Star Trail Creation – Step By Step

I dredged this one up out of the archives. Many people ask me “how do you do those star trails, Steven.”  If you want a grand overview of the process, my Treatise on Star Trails is a good read. However here I reveal step-by-step how I create a star trail image from the first shots to the finished image. So here we go.

one hour payday loans

South Side [C_009842-75br]

Photo 1: Quickly stacked image which is a composite of thirty-three 6-minute 500 ISO exposures and one 30 second, ISO 2000 exposure.

The above is my first quick attempt at creating a star trail, and following is summary of how I created it.

It starts with the test image and continues with the exposure set. For background on how to navigate the various shooting choices, see: the summary Stacker’s Checklist. The theory about selecting exposures may help.  And there is a two-part series that addresses the difficulties you may encounter. See Part 1 and Part 2. If you’re curious how I get the stars to form circles, this article will provide the information.

I usually start with a short (30 second or less), high ISO exposure to gauge several things: 1. How well framed my subject is, 2. How sharp the focus is, and 3. What I may need to adjust to control the sky-glow.

Photo 2: First, image: ISO 2000, f/2.8, 30 seconds.

After taking the first image, I realized two things: one is that the trucks passing were providing helpful light on my foreground – but not always illuminating the entire thing.  The other is that the exposure (ISO 2000, f/2.8, 30 seconds) was under exposed.  I needed to at least double the exposure to 1 minute. Let me stop for a moment. Those of you who don’t do much night photography are thinking “Whoa from 30 seconds to 1 minute is a huge difference.”  But no, it’s not. It’s only one f-stop. It is no different from changing a daylight exposure from 1/200 of a second to 1/100 of a second.

Starting with an exposure at 1 minute, 2000 ISO f/2.8 as a starting point I calculated an  ISO 500, 6 minute exposures at f/3.5. Here is how: A 1 minute exposure at ISO 2000 is equivalent to a 4 minute exposure at ISO 500 (500 is 1/4 of 2000).  Changing the aperture from f/2.8 to f/3.5 drops the light by about 33%, so I increased the exposure from 4 to 6 minutes.

I set my camera to record in RAW and my interval timer to take 5 minute, 59 second exposures every 6 minutes. I pulled out my reclining beach chair, a sleeping bag and slept while the camera clicked.  Below are a few of the shots. Note how the light changes from passing trucks!  You can also see the counter-clockwise rotation of the Milky Way. The last shot was taken as twilight approached is too bright to use because the sky is losing contrast and the light on the cliff is looking flat. I did not include that last shot in the stack.

Photo 3: Collage of some of the photos used in the stack.

I downloaded all the images from my card to my Incoming folder which is organized by date.  I used Digital Photo Professional to pull up the images, applied a bit of contrast enhancement, a slight exposure increase (1/3 of an f-stop), and a very slight noise control over the entire image. I exported in Landscape style which adds a slight saturation increase (Photoshop Saturation and Vividness) and modest sharpening. I cloned the recipe to all the photos and exported them into a “RedRockEast” folder in a temporary directory.  I could have done all these things with ACR (Adobe Camera Raw) or Lightroom.  In this case I didn’t have to do any white balance adjustments because I had preset the camera to approximately 4100K.

I then dragged and dropped all the exported (JPG) images onto Image Stacker which took about 3 seconds per image – less than two minutes to create a result. My option for Image Stacker was “Brighten” mode. I could have used the Star Circle Academy Stacking Action in Photoshop instead and the result would have been identical.  The stacking action takes about the same amount of time. (Note: StarStax is a newer program that supports Mac, Unix and Windows and works well, too!)

Photo 4: First Results stacking 34 images in Image Stacker by Tawbaware

The result was a little dark and flat so I used Picasa 3 to increase the exposure (called Fill Light), highlights and shadows – each by about 1/4 of the scale, and I warmed the photo by slightly tweaking the white balance (Color Temperature). That was all I needed to get the image shown in Photo 1.

Screen Shot 1: Picasa Adjustments

One obvious problem with the result is that the combination of the early short exposure with the sequence of shots left a gap. There really was no reason to include the first shot.

