Messier 8 - The Lagoon Nebula In the Hubble Palette
Date: August 5, 2021
Cosgrove’s Cosmos Catalog ➤#0079
Messier 8 - The Lagoon Nebula - taken with a mono camera and rendered using the Hubble Palette (click image for full resolution via Astrobin.com)
Table of Contents Show (Click on lines to navigate)
About the Target
Messier 8, also known as the Lagoon Nebula and NGC 6523, is a huge cloud of molecular gas and dark dust located about 4000-6000 light-years away in the constellation of Sagittarius. This is a large emission nebula that is bright enough to be seen with the naked eye. Within the cloud is the open cluster NGC 6530, and a bright central feature is known as the Hour Glass.
The Annotated Image
Annotated image of Messier 8 - done with Pixinsight ImageSolve and AnnotateImage Scripts.
The Location in the Sky
IAU/Sky & Telescope Constellation Map for Sagittarius - Messier 8 is indicated with the yellow arrow.
About the Project
The data for this image was captured on the nights of July 4th and July 14th, 2021. I had not processed this image until now because I thought it was a lost cause. why lost, you ask? Well, two things - I discovered part-way through collecting my subs that I had the camera setup for 3x3 binning instead of the normal 1x1 binning; Second, clouds had moved in on the last night and shut me down after only getting 4 5-minute subs with the Sulfur II filter.
Why is this bad? Well, normally, I shoot at 1x1 binning, which gives me the camera's full resolution. A bin of 3x3 takes a group of 9 pixels and maps them into one BIG pixel. So while this does effectively gather photons more efficiently, it reduces my camera resolution to 1/3 what I normally shoot. This was an accidental bone-head move on my part. And this meant that a lot of Ha frames would be useless to me.
When the clouds limited me to only gathering 4 subs, I became data-starved for the Sulfur II filter. This would mean I would have a relatively weak signal and a lot of noise compared to the other two.
Also, keep in mind that during July, we had a real problem with smoke and thin cloud issues.
But I decided to process the images and see what I was possible. I started with the following:
11 x 300-second Ha subs
15 x 300-second OIII subs
4 x 300-second SII subs.
Blinking the Light Frames
The first thing I do is blink all of my light images to see what I will have to deal with.
Here is a video of the Ha filter subs being blinked:
A few things to note in this video. Notice how the images seem to jump around? That is caused by “Dithering.” Basically, before each exposure, the mount shifts the framing slightly in a random direction. This causes the image to shift relative to the sensor pixel layout. This actually enhances the resolution over many frames as features get laid down on the pixel pattern such that the differential placement causes different edge sampling. It also prevents certain nose and interference patterns from forming.
Also, note that one frame is upside down! What happened there?
I am using a modified German Mount. As these track across the sky and hit the zenith point, they can't go much further without having the scope, or the camera hit the mount. To avoid this problem, the mount is set up to do a “Meridian Flip.” When the sequence starts, the scope and camera are on one side of the mount, and the counter-weight is on the other side as the target nears the meridian (the halfway point of traveling across the sky) - the scope/camera and the counter-weight then swap positions on the mount. This causes the image to flip 180 degrees. This does not cause an imaging problem because the star alignment during image-stacking detects and compensates for the rotation.
You may also see that two of my Ha frames have very poor focus! This is not usual - the autofocus methods that I use are normally very accurate - but something failed here, and two images were taken with an unacceptable focus. I needed to remove these from the mix, so I went down from 11 subs to 9!
Then I did the blink operation with the OIII filter :
Here you can see the dithering as well, but no frames were flipped on this set. But you can see that some thin clouds came through on some frames, causing some of them to be dramatically attenuated. So I had to more frames. I dropped out of the mix - going from 15 subs down to 13.
Then I blinked the SII subs. Unfortunately, I only had four - I hoped not to lose any of those. Luckily I did not, and I was able to see that all four images here were good!
Pixinsight Image Processing
Now the images for each filter are calibrated, aligned, and integrated into linear masters. While the image was still linear, I did a slight crop, some careful deconvolution to restore lost detail to the images, and finally, I did some light noise reduction.
Most recently, I have been using Mure Denoising during this phase of processing. Mure measures and models the noise characteristics of your sensor by examining two dark calibration frames and two flat calibration frames. It then uses some key metrics of this measured noise profile to minimize the noise in the linear images without negatively impacting the signal. I applied this to all three master images, which was all I needed for the Ha and OIII image. Because there were only 4, the SII image was improved but still had some residual noise issues, So I hit it again with an MLT run to take a bit more off.
You can see the great job that Mure does in the comparison images below.
A section of the OIII linear image showing the noise pattern.
After applying Mure Denoise - a really nice improvement (ignore the line - its from a preview box I forgot to remove before I did the scren grab)
Finally, it was time to delinearize the images and combine them to form a Hubble Palette color image.
