Whether for pleasure or for measuring the length of the sidereal day, photographing star trails is a beautiful art. There's a lot of detailed information out there about photography tips specifically pertaining to the night sky and how to make accurate and aesthetically pleasing astronomical observations. I thought it might be helpful to fellow enthusiastic stargazers who would like to capture star trails if I give here a summary of the essentials that one must consider to make one.
Contents:
What to Consider:
To record star trails, the minimum requirements are an SLR camera (ideally digital), an intervalometer and a good, sturdy tripod. The key components to achieving a good final image though are as follows:
The “Bulb” setting of a digital SLR camera can be used to take long exposures of 1 hour or more. Alternatively, a series of 5-minute exposures (12 per hour) can be taken and then digitally superimposed or stacked to produce a composite image of star trails. This latter method is what I used in order to reduce digital noise in each exposure.
The F-stop (regulating how much light is allowed through the aperture) and the ISO setting (regulating light sensitivity) parameters must be optimised to capture images of bright as well as weak stars while keeping the background relatively clear or free of digital noise.
It is also necessary to ensure that the camera is stable and vibration-free so that long exposures stay in focus. A heavy and sturdy tripod should be used to mount the camera. If windy, the tripod must be placed in a position shielded from the wind to ensure sharp focus.
A remote shutter release must be used to ensure vibration-free imaging and a programmable intervalometer should be used to ensure consistent opening and closing of the camera shutter when a series of images are to be captured.
Additionally, the position of the Moon and the moon-phase must also be considered since light reflected by the Moon can increase ambient light and stars with high apparent magnitude (lower brightness) could be harder to see. One should aim to capture star trails, therefore, either when the Moon is at a low altitude / below the horizon or when there is a New/Crescent Moon. However, star trails can also be recorded when the hour angle of the Moon (the amount of time after the Moon is highest in the sky) is significant (e.g. 3 – 5 hours, to ensure a low starting altitude of the Moon, which would limit the light contamination from the Moon) and positive (to ensure that the Moon will not rise any further during the experiment). Since the Moon was forecasted to be below the horizon and in a Waning Crescent phase for the duration of the star trail experiment presented here, there was no need to know the hour angle in foresight, since even on culminating, the Moon was below the horizon. Another source of ambient light is light pollution in urban settings, hence a rural setting, away from motorways and flight paths, is ideal.
It is vital to consult the weather forecast to ensure that the star trail recording is undertaken when the night sky is cloud-free and clear (and the atmospheric turbulence is somewhere like 1-3 on the Antoniadi Scale).
My Procedures:
To optimise aperture and light sensitivity on the camera Bulb setting, 5-min exposures were taken at a few different aperture and ISO setting combinations as a pilot experiment. The settings for the best image were noted and used in further experiments.
There were very few nights which satisfied the above considerations in rural north Cambridgeshire. Two experiments were performed to try to capture a one-hour star trail. These exposures produced images that were not suitable for use since there was a high amount of background noise. Instead of long exposures, a series of 5-min exposures were taken over three hours in further experiments and results from these latter recordings are presented here.
The Pole Star (Polaris) was identified using the “pointer stars” from “The Plough” asterism:
To capture longer celestial arcs within the image frame, the position of Polaris was deliberately chosen to be about 4/5 of the height of the image frame from the top of the image. Another benefit of this was that there were more arcs above the Pole Star. The light from a star A is less refracted by the atmosphere than a star B when star A’s altitude is higher than star B. This results in more circular star trails for A than for B. Given that there were more stars above than below the Pole Star, more trails were more circular.
My Apparatus and Settings:
Camera: Nikon D5600
Lens: AF-P NIKKOR 70-300mm DX VR
Camera Settings:
Zoom: 70mm Shutter speed: Bulb mode
F-Stop: 4.5 ISO: 250
Remote Timer: PIXEL TW-283 DC2
Timer Settings: i. Delay between repeats: 1 sec (the minimum possible value for the delay).
ii. Shutter Open iii. Timer to 5:00 min iv. Shutter Close v. Repeat i to iv: 11 more times (for 1 hour); 23 more times (for 2 hours); 35 more times (for 3 hours).
Tripod: K&F Concept 72”
Image Preparation:
Depending on the length of the exposure, either 12, 24 or 36 5-minute exposures were stacked using the following two methods.
Lynkeos Software: Images in sequence were added, aligned, analysed and stacked. The resultant image was made bright and sharpened within the Lynkeos platform and saved as a TIFF file. The alignment protocol of this software created an artefact. However, Polaris was made invariant, hence its position in the stacked image does not change, giving the erroneous impression that Polaris is at the North Celestial Pole. Given that the normal procedure for measuring the sidereal day is dependent on the angle subtended by the celestial arcs on Polaris, these images could still be used for this purpose. Representative images processed in this manner are shown in Figure 1.
