Comet Neowise
Wide-field image of Comet Neowise, hovering over the fields of north Groningen
On March 27, 2020 a new comet was discovered by the Wide-field Infrared Survey Explorer (WISE) space telescope. If was coded C/2020 F3 NEOWISE. At the time it was still an 18th magnitude object, 250 million kilometres away from Earth. But it increased significantly in brightness as the comet was approaching the sun. At its perihelion (closest distance) on July 3, the magnitude had increased to 1. It became the brightest comet in the Northern Hemisphere since Hale-Bopp in 1997. A truly spectacular object with its long tail. It was clearly visible with the naked eye throughout the month of July at dusk and dawn low over the Northern horizon. Towards the end of the observable period, the core of the comet broke into pieces, and the comet quickly lost its brightness. But what a spectacular event this has been.
Close-up image of Comet Neowise, with a clear view of the bright core and faint nebulous tail
Planning
Object
Visual Magnitude: 2.9
Apparent size : approximately 15º
Apparent motion Az: 9.6”/s
Apparent motion Alt: -3.6”/s
Conditions
Astr. night: n.a.
Astr. dawn: n.a.
Moon: 24.2%
Moon set: 17:51
Comet Neowise was a circumpolar object with highest altitudes reached during daytime. Best observation times were either in the evening hours around 22:00-23:00h, when the comet was roughly 15-20º above the horizon. Alternatively, early morning hours like 04:00-05:00h were also very good time windows to observe the comet.
On July 15, the first photos were taken from the backyard. The backyard is south of the city of Groningen, so the comet was showing against a bright glow in the sky from the city. On July 16, a short drive to a rural place north of Groningen was sufficient to have a darker and less obstructed view of the Northern Sky. Unfortunately during the second night fog appeared after the sun had set, so conditions were not optimal. But still both observation nights resulted in usable images.
Capturing
In astrophotography terms, the size of a comet like C/2020F3 Neowise is gigantic. With the length of the tail being roughly 15º, you need an extreme wide-field telescope. The FSQ-106 + ASI1600 combination has an image diagonal of approximately 2.4º, not even near enough to properly frame. Instead, images of comets are much easier taken with a regular photo camera with a medium to long telelens.
In this case the Leica SL2 full frame camera was used in conjunction with the Leica SL 90-280mm tele zoom lens. The camera was placed on a regular tripod with ball-head, so there was no tracking possible. The maximum exposure for non-tracked images is typically calculated with the ‘500-rule’. It means dividing 500 by the focal length of the camera is the maximum duration in seconds. In this scenario that would have meant 500/280 = 1.8s for the far tele-end of the zoom and 500/90 = 5.5s for the short tele-end of the lens. In reality both these numbers gave clearly visible star trailing. Clearly, for objects so low above the horizon, the 500-rule is much too liberal. For the images described here, exposure was set to 1s for 280mm focal length and 3s for 90mm focal length. This resulted in perfectly round stars.
With the shutter speed determined by the earth’s rotation, and the aperture defined by the lens (maximal), there is one parameter from the exposure triangle to be set, and that is ISO. A value of 3200 was selected here as a reasonable balance between noise in the image and overall exposure. Noise was later reduced in post processing by stacking multiple images.
Focusing is a bit of a challenge, as little stars in almost complete darkness is an impossible challenge for any autofocus system. Luckily modern mirrorless cameras have some tricks up their sleeves that are very helpful. First, the exposure simulation function can crank up the visibility of dim objects in the viewfinder or the screen quite a bit. Secondly, an almost infinite zoom in the Liveview image allows for manual focusing on one individual bright star. In combination these techniques worked quite well.
Comets move relative to the star background at a reasonably fast speed. So if multiple images are stacked with stars aligned, the comet will move. In order not to loose sharpness in the comet, time between first and last image from the stack should not be too long. The comet moves approximately 10.3”/s relative to the background. In one minute that is 10 arcmins. It was decided to not go any longer than that. So the maximum number of frames in the stack for the closeup image was 20. Eventually 19 frames ended up in that stack.
Technical details
Leica APO-Vario-Elmarit-SL 90-280 mm F2.8–4
None (stationary ball-head)
Leica SL2
Ambient (appr. 11-12ºC)
Telelens
Mount
Camera
Sensor Temp.
Exposures
10 x 3s @ ISO3200, f/4.0, 90mm
19 x 1s @ ISO3200, f/4.0, 280mm
Wide-field image
Close-up image
Processing
Image processing is following a sharply different workflow than regular astrophotos. In a first step all 670 images were reviewed in Capture One software. Based on criteria related to object (wide-field vs close-up, foreground, etc) and technical performance (exposure, trailing stars, etc.) sets of images were selected that would end up in the final image stacks. Only minimal adjustments were made. Typically exposure was increased and white-point was decreased a bit. From here, 16-bit TIFF files were created for stacking.
Both Photoshop and Affinity Photo were tested here to do the stacking but the latter was by far superior, both in terms of flexibility as well as ease of use. While image processing software typically has built-in image stacking functionality, for these astro-images that did not work. A few small white pixels on an otherwise black background just could not trigger the proper stacking algorithms. Therefore stacking was done manually.
Under File > New Stack.. there is a pop up menu that allows for selecting the frames that should be part of the stack. With ‘Automatically Align Images’ turned off, click ‘OK’, and a ‘Live Stack Group’ is created. Each individual image in that stack can be selected and moved, rotated, etc. The alignment process is an iterative process in which layer for layer will be aligned. First all the layers are turned off, with the exception of the bottom layer. Then the second layer is turned on and selected. With this layer selected, the cursor keys are used to move this layer relative to the bottom layer to align the stars. Best is to do this while heavily zoomed in. If the blend mode is set to ‘Standard Deviation’, the stars will disappear when they’re aligned (see image below). This allows for more precise alignment than when the stars remain visible. Once layer two is aligned it is turned off and layer three is turned on and selected. In a similar manner layer three is aligned to the bottom layer. This process is re-iterated until all layers are aligned to the bottom layer. Then all layers are turned on and the blend mode is switched to ‘Mean’ or ‘Median’. This is the final image, but with a lot less noise than the individual frames (see below).
The close-up image looked pretty good after the stack. Only a curves adjustment to boost the mid-tones was applied to show a bit more detail in the tail of the comet.
The wide-field image still needed some extra work. The stack was aligned based on the stars. But obviously this blurred out the foreground. To fix that, a second stack was made. This time no alignment was applied. The blend mode ‘Mean’ created a virtually noise-free sharp foreground. A mask was applied to cover the stars in this stack. Both stacks were now combined into one final image.