Adventures With Urban Star Trail Photographs Using a Canon 10D Digital Camera


trails

September 18-19, 2006
9PM to 6 AM

Click here for larger image


Notice the Iridiun flare at about 4 o'clock from Polaris and two smaller Iridium flares at about 1:30 from Polaris. The ubiquitous airplane trails are unavoidable at my location.


I have always loved star trails. Some people think images of the stars circling around the pole are boring, but I find a kind of peace with the stars eternally circling the celestial pole. I see it as a sort of permanence in an always changing world. And, if you look closely, there are always surprises, like an airplane, an Iridium satellite flare, www.imitationraybanuk.com or a drifitng cloud high in the atmosphere. In the case of my star trail adventures shot from my garage roof, I also discovered what I thought was a solid, imovable building structure actually moves and dances with a life of its own. But more of that later... for now, lets see how we can create beautiful star trails from our own light-polluted urban backyard.

It was a delight to discover early in my crossover from film to digital astrophotography that when using my Canon 10D camera from a brightly illuminated deep urban location, I could actually take better star trail pictures that I could with my Nikon film cameras at a dark-sky location. One would think that urban start trails would be fewer and dimmer than dark site star trails, but the reality is both are the same. Since the stars move relative to the fixed camera used to shoot star trails, stellar exposure is limited by the time a star image remains on one point of either the film plane or digital sensor. So the stars are going to expose the same, regardless if film or digital cameras are used.

The difference between deep sky site film star trails and urban digital star trails is the length of the exposure. With film, we have to record the entire star trail with a single exposure lasting minutes or hours. In an urban location with a bright sky and bright street light-illuminated foreground, the exposure is limited to a short period before the brightness overwhels the star images. In the case of my location deep within San Antonio, Texas, if the humidity is low and sky transparency is high, 30 seconds at f/4 at ISO 400 will result in an image that is not completely washed out by sky glow. The resulting image can be processed in Photoshop to produce pleasing colorful stars with a dark sky background. But a 30-second exposure does not a star trail make... However, with digital astrophotography, we have the tools to combine a sequence of shorter exposures to synthesize a longer exposure.

Creating One Long Exposure from Many Short Exposures

Sequential short exposure star images can be manually combined using the tool in Photoshop to show one long star trail. The beauty of an image created this way is that both foreground and sky brightness never exceeds that of the brightest individual image in the stack, yet star trails grow in length proportional to the total number of short exposures. However, if you do the math, a one hour star trail taken with 30 second exposures requires combining 120 images. A dusk to dawn star trail can contain over a 1000 individual 30-second exposures. This quickly renders lengthy urban star trails into a labor intensive boring drudge.

However, astrophotographers are a clever lot and find ways around this labor bottleneck. There are two easy ways to automate the creation of digital camera star trails taken with the sequential short exposure technique. Users of Photoshop CS with find the Star Trail Photoshop Action file created by Chris Schur will allow Photoshop CS to automatically combine a whole folder full of short exposures into a single long star trail. Chris' Phtoshop Action file can be found by

clicking here

However, if you do not own Photoshop, there is another even simpler way to create star trails from sequential digital images. Achim Schaller from Freiburg, Germany, has created a freeware Windows program called startrail.exe that will read a folder full of sequential short exposure star images and combine them into a completed single image similar to the way Photoshop does. This program can be found by

clicking here

Automating Sequential Exposures for Star Trails

To create smooth star arcs in digital star trail images, the individual images in the sequence of short exposures must have a minimal interval between successive exposures. If the time between exposures is too long noticable gaps will appear. The resulting star trail will then resemble a series of dots instead of a smooth arc. Each image, therefore, must be recorded quickly and the next image in the sequence started as soon as possible. To speed the recording of sequential images, use the camera's .JPG image setting instead of .RAW. For star trail images, the .JPG setting provides adequate quality and images are recorded up to 10 times faster than with the .RAW setting.