Image Combination and Color Treatment
I use the MaskedStretch process to apply a non-linear transform that increased the image contrast while protecting the star images from being blown out. Then I used the CurvesTransform process to apply an S-curve to shape things a bit more. Finally, I did a rough visual match of the Ha, OIII, and SII Non-linear images so that the combined image would not be too far out of balance.
Here are the three images before combination was done:
The Ha image (click to zoom)
The OIII image. (click to zoom)
The SII image (click to zoom)
When I was happy with the three images, I used the Channel combination tool to map the Sodium II image onto the red channel, the Ha image onto the Green channel, and finally, the OIII image onto the blue channel. When this was done, the image looked like this:
The Combined SHO image.
It looks really green because the most dominant image - the Ha - is mapped onto the green channel, so of course, the image comes out with a strong green balance.
I addressed this by using the SCNR tool to remove excess green in the image. Which gave me this image:
After the excess Green was removed with SCNR.
You can see we have some funky magenta stars and other magenta areas in the image. I do a straightforward operation to handle this - I invert the image, giving me a color-negative-like image.
The inverted image, where the magenta sky and stars now appears green and easily fixed.
Note that the magenta areas and stars, now inverted, appear green - so once again, I use SCNR to remove the excess green, re-invert the image and end up with this:
Now I have a nice clean image with neutral stars, and that classic Hubble SHO look.
After boosting the contrast and the color saturation a bit with the CurvesTransform tool, I ended up with a good foundation image to work from:
At this point, I did the following:
ACNR to bright down Noise some more
Histogram equalization
Masking to work on some regions of the image selectively
MLTP at the 128-pixel band to bring out larger details better
EZ- Star minimization.
And so on.
One problem I had was elongated stars. I normally don't run onto this with my mount, but I did here. I am not sure what caused it. There are ways to address this, but I also found that these "fixes" were causing other and worse problems to the image, so I decided not to mess with it anymore. There are not noticeable unless you go pixel peeping.
Given how bad those nights were - and how little data I ended up getting - I was very surprised that I got an image as good as I did out of it.
Earlier Efforts
I have shot this target twice before.
My first image was of M8, and you can see the posting of my first effort HERE.
The second effort was a traditional RGB - you can see that project HERE.
Here is a three-way comparison:
My first image was of M8 (Click to enlarge)
M8 in RGB (click to enlarge)
The new narrowband SHO effort (click to elanrge)
More Information
Wikipedia Entry: Messier 8
NASA: Hubble Telescope Images of portions of M8
The Live Sky: Messier 8
Capture Details
Light Frames
Taken on the night of July 4th & 14th, 2021
Ha Filter: 9 x 300 seconds, bin 1x1 @ -15C, gain 100.0
OIII Filter: 11 x 300 seconds, bin 1x1 @ -15C, gain 100.0
SII Flilter: 4 x 300 seconds, bin 1x1 @ -15C, gain 100.0
Total of 2.0 hours
Cal Frames
Darks: 30 x 300 seconds, bin 1x1 @ -15C, gain 100.0
Flats: 12 Ha, OIII, &SII
Flat-Darks: 30
Capture Hardware
Scope: William Optics 132mm f/7 FLT
APO Refractor
Focus Motor: Pegasus Astro Focus Cube 2
Cam Rotator: Pegasus Astro Falcon
Guide Scope: Sharpstar 61EDPHII
Guide Focus Motor: ZWO EAF
Mount: Ioptron CEM 60
Tripod: Ioptron Tri-Pier
Camera: ZWO ASI1600MM-Pro
Filter Wheel: ZWO EFW 1.2 5 x8
Filters: ZWO Gen II 1.25” LRGB,
Astronomiks 6nm 1.25” NB
Guide Camera: ZWO ASI290MM-Mini
Dew Strips: Dew-Not Heater strips for Main and Guide Scopes
Power Dist: Pegasus Astro Pocket Powerbox
USB Dist: Startech 8 slot USB 3.0 Hub
Polar Align Cam: Polemaster
Click below to visit post about the Telescope version used for this image
Software
Capture Software: PHD2 Guider, Sequence Generator Pro controller
Image Processing: Deepsky Stacker, Pixinsight, Photoshop - assisted by Coffee, extensive processing indecision and second guessing, editor regret, and much swearing….. Given the problems on this image, more than the usual whining….
I was very excited to get the ZWO ASI2600MM-Pro camera a while back. I ordered it when it was announced and then prepared to wait a long time to get it. When I did get it - I decided to put it onto the AP130 platform. That meant that I could move the ZWO ASI1600MM-Pro, along with its filter wheel over to my William Optics Platform. This now means that all of the platforms have been moved over to a mono camera and my ZWOASI924MC-Pro is now not in use.