to the different aspects of the image data you can focus on]
You're gonna hate me but I kinda like the first one best. In the last version the Lagoon region lacks detail, looks smokey and is washed out compared to the first.
I absolutely will not hate you Kevin. That's why I like this forum, to learn what is good and what is not. It's the only way to improve, so thanks for the feedback. I too wasn't sure about the white cloudy look it gave it but the extra detail won it for me.
Quote:
Originally Posted by rmuhlack
this is where the LocalHistogramEqualization process is your friend (try contrast Limit of 2-5, Amount of 0.1-0.2). Then Curve adjustment to restore some colour saturation.
Here's my quick edit using Rex's third version.
[edit: looking at this again on a different screen and I've probably overcooked this a little for some tastes. working with just the lowres jpeg doesn't help. still, it gives an example as to the different aspects of the image data you can focus on]
Thanks again Richard, I should have known to use LHE to tone it down a little because I've used LHE before. Late at night, tired, just a dumbass lol. that's enough reasons.
Quote:
Originally Posted by multiweb
Nice mosaic. Great framing.
Thanks Marc. I still have a lot to learn, but am slowly moving into the more challenging areas, and having a blast the whole way.
that is great Rex!!
Such a nice vista - well done and you don't see it that often.
Just to throw another spanner in the works, I too like the lagoon in the first but the fainter details in the third
that is great Rex!!
Such a nice vista - well done and you don't see it that often.
Just to throw another spanner in the works, I too like the lagoon in the first but the fainter details in the third
Rusty
Thanks Russel, and no spanner mate, all good. As I've said that's why I love posting here. I have attached another image to my first post that hopefully combines the best of the first and the third version.
This result would not have been possible without the coaching from Richard, so thank you very much for all your help Richard. I have learnt a lot about the finishing touches that are possible in a very short time.
That's lovely Rex, but I seriously dislike modded DSLRs. Everything is red, and the other colours are overwhelmed, until the image becomes almost monochrome. There are some objects that benefit, such as the horsehead, so if I could afford two DSLRs I would have one modded.
raymo
That's lovely Rex, but I seriously dislike modded DSLRs. Everything is red, and the other colours are overwhelmed, until the image becomes almost monochrome. There are some objects that benefit, such as the horsehead, so if I could afford two DSLRs I would have one modded.
raymo
Raymo, I guess what your saying is its horses for courses (excuse the bad pun).
However, 70% of the nebula in this shot is invisible to our unmodded cameras as the light in the right hand side of the nebula is all in the Ha wavelength. The modded camera is showing us that there is ionised dust and gas that is invisible to the rest of our images. This is indeed why those that
A: are experienced enough
B: can afford the gear
shoot in LRGB Ha using a mono ccd.
You are right though some objects display more pleasing to the eye with un-modified DSLR's and there are others that are much better in Modded cameras.
Last edited by BruceG; 22-08-2014 at 03:24 PM.
Reason: typo
Really nice Rex, the last rendition worked nicely. Didn't find any blue hiding in there hey? (I like a bit of blue in my nebs, a bit like Justin's image Here)
Over all great job
(I like a bit of blue in my nebs, a bit like Justin's image 