I want the cliff to pop a bit better, so my next course of action was to work on improving the foreground.  I found all the brightest shots of the cliff face (e.g. when the trucks were lighting them), and combined them using additive stacking to brighten them and averaging to reduce the noise. Remember that “brighten mode” (Lighten in Photoshop) does not brighten anything – what it actually does is select the brightest pixels at each location from each of the images in the stack.  The brightest pixels may also be noise! Using averaging reduces the noise significantly – but it will not remove “hot” pixels; we will address those later.  Fortunately Image Stacker has an option to stack and average. All you need do is specify the divisor.  If you have 10 images and specify a divisor of 10 then you are simply averaging. But if you specify a divisor of, say 5, then you are averaging AND effectively increasing the brightness by about 1 stop.  I used 12 images and a divisor of 3. And I made the same adjustment to the result in Picasa as I showed in Screen Shot 1. But I wasn’t happy with the result – the foreground still wasn’t bright enough.

Next I took 10 of the brightest images and Stacked them (additive).  After tweaking shadows and brightness in Picasa I got this:

Photo 4: Additive stack of 10 images.

Now my foreground is better, but I have created a new problem. The sky is over-bright and the hot pixels and the noise are significant as shown in a 100% crop below.

Screen Shot 2: 100% view showing Hot Pixels and noise (white speckles)

The hot pixels here have a purple fringe to them. Sometimes hot pixels are tinged red, green, blue, white or gray. I will fix hot pixels in my next to last step using the clone stamp (Picasa’s retouch) or the healing brush in Photoshop.

While the noise is obvious at 100% I think it will be fine so I am not going to address it.  If I later find the noise intolerable I will go back and stack more images and average them. Or I may return to the original images and apply stronger noise reduction in Digital Photo Professional and re-export them.

My next task is to remove the over-bright sky from the Photo 4, above. Sky removal is rather easy with the wand selection tool in Photoshop. I select all the sky and fill with black after making a few more tweaks to contrast and color.

Photo 5: Sky removed and replaced with black.

Since I now have a black sky version with the foreground as I like it, I can include this frame in any other stacks I make, and my foreground will be just as I want it.

To complete the process, I restacked 33 images together with the sky-less foreground image (Photo 5). Some more minor shadow and color temperature tweaks and some spot corrections of the few hot pixels (there were about 15), an addition of my copyright and this is the result:

Photo 6: Final Image

Since I had all the images for the stack, I was challenged on Flickr to also make a time-lapse video. This video below also helps to illustrate how stacking works. I collected the original thirty-three 6 minute exposures and cropped them to HD format (1920 x 1080). I then created and a sequence of stacked images using a modified Star Circle Academy Stacking action and joined them into an animation complete with a lovely snippet of the song Kidstuff by Acoustic Alchemy. In my next column, I’ll show how to create the time-lapse animation.


Red Rock Dancing a video by Steven Christenson on Flickr.
If you would like hands on experience and instruction, you can join us at a StarCircleAcademy Workshop

Stratospheric Exercise for Moonatics

The moon setting behind the US Capitol Building, Washington, DC

If you’re going to chase the moon (or the sun), there are problems that you need to solve.  Here are some exercises to hone your sun and moon chasing skills so you can turn the chasing into catching.  The questions get progressively harder.  Those who have taken our Catching the Moon Webinar will find the answers and much more detail on the private course materials page.  The tools you will need to solve the problems include

You might also want to read some of our past articles on the topic, especially part 1 and part 2.

The Stratosphere Tower, Las Vegas, Nevada

I’ve picked a place that I hope few people are intimately familiar with.  Many of use have been to Las Vegas, Nevada and know that there is one of the worlds tallest towers there. I’ve even had the thrill of hopping on the “Big Shot” ride – the tallest ride in the world.  The Stratosphere Tower is second in the Americas in height only to the CN Tower in Toronto.  The Stratophere’s height above relatively flat surroundings makes it an easier target for catching a sun or moon set or rise from a distance far enough to make the tower seem small.  Even though the Stratosphere is tall, there are complications – including surrounding buildings and surrounding mountains. The farther you move away from the tower, the more significant those obstructions and potential obstructions become.