My older Canon 10D is a USB 1.1 device, meaning downloading individual images to a computer takes a prohibitively long time for star trail photography. Recording the images on the camera's internal Compact flash memory card is much faster. However, my memory cards are inexpensive (translation.... slow!). So, to speed up recording images, I use the 2048x1360 pixel image size instead of the camera's maximum image size of 3072-pixel width images. The smaller image size retains adequate detail to resolve all star trails and has the added bonus of much faster image recording to the memeory card. Yet another plus is the fact that over 1000 images can be recorded on a 1-Gb memory card, alowing dusk to dawn star trails to be created on a single processed image.

Triggering each exposure in the star trail sequence can be done with a variety of devices; connecting the camera to a control program in a laptop computer or a programmable remote shutter switch. However, with my Canon DSLRs, and most other DSLR cameras, the simplest way to take sequential images is to set the camera to "burst mode" and keep the shutter depressed. With this setting, as long as the sutter is depressed, the camera will keep taking pictures until the memory card is full. The works with any sutter speed that is programmable within the camera, including the 30-second exposure speed.

window Setting the camera to "burst mode" and continuously holding the shutter open with a remote shutter switch allows the camera to take repeated 30-second expsoures until the memory card is full.

The star trail images shown here were taken during the new moon period in late September, 2006 using either a 35mm F/2.8 or 28mm F/2.8 lens, both stopped down to F/4. For some unknown reason the sky gods blessed San Antonio with a series of cloud-free nights alowing dusk-to-dawn astrophotography. Beleive me.... this is a rarity!


trails

September 19, 2006
10:30 PM to 6AM

Click here for larger image
Airplanes galore! You can tell I live near the San Antonio airport.
Another bright Iridium flare is seen at the bottom of the image at about 4:30 from Polaris.


traisl

September 20, 2006

Click here for larger image


And finally, a shot with not too many airplanes. I guess there are fewer flights on a Wednesday night.


The Strange Case of the Dancing Garage....

The images below were taken with the same setup as the ones above. In fact, one image, September 19th, is the same as above, exept the full duration of the exposure is shown below while the first 90 minutes or so has been eliminated from the above versions. My Canon 10D was firmly clamped to the solid structure of my garage roof using a ball-head adapter and C-clamp. (See the photos below) This same device has proved to be a stable camera platform when clamped to other solid objects like the structure of my old observatory building. However, when attached to my garage, the first portion of the star trail sequence is twisted like the camera platmorm is moving. I have not measured the exact amount of movement in the trails, but by eyeball it appears they stars have shifted by about half a degree.

My only conclusion is that the Texas summer heat warms and expandes the structure of my garage and after sundown there is a cooloff period where the structure shrinks again. The cooling and shinkage begins on the east side, away from the sunset side which remain hot longer, and thus creates an uneven warping of the entire structure for several hours after sundown. Considering how much the stars are shifted in the early portion of the star trail arcs, I find this process amazing! My garage is a typical detached two-car garage about 20 by 30 feet in size. It was built in 1951 and I have lived at this address for 31 years. The garage is apparently as solid as the day it was built. There is no evidence of nails or fasteners backing out of the wood, no cracking or spliting of the wood, and no dificulty in opening any doors that would indicate structural movement.

So the stars show what we think are solid and unmoving structures my actually creep and shift a bit after sundown. This is useful information considering I am thinking of putting a telescope mount in my garage roof to get above the local trees. Now I know the mount will have to be supported from the concrete floor, not the building structure or the scope's polar alignment will be creeing around the pole for several hours after sunset.


The camera hung from the edge of the garage roof by a ball head adapter and C-clamp The camera strap is secured to hooks under the garage roof in case the C-clamp slips

twisted

September 19, 2006
9PM to 6AM


Click here for larger image


This image is the same as the September 19 image above, but includes the first hour and a half of the exposure. Notice the difference in sky color due to business lighting to the north of me that shuts down after about 10PM. Also notice there are a LOT more airplanes just after dark.


twisted

September 21, 2006

This attempt was interrupted by clouds after only 155 minutes of exposure, but showed the warping of my garage roof continued for well over two hours after sunset.

Click here for larger image



twisted

September 25, 2006

Once again, approximately the first 90 minutes of the exposure was ruined by movement of the entire garage structure.

Click here for larger image



Back to ------------------ Table of Contents