NGC7822 - Cederblad 214
NGC7822 - Click here for full resolution
NGC7822 (R.A.: 0h 04m 43.68s, Dec: +68º 44’ 19.9”) is a large nebula in the constellation Cepheus at a distance of 3,000 lightyears from Earth. It is also known as Cederblad 214 and Sh2-171. At the heart of the nebula, illuminating the gaseous area, is a group of young, bright stars. The brightest of these stars is 100,000 times brighter than our Sun. Parts of the nebula are some distinct formations of Elephant-trunk style cloud structures. These are molecular clouds that are typically formed near very big O- or B-type stars. Other examples of these structures can be found in the Rosette Nebula and the Eagle Nebula (Pillars of Creation).
Sky-plots with a FoV of 50º (left) and 5º (right). Click to enlarge
Conditions
Images were taken on three nights in early March 2022, from the backyard in Groningen, The Netherlands (53.18, 6.54). Moon was at mostly absent. NGC7822 is circumpolar and actually best observed in autumn, early winter time. But also in early March, visibility from this location wat pretty good with altitudes of 30-50 degrees above the northern horizon.
Visibility charts showing 22:00h altitude throughout the year (left) and throughout the session on March 04, 2022 (right).
Weather was nicely cold and crisp, with temperatures around freezing point. During late February, early March there was a fairly stable high-pressure area, making for uncommonly long periods of good observing weather. The telescope system could be left outside several times for multiple nights. Visibility was reasonably good with SQM values just touching 20 mag/arcsec2.
Capturing
The image was captured using the Takahashi FSQ-106 in combination with the ASI6200MM-Pro camera. The Field of View of this combination is quite large, ideal for a pretty large object like this.
Telescope
Mount
Camera
Filters
Guiding
Accessoires
Software
Takahashi FSQ-106, Sesto Senso 2
10Micron GM1000HPS, Berlebach Planet
ZWO ASI6200MM Pro, cooled to -15 ºC
Chroma 2” LRGB unmounted, ZWO EFW 7-position
Unguided
MacMini 2018 (MacOS 10.14.6), Pegasus Ultimate Powerbox v2
KStars/Ekos 3.5.6, INDI Library 1.9.3, Mountwizzard4 2.1.2, SkySafari 6.8.2, openweathermap.org, PixInsight 1.8.9
The image was captured using 3nm narrowband filters for H-alpha, OIII and SII. For star colours, also a few images were taken using R,G,B filters, but they were not needed in the final processing. Below are the frames listed taken in each session that made it to the final image.
Image
NGC7822 is a fairly faint target and requires a lot of exposure time. In total 22h of SHO exposures were collected that made it to the final images, one of the longest exposures run at the observatory until now. Nebulas like this are ideal targets for narrowband imaging. In processing this gives a lot of options for artificial colour palettes, some of which have been used for this image (see under Processing).
After a little crop to straighten the edges and a final crop to get the nebula in a proper composition, the final image has a resolution of 8712 x 5820 pixels, or 50.7 Megapixels. It covers a field of view of 3.53 degrees horizontally.
Annotated image showing other deep sky objects, stars brighter than mag. 9 and the orientation.
Processing
Flat frames were calibrated with Bias (100). Light frames calibrated with Dark (50) and Flat (25) frames for each channel. Calibration, registration and stacking have all been done using the WeightedBatchPreprocessing script. This was the first processing using the new version of PixInsight, version 1.8.9, which has seen some significant upgrades in the WBPP script, including updated weighting algorithms, improved local normalisation and additional rejection algorithms. A new option of presets was set to ‘Maximum quality with no compromises’.
Flats were taken after the sessions, and also after an upgrade of KStars/Ekos to version 3.5.7. Unfortunately this new version does not always reliably write the FILTER keyword used in the FITS header. And without that keyword, WBPP does not know how to match the various frames. After some searching a workaround was found. The BatchFITSKeywordEdit script allows for bulk adding and changing keywords in PI. A very valuable tool.
In all three channels background patterns were removed using DynamicBackgroundExtraction. In nebulous images like this, the standard grid of sample points often introduces more problems than it solves. Sample points (only few) were manually placed in areas without nebulosity. The process was then applied using the lowest continuity order of 2, to keep the mask as simple as possible. After straightening the edges of the images, noise reduction with the MMT-method was applied to only the OIII and SII channel. The Ha channel had only very little noise, and the goal was to use Ha as a luminance, so maximum sharpness was to be retained. Median OIII and SII signal were adjusted to the Ha channel before they were stretched using standard HistogramTransformation.