If you have never seen the tower, above is a relatively close up shot captured from Google Street View. Take note of the height of the mast above the “bulge” in the tower – that’s where you find the ride “Big Shot.” To my thinking it would not be terribly interesting to get an alignment with the sun or moon behind the mast of the tower. On the other hand, if the moon/sun diameter is not as large or larger than the bulge, the shot may not be all that interesting either.

On to the questions, starting from the basic data you need to collect, and on through to solving a “real life” alignment problem.

  1. What is is the correct GPS location for the Stratosphere Tower in Las Vegas, Nevada?
  2. The base of the Stratosphere Tower is at what elevation?
  3. Looking west from the base of the Stratosphere Tower, what is the azimuth and altitude of the tallest natural obstruction in the range of West, south west, to west north west (235 to 295 degrees)?
  4. From the tower at ground level: sunset on Tuesday, August 28, 2012 occurs in line with which of these natural features:
    1. La Madre Mountain
    2. Griffith Peak
    3. Lone Mountain
    4. Frenchman Mountain
    5. Mt Charleston
  5. On what day in August, 2012 will the sun appear to set on (not behind) La Madre Mountain peak?
  6. How far is the summit of La Madre Mountain from the base of the Stratosphere Tower?
  7. Can the Stratosphere Tower be seen from the intersection of Boulder Hwy (Nevada Route 582,aka Fremont Street) and East Sahara Avenue?
  8. If the tower is visible from the above intersection, which part of the intersection provides the least obstructed view?
    1. East
    2. North
    3. South
    4. West
  9. How tall is the Stratosphere Tower (excluding the antenna/mast on top)?
  10. How far is the Stratosphere Tower from the Fremont Street/East Sahara avenue intersection?
  11. What is the difference in altitude between ground level at the Stratosphere, and the ground level at the intersection?
    1. The intersection is 279 feet lower
    2. The Stratosphere is 279 feet lower
    3. No change
    4. The Stratosphere is 1,402 feet higher
  12. What is the altitude (angle above ground) from the intersection to the tip of the mast of the Stratosphere?
  13. On Wednesday, December 19, 2012 from the intersection, the moon will pass closest to the Stratosphere tower at what time:
  14. From the intersection the apparent moon size is about:
    1. Equal to the tower height, excluding the mast
    2. Half the height of the tower, excluding the mast
    3. 1/6 the height of the tower, excluding the mast
    4. Twice the height of the tower, including the mast
  15. On Wednesday, December 19, 2012 at the time calculated in question 13 the moon will:
    1. Pass just under the bulge in the tower
    2. Pass just over the bulge in the tower
    3. Pass behind the bulge of the tower
    4. Pass through the mast of the tower
  16. As seen from the intersection: what is the first day after June 13, 2012 when a nearly full moon (at least 95% illuminated) will appear to set behind the Stratosphere Tower?
  17. What is the NEXT day after the date found in the previous calculation that a nearly full moon will appear to set behind the Stratosphere Tower? (Hint: it’s more than a year later than the previous event).
  18. You want to catch Venus crossing the face of the sun as the sun sets behind the Stratosphere tower on June 5, 2012. In what publicly accessible location would you stand, and at what time so that:
    1. The sun is as large as possible relative to the tower (i.e. you’re standing as far away as practical).
    2. You are confident there is a visible line of sight to the tower.
    3. There are as few obstructions as possible in your line of sight.
    4. There is no mountain, hill or other building behind the tower along the sightline.
    5. You have at least a little bit of room to move to correct for misalignments in your calculations (e.g. standing on a manhole cover in the middle of the freeway is not advisable!)

Good luck!

PS If you’re stumped, I recommend our Catching the Moon (and Sun) Webinar.

NOTE: You are free to ask or answer any of the questions in comments, but those comments will remain private so that those who come along later won’t be tempted to cheat!

