We went up to the Look Rock parking area on the Foothills Parkway for the Perseid meteor show last night. We got there about 12:30 AM and stayed until about 3:30 AM.

I was surprised by the number of people and all the traffic. There were people at just about every pullout, and some of the lower ones were almost full. We saw several people with cameras.

The difference in the number of stars you can see in the darker skies is amazing as compared to the suburbs. The Milky Way is so bright you almost don't need a flashlight. There were so many stars and they were so bright, the Mrs. (our star navigator for the mission) had trouble orienting the various constellations at first. It helped when we finally got out a compass and figured out that North was the opposite direction from what I thought. (I believe I have mentioned before that I am directionally challenged.)

The meteor activity seemed to peak between 1:30 AM and a little after 2 AM. We saw dozens of meteors -- I'd guess around 50, but we weren't counting. Some were quite spectacular. I saw at least one double and a couple of rapid fire bursts of two or three.

I took about two hundred continuous 30 second exposures (actually more, but I deleted a bunch of non-productive shots along the way) using a cable release with a locking trigger and continuous shutter mode, and managed to capture exactly four meteors that I can actually see in the images. There were several more that I was sure I had captured, but I guess they just weren't bright enough. Click read more for the two best, which aren't all that great, and some notes on what I learned.

ISO 640, 28mm, f1.4, 30s, 1:26:12 AM

ISO 320, 28mm, f1.4, 30s, cropped, 2:56:28 AM

I tried a few different settings, and after studying the results here is what I have figured out.

Because the meteor trails are so faint, you probably need to go with the fastest, widest aperture (smallest f-stop) you have. In this case I went with a 28mm f1.4 lens and opened it all the way up to f1.4. You sacrifice some sharpness, but that doesn't really matter for these kinds of shots.

Because the meteors are moving so fast (they are only visible for a fraction of a second), it doesn't matter if you expose for 1 second, 30 seconds, or 30 minutes -- the exposure for the meteor is going to be the same fraction of a second when it appears.

The disadvantage of a longer exposure is that light from the stars and the Milky Way get overexposed and overwhelm the meteor's faint streak of light. You also get a lot more noise. If you try to adjust the exposure compensation afterward, you get even more noise. You also have a greater chance that a passing car or other incidental light will fog out a shot that just captured a meteor. Also, at 30 seconds the Earth's rotation elongates each star's point of light (an effect that is used to make "star trail" photos with much longer exposures measured in minutes and hours). You can see this effect in the second photo above which is cropped (i.e. "zoomed in"), making it more visible.

A wider lens would cover more sky and give you a better chance of capturing a meteor in the field of view, but the wider lens I have is limited to f2.8 and I don't think these would have registered. You would have to shoot at a much higher ISO, and images from my camera become almost unusable at ISO 800 and above for dark exposures like this because of the noise. (Yeah. I know. Get a Canon.)

Noise is the problem. In the second one above, there are a lot of red specks when viewed at 100% (which are not as visible in this reduced size image). I believe this is noise from the overheated sensor. (This was ironically the very the last shot I took, and the camera had been exposing almost continuously for nearly two hours. The battery held up for the duration, though, which surprised me.) There is also lots of other color and luminosity noise, some of which I was able to filter out in processing. Unfortunately I didn't take occasional dark frames to subtract out the sensor noise. I'd probably do that next time.

Some cameras have a "high ISO noise reduction" feature that automatically takes a dark frame and does the subtraction in camera when it stores the image. The problem with this is that it takes twice as long to process (in this case, 30 seconds for the exposure and 30 more seconds for the dark frame), reducing your chance of capturing a meteor by half.

All of this leads me to conclude that shorter exposures at the widest aperture and the highest ISO you can shoot without too much noise is the best formula. Next time I will try continuous 10 second or less exposures. Notice that I was able to capture one at ISO 320 in the second photo above, which yields less noise than even ISO 400 in my camera.

The disadvantage is that this fills up your camera's memory card a lot faster, in this case three times as fast for 10 second v. 30 second exposures (the file from a 1 second exposure is roughly the same size as the file from a 30 second or 30 minute exposure).

The solution is to shoot JPEG, which results in smaller files than the RAW format I was using which produces enormous files. But that means you have to get your exposure and white balance and other settings correct in the camera, because there is very little ability to adjust it afterward (i.e. you have to know what you're doing). But then you will have a lot more images to sort through afterward. On the other hand, they are faster to review than RAW images, and there's virtually no processing -- they're done and you either got the shot or you didn't.

If anyone else got any captures I'd love to see them. And any advice on getting better captures would be appreciated for the next time.

UPDATE: Here's a gallery with photos from around the world...