Quadrantids Shoot Questions

Thanks all for the sharing of the post, tis much appreciated. I've received a few questions about the images and figured I'd take a moment or two to answer them to the masses even though I've responded directly already. The first question I'll address here was actually the most recently received, but for a bit of background on my process, I decided to answer it first.

What is image stacking, and how does the PhotoShop action work?

Essentially – image stacking is layering a series of images together so that you can then process them as a set. Because I shot from a tripod, and the illumination on most of the foreground objects stayed the same throughout the shoot, the sides of the houses, the chimneys, the trees, all have a consistent illumination. The sky is mostly black, except for where the stars are in any given shot. In each shot, they move just a little bit, not doing much more than adding a bit of blur to the star rather than a discernible trail in any given image. By stacking 683 of them, that lets the stars move through their arcs, the process pulling the brightest pixel from the column of images.  Now, PhotoShop doesn't have to actually stack all 683 at once, you start with a new PS document and fill it with black. As you run the automated batch process against your folder of images, PhotoShop grabs each one, and copies it into the "new" file, and then applies a lighten action to get the brightest pixel at each location. So, after processing the one image, the "processed" image looks just like the first one… from here, the changes are more subtle. The action closes that first image and works through each successive image, merging the new layer into the previously processed image and brightening fairly small areas.  Over time, this yields the streaks. 


Why 683 images? Why not a 3 hour exposure?

Okay – so that is a shortened version of the question, but the essence is the same. Why do folks stack rather than using longer exposures or a very long exposure?

I've not searched for an answer but a few quick reasons come to mind. First off, early digital cameras had more noise the longer the sensor had to gather data to process, so to avoid this extra noise, you had to keep your exposure times short. Even now, there are settings in DSLRs that enable or disable long exposure noise reduction on images; so there is still some noise generated by a longer exposure.

Another practical reason is that most cameras' longest exposure time (excluding bulb) is 30 seconds; but again this changed with the advent of intervalometers (such as my canon TC-80N3) – devices that allow you to set exposure time, time between exposures, and number of exposures, greatly simplifying time lapse photography. There are even firmware builds (such as Magic Lantern unified, available on several Canon bodies) where countless additional features can be added to your DSLR. Without one of these devices (or a firmware modification), it is easy (as I did, see yesterday's post) to set the camera on continuous shooting and just slide the latch to lock the shutter into the "shoot" position… the camera dutifully shoots till it runs out of battery or storage.

Test shots are much easier to evaluate in shorter exposures. You can always do the math and calculate how a longer shot needs to either have the aperture cranked closed, or iso lowered, or both; but why not just go ahead and do 25 or 30 second exposures?

Flexibility in post processing pops to mind as well. I is fairly easy for me to process 180 or so images:

instead of the full 683:

(Click either to enlarge)

It nets a very different result. I could also skip sections of images to add some gaps… hmmm… spelling out a secret message in morse code anyone?


So, did you actually capture any meteors?

Yes, I think… well, one… maybe. How's that for decisive? LOL  Well, based on info discussed in yesterday's post, I think I ended up with one meteor, and one satellite.  Both came from the lower portion of the image towards the upper portion… both were caught in multiple frames. The first item I saw was what I think was a meteor:


(click to enlarge)

and what I think is a satellite – tho it may be a faint meteor. It is a longer trail, but the trail is pointy at the end, so it may have been a more glancing blow to the atmosphere – but while it is faint – the trail goes all the way into the tree… so perhaps tis a satellite after all:


(click to enlarge)

There are gaps in both due to them being captured in two frames each. There is a slight pause between closing the shutter and re-opening it. I got "lucky" both times – LOL

If anyone can give me a bit more info on meteor vs. space junk, I'd appreciate it.

All for now – keep those questions coming!

Thanks for reading!

Quadrantids Meteor Shoot

Ahh… celestial photography. A fun topic that I've rarely had time to explore. I've done occasional shots of the moon, usually to test a new long lens or stacked teleconverters, but have never tried to shoot a meteor shower or star trails.  I've done some other time lapse stuff lately, and other long exposure photography, so I've had it at the front of my mind. As I was surfing around the other day, I saw a post about the Quadrantids meteor shower, and how it was that night. I did a bit of surfing about where and when it would be visible but didn't think too much of it. As luck would have it, I got stuck at the office late after a fairly long day and noticed it was crisp and clear in Austin as I headed to the car.  On the drive home, tinkering with some long exposure photography seemed a great way to relax and unwind.

When I got home I went into the back yard and realized that a) I'm to the south of downtown Austin, so the views north to northeast from the back yard were polluted by light; and 2) [sic] that with the houses of my neighbors also cluttering up the sight lines, I'd be pretty much guaranteed to get squat for meteor shots.

Undaunted, I figured I could use this as a great way to tinker with exposures, apertures, ISOs, and lens selection. I soon had my 5D2 in the back yard with my 17-40L f/4 mounted atop my tripod and ballhead. I wanted to go for a longer exposure (figuring I'd go "simple" with my camera's shutter options) but didn't know what ISO, f-stop, etc. I ran a series of tests with the lens wide open at f/4, and thought I'd be happy with an 800 iso 25 second exposure @f/4, but ran a few other tests while I was at it.  I played with different shutter speeds at 400 and 1600 ISO, as well as halving the exposure time and bumping a full stop to 5.6.  I was a bit surprised at how much I'd liked the 800 ISO f/5.6. It was definitely sharper, and with some lingering humidity in the air (along with the light pollution to the north), the reduction in overall exposure made it all look so much crisper. I had thought I'd want to be a touch overexposed to ensure I caught every possible trail, but with the reality of the houses and lights, I figured this might be a better star trails exercise, so I leaned towards a slightly darker image.