Bending Reality

Many of those who follow my work and my webinars already know that I’m passionate about all things astronomical and night photography.  I’m the kind of guy that will go 10 times to the same location over a 4 year span to capture a shot that requires the elements to all be in place… the moon, a soon-to-rise sun, a coastal lighthouse, and clear weather.

A Perfect 10 [5_057646]

Why do I do it? Because it involves embracing challenges from several disciplines – mathematics, astronomy, and some hairy technical aspects of photography. I don’t know what to make, however, of the super cheesy way to get the moon where you want it using photo editing techniques.  Take a nice clear picture of the moon with a 200 mm lens, and put it in a landscape taken at 20mm.  You get a moon 10 times larger than it should be, but will people notice?  Not very often, it turns out! I submitted an “e-ticket” for a charity. The E-ticket allowed the bidder to select any of the items in my image library for an 18 x 12 print. To my chagrin the image chosen was this one – a complete fabrication.

My What A Lovely Moon You Have There! [C_039349+]

The patron knew it was a fabrication, but loves Yosemite and likes having the moon, a moon bow and a waterfall co-mingled together.  Making compelling composite images certainly falls in the realm of art which I highly value. But sadly, I created this image to educate people about forgeries.

One distressing trend that I see are photographers who run workshops and draw in participants by exhibiting photographs that are composites – not reality.  I know how I would feel if I booked a particular hotel because a lovely photo of the property showed a pristine beach just yards away but upon arrival discovered that the pristine beach was actually four blocks and a freeway away. Angry.

And what about those aspiring photographers who wonder how they too can get a photograph of a fantastic huge moon behind Yosemite Falls when you can’t – it’s impossible!

I have spent substantial effort finding, researching, learning about and writing tools to aid in achieving alignments of the moon (and sun) with various landmarks – but it’s quite unfair to compare the days weeks and months of calculation and waiting for the right date and weather to arrive to those photographers who recycle their stock moon, lightning, or cloud photos into whichever photos they think they will look best.

And it’s not a moon-only phenomenon.  I’m amused when I see folks exhibiting impossible star trails, improbable eclipses, and many other manufactured phenomenon.  So I am seeking your help. I want to learn from you what you think the boundary between fantasy creation and photography should be. And I’ll also give you some tools to help spot forgeries.

Here is what I think must be true for a photo to be an Authentic Photograph. Some editors are even more stringent about what they allow.

  1. Photo(s) used in the final image must be taken at the same focal length and using the same equipment.
  2. And within the same 24 hour period.
  3. With no introduction of or removal of elements except those that are “small distractions”.  E.g. noise, an overly bright or overly dark element such as shiny trash or a tripod shadow,  Or cloning out an object that moved between exposures as in HDR.

I find these acceptable:

  • Cropping – any amount.
  • Sharpening, or Blurring (smoothing)
  • Color correction, saturation or desaturation (but not color change. Green eyes should not become blue ones – though “red eye” correction is certainly ok).
  • Selective coloration, including black and white, duo toning, etc.
  • Perspective or lens aberration corrections.
  • Vignetting
  • Framing
  • HDR, or bracketed exposure combinations together with tonal compensation.
  • Contrast enhancement

For me the following cross the line from Authentic Photographs into Composites

  • Use of any elements taken a different focal lengths or with different equipment  unless those elements are resized proportionately and placed in their correct and actual location.
  • Using elements taken at different dates or from different directions (e.g. combining a photo of lightning with anything that the lighting did not actually strike)
  • Moving elements in a single image to other locations (except incidentally to clone out or cover over issues).

When people violate my personal ethical boundaries, they usually create physical conundrums that are often easily spotted by a trained eye. For example in this photo, and this photo the moon is impossibly large though only the second photo clearly feels wrong.  The moon is too large in this photo, too. The impossibility can be determined either from experience or through some mathematics (which I’ll show later).  Sometimes the moon placement just doesn’t match physics. For example in this photo and this photo, the moon is illuminated on the wrong side. The moon is always illuminated by the sun so if the sun is setting at the left and the moon is illuminated on the right – well that’s impossible. Another common mistake is when people put a Full moon anywhere near a sunset or sunrise. The full (or nearly full moon) is always located on the opposite side of the sky from the sun. Any photo showing otherwise is doctored.  Sometimes the doctoring is laughably obvious as in this photo.