The Ha channel was treated with a deconvolution to enhance sharpness. The EX Decon script was used this time. A while ago when this script was first tested, it did not give very pleasing results, so it was abandoned in favour of regular manual deconvolution. This time however, the EZ Decon script worked very well and turned out to be a very easy method to deconvolve the Ha image.
After the Ha was stretched, channels could be combined. A typical channel-to-colour assignment is known as the Hubble palette (SHO). So far when applying the Hubble palette, images were mapped 1:1 to each of the R,G,B channels. But many people experiment with weighted and mixed allocation of the various channels to get slightly different colour palettes, with very nice results. Since these images are false colour by definition, there is no right or wrong, and finding a nice colour mix is more art than science.
One of such alternatives is referred to as the ForaxX palette and is introduced by ColdestNights website. The key of this palette is that channels do not have a fixed balance between input images, but a dynamic one. One of the input images functions as a multiplying factor for the input images, making the relative contributions of one image dependent on another. The palette used to process NGC7822 is as follows:
R = (O^~O)*S + ~(O^~O)*H G = ((O*H)^~(O*H))*H + ~((O*H)^~(O*H))*O B = O
Here you can see that the R channel consists of a mix of SII and Ha, with the OIII signal influencing how much SII and Ha contributes to each pixel. The amount of variations on this theme are endless of course and it is difficult to predict which change would have which effect. So for the processing of NGC7822, the ForaxX example was applied unaltered. The resulting image was very dominant in two colours, a rusty-brown-red colour in the SII regions and a teal-blue colour in the OIII and Ha regions. In addition to the ForaxX palette, also an image with a traditional SHO palette was created. The final image was created by making a blend between the two images, with 60% SHO and 40% ForaxX. The SHO image had nice star colours, the ForaxX image just had white stars. In the final blend star colours were nicely maintained.
The resulting image has a warmer tone palette than the traditional SHO palette, but still retains the characteristic look of the Hubble palette. Also the blues were more natural blue and required less stretching to really stand out.
The next step was to enhance the sharpness and brightness of the image by bringing the Ha in as Luminance. This was done using LRGBCombination, this time with a slight enhancement of saturation (0.4; anything lower than 0.5 increases saturation) and with chrominance noise reduction turned on. Not only was the brightness of the image markedly enhanced, also the stars tightened up very nicely, eliminating the need for any star reduction later in the process. The resulting image had again a slight green-cast around stars. SCNR was applied to the inverted image to eliminate this.
In typical SHO narrowband images, there is a lot of pulling and pushing of individual colours needed, with masks applying those effects to only selected areas. In the processing description of the Rosette Nebula, this is well documented. In this case, that stretching of colours was kept to a minimum. Using the ColorMask script two masks were created. With a Blue mask additional brightness was applied to the blue areas. With a Yellow masks, an S-shaped contrast curve was applied both to the luminance in general and the Red channel in particular. This darkened the background a bit while enhancing the yellowish-bronze colours.
As a next step, mild noise reduction was applied by using the ACDNR process. This was a relatively new tool to use, and was applied out of curiosity what it could do. A luminance mask was created in a similar way as is done for the MMT method. Under inverted protection of this mask, ACDNR was applied for both Lightness and Chrominance, each with the Lightness mask selected and a StDev of 2.0. It was easy to apply and the resulting effect was very pleasing, so this will likely remain a tool to be applied in future images as well.
Finally some tweaks were made to give the image a little bit more ‘pop’. First the LocalHistogramEqualization tool. This can bring out some additional contrast in smaller structures. It is a tool that is easily overdone. After some experimentation, it was found that the maximum Kernel Radius in combination with a small Contrast Limit of 1.2 gave a nice overall contrast enhancement without any signs of overdoing it.
The last step was the script DarkStructureEnhance. Also this is a tool to be careful with, but for this image, the default settings gave a nice little extra boost to some of the darker areas in the image. To finish off, a little finishing touch to darken background was applied, and a crop was applied to get a more pleasing composition.
A set of RGB images (18x60s for each channel) had also been taken. But when combined, the star shapes were not as nice and tight as from the SHO image. And because star colour had been largely retained in the SHO image, it was decided not to swap out the stars with an RGB version.
Processing workflow (click to enlarge)
This image has been published on Astrobin.