As for which "exact" direction to point the camera, I leaned on (as I say it) the AKASG – all knowing all seeing Google – In this case – the Google Sky Map app on my Android based device (tho I'm pretty sure tis free for iOS and Blackberry devices). I figured I could use it to help ensure I knew which way was NNE, but I was surprised to see that meteor shower focal points (essentially the point at which it looks like the meteors are coming from, they radiate out from the focal point) is a selectable option to include in the view. So it was very easy to see where the shower currently was (at this point, it was bout 11:00 PM and the shower's focus was still below the horizon) but knowing where Polaris (the north star) was, I could figure out where the shower's radius would "rise" and how it would track across the sky.

I'll admit I got lazy and didn't over research the ability of the 5D2 to lock the mirror up and stay up between images. I thought I could "fool" it by going into Live View (which locks the mirror up so the sensor can read real time) but wasn't sure if I'd be wasting valuable battery juice powering the back LCD panel. As I thought through the mirror lockup, I decided that since the up and down of the mirror would be such a small portion time wise of the 25 second exposure, it wasn't worth fooling with. Overall battery life was indeed a concern as I'd no idea as to the duration of a gripped 5D2 with two batteries in the cold 35 degree Austin night, so I was trying to err on the side of caution. I wasn't sure if the cold and draw of the long images would run the battery out before my 16gb CF card filled up.

As for how to trigger the sequence. I did have a few options. I have a nice long USB cable with a repeater built in (for longer than standard) runs, and I could have run the cable from where the tripod was in the yard, through the dog door, and to my laptop on the counter, but – the dog door – that's the rub.  I didn't want the dogs rushing out in the middle of the night chasing some random noise and tearing into a cable… destroying either my camera, my laptop, or both. So I went (again – sense the theme for the evening yet?) easy… I just used my TC-80N3 timer control, but dumbed down. Rather than setting an interval, I simply set the camera on repetitive "rapid fire" (a bit humorous when doing 25 second exposures) and I just slid locked the trigger in the "on" position.  This would have the camera repeat 25 second exposures one after the other, knowing it would take shorter than 25 seconds to write the previous image to the card, so no fear of the buffer filling. It would just be a race – what would run out first? The battery (knowing it was hampered by the cold) or the card capacity?

My final setup was as follows: Gripped Canon 5D2, EF 17-40 f/4 – 17mm for 25 seconds @f/5.6 ISO 800

Hurry up and wait, I had topped off my spare pair of batteries as I was doing my tests, and then I kept them in my pockets to warm them in hopes of squeezing out every shot possible. I finally started the process about 1:30 AM Central time (the focus was just coming over the horizon and was still blocked by houses and trees), went off to bed, and figured I'd have at least 2 hours worth of exposures.

This morning, I found that the camera still had about 30% battery left, I'd filled up the memory card first. Yes, I had considered reducing the resolution, but not knowing how faint or bold the meteor streaks would be, I went for resolution. 683 images was the final tally.

(click to enlarge)

I took a quick peek after importing them into Lightroom, I did a quick scan and found some cool airplane traces but didn't have time to examine all the images. One that caught my eye (above) had an airline trace that curved; I can't recall seeing a curved trace in a sky shot of mine before. The images looked great, but I bumped the blacks to 15, and pulled the color temp back to 3750 (from an as shot of 4500 or so) which brought the tone from orange-y, and more like a nice cool toned night shot. The adjustments applied, I kicked off an export for some testing – 1600 pixel wide 70% quality jpg images.  I had some coffee, hopped in the car with the laptop on the passenger seat and motored up to the office. It took about an hour to render the images from the 21mp raw files.

On the drive, I figured I'd try two different ways to locate any meteor traces. I'd try a "star trails" stacking of the images, hoping that any non rotational trails would stand out better that way; or, alternatively, I could convert the sequence of images into a time lapse movie and look for the streaks by "scrubbing" the positional control back and forth as I looked at different areas of the screen.

Another visit to the AKASG brought me to a star trails action for PhotoShop CS4. A quick install later, I followed the instructions on the linked page and found what I thought to be the only meteor in my shot:

(it is diagonally up-left from the center of the swirl – click to enlarge)

Using Apple Quicktime to import my sequentially numbered sequence, I was soon happily scrubbing through a 6 fps frame rate version; watching the big dipper rise and set, as I looked for the meteor trail I'd seen in my stacked image. It wasn't hard to see that the image was near the end of my sequence. I referenced other more visible cues and soon knew it was after my neighbor switched on a light in a bathroom (which threw a vertical broad flare at about 1:44 in) and then right after a distinct long trail from chimney to tree to the right of the house. Within a minute or so, I'd found my image. Another "scrubbing" session showed another trail, a longer faint one just to the left of the tree (about 1:48).


After a bit of well timed research (thanks, via Google+, Jennifer Yu) seems to point towards my capture of one meteor (the first one I spotted), and one satellite (the longer, thinner trace). Jennifer's top tip – meteors are usually in just one frame, while satellites and other space junk will be in more than one. Also – meteors will usually be pointy at one or both ends while the satellite or other space junk will be a more even intensity across the length of the trace… Sorry to keep you in suspenders, but I've only the time and the energy to fight off the sandman long enough to post the images I've already referenced. I'll post a set of shots of the meteor and space junk later.

Thanks for reading, and please ask questions early and often either through comments on this post, or via the contact link in the header above!