ImageJ and PowerPoint: Images were imported into ImageJ software, converted to an 8-bit monochromatic image each and the threshold of this image was adjusted (0 – 185: Black; 185 – 255: White), with these parameters kept constant for all images, to capture faint stars but also reduce star trail interference. (Startrailers must feel free to experiment with the threshold values since these given above were optimal only for my specific environment that night.) The images were then inserted into PowerPoint. The black background was then made transparent for every image except the first one which was placed at the bottom of the stack. All the images were aligned by length and height. In this process it was seen that Polaris does produce a small, but noticeable arc attesting to the accuracy of the standard process for measuring the sidereal day. Representative images processed in this manner are shown in Figure 2.
<<<Note: these softwares Lynkeos and ImageJ are FREE!>>>
Figure 1: Star trails generated by stacking of 5 min images totalling 1 (A), 2 (B) and 3 (C) hours of imaging in Lynkeos. The position of Polaris, shown by an arrow, appears to be at the celestial axis. The inset image in the top right corner of A is a magnified view of the area of the image included in a small white rectangle. The arrowheads in the inset show where sequential star trails overlapped in this stacking protocol.
Figure 2: Star trails generated by stacking of 5 min images totalling 1 (A), 2 (B) and 3 (C) hours of imaging in ImageJ/PowerPoint. The position of Polaris, shown by an arrow, appears to be following the circle in red, centred around a virtual North Celestial Pole (vNCP).
Evaluation:
Polaris is one of the brightest stars very close to the North Celestial Pole, easily identifiable by a virtual line extended from the pointer stars of Ursa Major or “The Plough” asterism. While Polaris appears to occupy an invariant position, star trail images reveal a bright celestial arc with a very small radius.
In order to circumvent the problem of ambient light pollution occluding stars in a long exposure, a series of 5-min exposures were taken in this study with the aim of “stacking” these constituent images into one composite image. To achieve imaging in this manner, a remote shutter release intervalometer was used which was programmed to take twelve consecutive 5-minute exposures. The intervalometer, by design, requires a delay between images and so a 1 sec delay, the minimum achievable, was set to image the star-trails with minimum interruption.
As evident from the images of the star-trails obtained, stacks produced by the Lynkeos software (Figure 1) appear more realistic and are quicker to produce but there are errors such as discrete overlapping of star trail segments (Figure 1A) arising out of the software’s forced alignment at Polaris. The images produced by ImageJ (Figure 2), however, show the star trails in greater clarity and have a higher degree of accuracy.
Results:
Here are two star trail GIFs I made. GIF1 "Star Timelapse" is composed of the original long exposures from the Nikon D5600, and GIF2 "Star Trails" is created from the exposures processed by ImageJ.
GIF1 "Star Timelapse".
GIF2 "Star Trails".
Measuring the Sidereal Day.
As a result of the Earth’s rotation around its axis, stars and constellations also appear to rotate around the Pole Star. Stars closer to the celestial poles rotate minimally whereas those farther away rotate through a greater distance. The angle subtended by the stellar arcs at the Pole Star, for a given time interval, would, however, be invariant. The time taken for one complete rotation, 360 degrees, by any star, is the length of the sidereal day (the time it takes for Earth to complete a rotation on its axis). It follows, therefore, that if stellar arcs for a given time can be observed and recorded and the angle subtended by the stellar arcs at the Pole Star are measured, the length of the sidereal day can be calculated as follows:
(SD --> sidereal day; Δt --> exposure time; θ --> average angle subtended by stellar arcs)
While Polaris appears to occupy an invariant position, star trail images reveal a bright celestial arc with a very small radius. Measurement of the angles subtended by celestial arcs to Polaris is therefore likely to generate an erroneous estimate.
Lynkeos artefactually places Polaris in the North Celestial Pole. As a result, the measurements will have to be of angles subtended by the celestial arcs at Polaris. The estimates of the sidereal day from these measurements are therefore found to be erroneous, as expected. The percent error in this measurement ranges from 0.15 – 9.06%.
The ImageJ / PowerPoint method creates stacks of images aligned by length and height and therefore Polaris appears as a bright arc with a very small radius.
Below are calculated sidereal day lengths resulting from each image processing method (for each of the exposure times):
For each exposure time, I found a few arcs and measured the angles that the arcs subtended. I chose arcs distant from the centre so that the angles could be measured more accurately. You can see why measuring the angles from Polaris (as in Lynkeos) is very bad practice. It led to an estimate for the sidereal day of 26 hours (for an hour-long exposure)! Not only is that wildly inaccurate, it is astronomically impossible! Given that the Earth rotates on its axis and orbits the Sun in the same direction, the sidereal day must be shorter than the solar day. Click here to see why!
On the other hand, Image J allowed me to estimate the sidereal period with only 21 seconds error (for 1 hour).
Well, that's all folks! Happy Startrailing! Please share your comments and thoughts (and any star trail pictures if you like) and c u soon!