Math Reveals Forgeries

As I teach in the “Catching the Moon (and sun) Webinar” the moon is a well known and almost invariant size and its presence can be used to measure distances in the photo. Specifically the moon is 1/2 of a degree in angular size. An image with a angle of view of 50 degrees – as might be achieved with a 35mm lens on a 35mm camera – will result in the moon being exactly 1/100th the size of the image. Since the field of view of an image isn’t always obvious – especially in unfamiliar locations, finding something in the scene of identifiable size helps. For example in this image:

My What A Lovely Moon You Have There! [C_039349+]Some googling will tell you the height of Upper Yosemite Fall (675 feet). By eyeball Upper Yosemite Fall appears to be about 2.5 times the moon’s diameter. That means the falls are about 1.25 degrees in total (0.5 degrees times 2.5). Throwing some trigonometry in here, we can conclude that to get the moon sized as in the photo, the photographer had to be 30,934 feet away from the fall (5.8 miles). Yosemite Valley is less than a mile wide in this direction. Did we have to do the trigonometry? Not really! The rainbow gave us another huge clue. The arc of a rainbow is about 2 degrees wide from the top color (Red) to the bottom color (Violet). So the rainbow height here SHOULD be 4 times larger than the moon. The moon, however is much too large – just as we might have suspected.

What if the only identifiably sized object in a photo is a sand dollar?  No problem.  If the sand dollar and the moon are nearly the same size in the image, it’s easy to calculate how far away the photographer was from the sand dollar by multiplying the size of the sand dollar by 114.6.  Assuming that a 4″ sand dollar and an equally sized moon are in the same scene, a simple calculation reveals that the camera was 38 feet away from the sand dollar. Do many photographers get 38 feet away from the sand dollar they put in their foreground? No, they don’t!  Suppose we were wrong and the sand dollar is really only 2 inches in diameter. It would be just as unlikely for a photographer to be 19 feet away!  If you find more than one thing of an identifiable size relative to the moon you have a second point of reference.

For comparison here is an undoctored photo featuring the moon and a guy who is about 6 feet tall.  See how tiny the moon looks in this 20mm focal length photo? Did you even SPOT the moon? It’s at chin level on the left of the post.

Mountain Man - Steven [5_018682]

Realities about Moon Exposures

The moon, even the crescent moon is very bright. Any exposure showing a detailed full or half-moon and stars is immediately suspect because the 1/100 of a second exposure needed to keep detail in the moon will rule out the capture of any stars. The problem is a limitation in the dynamic range of the camera. Our eyes can see stars near a fully featured moon… but no contemporary camera can do so except when the moon is well veiled by clouds or hanging very near the thick part of the atmosphere at the horizon.   The presence of stars in the moon bow photograph, oops, I mean COMPOSITE, scream inconsistency – or at the very least some super-duper HDR processing.  Click the image and check out the observations that other people have made about what is wrong.  E.g. How can the moon create a moon bow, illuminate the face of the fall AND be behind all those things?

A Parting Puzzle

This image accurately depicts an Annular Solar Eclipse as captured by a series of images taken with the same camera pointed in the same direction and all at the same focal length. I wrote about it in my previous column.

Annular Eclipse Sequence [C_040079+5s]

But the following image I fabricated to look much like the many forged eclipse photos on Flickr and does not jive with reality. Can you spot why?  By the way I used a REAL image of an annular eclipse to create this photo but I combined it with another photo which had nothing to do with an eclipse.

Solar Corona + Keck & Subaru

If you can come up with a reasonable refutation (or two) you may win a free Catching the Moon webinar. See here for details about how to enter, also read the contest rules.

I’d Like to Hear From You!

Do you take exception to my exceptions? Do you resonate with my concerns? Did you spot something you’re pretty sure is faked, but can’t quite tell why? Please leave a comment!

Interested in more about faked photos?

See here or the Scientific American article on spotting fakes in general photos.

Sequenced Shots (How To)

How on earth did I end up with this:

Annular Eclipse Sequence [C_040079+fives]

What I started with was lots of shots that looked liked these first three images – i.e not much of anything.

As I went along I ended up combining the “specks” into the image at the lower left. I combined the sequence with a shot taken just after sunset (middle bottom) and the result is as shown in the lower right.

We will soon provide the explanation of how to create the result. First we would like to give some clues about how the shot was planned – because, as it turns out, planning is an important part of all sequences like this one!

Avoidable Technical Content

The May 20, Annular Solar eclipse was well documented. Particularly handy is Nasa’s map based application. Choose a spot on earth by clicking on the map and some useful data pops up:

See those highlighted numbers… they tell you that when the eclipse starts it will be 31.5 degrees high in the sky, and when it ends it will be 5 degrees high – about 27 degrees top to bottom.  Allowing another 5 degrees above and say 10 below we need an image that spans 42 degrees in one direction.  Looking at the Azi numbers  The eclipse begins at 270 degrees (due west) and ends at 292.2 degrees (WNW).  So to take that all in and allow a little breathing room we need about 30 degrees.   Thus we know our field of view needs to be somewhere around 42 degrees vertically and 30 degrees horizontally. Already it sounds like we would prefer portrait mode to keep the sun/moon as large as possible. Using one of the many online tools, like the Angular Field of View Calculator by Tawbaware. Canon people might prefer the “easy to click, but perhaps not so easy to understand Canon equipment specific calculator.”

On a full frame camera, the 50 mm lens comes out to 39 x 27 degrees. which would just fit the whole sequence.  I decided to use my 70mm lens – because I already had a solar filter for it. My plan was to wait until I could catch the sun in the upper left of the frame and the foreground I wanted at the bottom. When the sun arrived, I slapped on the solar filter and started automatic 30 second intervals between exposures.

Or Just Go with Luck

Perhaps my first attempt was not so well planned.

Mamma Glows, Baby Shines

I was too interested in keeping Mt Tamalpais in the picture and ALMOST didn’t get the whole moonset. I know better now! Over three years ago I described how I created the image.  The technique is an extension of my previously described Easy HDR method.

To Be Continued…

In Part 2 of this article, we will show you a few helpful little addenda to make the process easier to manage. We will reveal a Photoshop-only method to approach the problem, AND for good measure a nifty tool to make it easy as pie.

Meanwhile if you are intrigued by the moon, you might want to join us from WHEREVER you are on one of our fun, informative, and oh so reasonably priced Moonatic Webinars.  Or maybe the next Photo Manipulation webinar is just your size.

Learn from the HDR Pro

I’ve said it before, but it bears repeating.  Were it not for what I learned from my friend and mentor, Harold Davis, my Astronomy Photographer of the Year, 2010 win would not have happened.

Photon Worshippers  **Winner Astronomy Photographer of the Year, 2010 - People and Space **

It is really just not possible to capture the wide range of exposure latitudes without post processing. Moreover processing with traditional tools does not result in as pleasing a result. It took application of the Hand HDR blending technique I learned at one of Harold’s workshops (and from his book) to make the image possible.  I hope you’ve also noticed that the image above bears none of the hallmarks of “typical HDR” imagery.

Harold is teaching an HDR workshop on March 24, 2012 in Berkeley, California. For details please see “Photography with Harold Davis”  Harold has a book on the subject scheduled to publish in July, 2012. You can pre-order it now for a quite modest price.  The Hand HDR techique is also covered along with several other powerful tips in his book The Photoshop Darkroom: Creative Digital Post Processing.

A Line in the Sand [C_035604-8PSb]

Here is the same subject using automated HDR processing:

In Another Universe [5_019718-23hdr]

And another example that “screams” HDR with its “gritty” texture and surreal look.

That Sinking Feeling [C_035617-21de]

But you do NOT have to make images like that. You can if you wish. Harold will show you both!

If you’d like to see all of my images of and around the marvelous location, take a look here.