Alley in Pisa, Italy

by Lars Kotthoff

submit your photo


Hall of Fame
View past winners from this year

Please participate in Meta
and help us grow.

Take the 2-minute tour ×
Photography Stack Exchange is a question and answer site for professional, enthusiast and amateur photographers. It's 100% free, no registration required.

I was out last night trying to shoot the Geminids meteor shower. I was using a Canon 5D2, 17-40mm f4.0 lens set to 17mm and f4.0, ISO 400. The camera was mounted on a non-tracking mount (only because it failed in the field!) and I was doing 60 second exposures. So I expected star trails but what I didn't expect was dim meteors.

Acually, of the 90 or so shots I only grabbed 2, and that's what puzzled me. The skies were very active, at least 2 per minute and the camera was aimed at a good spot.

In particular, one meteor was noticeably bright and passed between Orion and Jupiter, right where the camera was pointing. I was pretty excited about that, it would be a great shot. I was disappointed when I got home. At first, I couldn't find it, it was so dim.

enter image description here

Look about halfway down and a third of the way over. The meteor is pointing at about 2 o'clock. It is really hard to see.

Why? It was easily as bright as Betelgeuse and Bellatrix, the two bright stars in top half of Orion, you can see Orion's belt at the far right.

I know the shutter was open when the bright one passed by. Are meteors really that dim?

References:

  1. meteor question on this site
  2. article about meteor photography
share|improve this question

6 Answers 6

up vote 16 down vote accepted

Meteors can be dim or bright, depending on the size, duration, and intensity of their entry. That is generally immaterial to the process of photographing them, however. The first concern of most wide field astrophotographers is ISO, and I think that leads to the frequent use of TOO LOW of an ISO setting. I was also out last night photographing the night sky, hoping to capture a few good photos of a large geminid meteor. I managed to capture a few, such as this one here:

enter image description here

It may surprise you to find out that the statistics for this shot were as follows:

  • ISO: 3200!
  • Shutter: 4.0s
  • Aperture: f/2.8
  • Focal Length: 16mm (EF 16-35mm f/2.8 L II)
  • Camera: EOS 7D

I pushed the exposure settings about as far as I could. I wanted to minimize star trailing (in case I printed any of them), so I wanted a shorter exposure, and 6s was about as high as I could go before encountering obvious star trails. I'd have chosen ISO 6400, however that blotchy red color noise really eats away at IQ, so I stayed at ISO 3200 with an aperture of f/2.8. As you can see, with the maximum amount of light down the lens and a very high ISO setting, the meteor (which in this case lasted about 2.5 seconds, and was moderately bright) clearly stands out.

The notion of "time-on-pixels" set forth by j-g-faustus in response to BobT's answer is critical here. You want the amount of time a meteor has on each pixel it covers to be similar to the time stars have on each pixel they cover. The greater the ratio is, the dimmer a meteor will appear relative to the stars. The trick is to expose such that the ratio of time-on-pixels for meteors is similar to that of stars. That means reducing exposure time, which necessitates pushing ISO (possibly much higher than you would normally think to use). At high ISO, read noise is effectively a non-factor. The primary source of noise is photon shot noise, and de-noise algorithms tend to be best at removing that kind of noise. If I'd had a Canon 5D III at my disposal, I'd have been shooting at ISO 6400, possibly even 12800 if it was necessary to reduce the time-on-pixels ratio of meteors to stars. One need not aim for a 1:1 ratio...meteor intensity is higher over a shorter duration, but one does not want a 50:1 or greater ratio.

The photo here has had some post processing. I applied a tone curve to improve contrast, adjusted white balance to bring out color, boosted saturation a bit and applied some noise removal (although not nearly as much as you might think.) Here are a few more shots, all taken with a shutter speed between 4-6 seconds, ISO 3200, f/2.8 aperture. Each has had similar processing. These meteors were all dimmer than the one above due to their shorter durations, however in terms of time-on-pixels the ratios are still fairly small. One of them was a momentary spark that lasted for less than a second, and this is fairly dim, but still visible enough.

enter image description here enter image description here enter image description here


Regarding hit rate, out of 90 shots, you got two. That is one for every 45 shots, and since your exposures were 60 seconds long, that means one every 45 minutes. In my area, Geminid entry rate was around 50-60 per hour during peak (12am - 3am), and my exposure and interval period was about 9 seconds (4 second exposures, 5 second interval most of the time). That means I'm making about six exposures every minute, and there was just under one meteor a minute. One would expect that nearly every sixth shot would have a meteor in it.

Contrary to the simplistic math, one has to factor in the ratio of the sky that your camera frame covers. At 16mm on an APS-C, my lens covers a 35° x 24° field of view, out of a sky that encompasses an ideal range of 360° x 180°. Accounting for the fact that meteors generally filled about 3/4 of the sky from their radiant source, I figure a range of sky of 270° x 135°. My frame covered about 13% of the horizontal and about 18% of the vertical, so quite a lot of sky that was potential area for a hit was out of the frame. Personally I saw about 1-2 meteors every 2-4 minutes, however only a fraction of those actually ended up in the frame. For every sequence of 100 shots (which took about 15 minutes at 9s per shot), my camera picked up 2-3 meteors.

Not knowing the region you were in (there were good, fair, and poor regions for viewing the Geminid showers), I can't calculate the rate at which your camera should have picked up a meteor. In a poor area, the rate could be anything from zero to a few dozen per hour. In a fair area, where I was, the rate was easily 50 per hour or more. In an area with good viewing, the rate was 120 to 190 per hour, which would have been 2-3 per minute. With an exposure time of 60s (too long, in my opinion...a higher ISO and shorter shutter would generally be better for wide field) in a fair viewing region, you should have picked up a lot more than two out of 90 frames with meteors in them.

share|improve this answer
    
thank you for sharing and explanation! –  peter_budo Dec 15 '12 at 16:56

I think your main problem is not that they are dim, but that they are so fast & brief.

For a 60 second exposure I'm surprised how dim the starts are.

I would personally run at a higher ISO and shorter exposure in order that the brief moment of the meteor is better captured. For example if you ran at ISO 800, and a 30 second exposure you would benefit from:

  • Half the star trail length

  • the same star brightness

  • a much brighter meteor

The potential down-sides are more exposures needed, and possibly more noise.

share|improve this answer
    
but look at the meteor trails in the 2nd link I posted, they are so bright! –  Paul Cezanne Dec 14 '12 at 12:42
    
true, but not all meteors are the same brightness, if you say it was only as bright as a star then it was indeed rather dim by meteor standards. –  Darkcat Studios Dec 14 '12 at 12:49
1  
@PaulCezanne: And the copyrite dates on those two bright shots were two years apart. Those were his best two out of likely tens of thousands of exposures over a long time. Catching a spectacular meteor shot is a numbers game. –  Michael Clark Feb 21 '13 at 12:10

Also, regarding that link, here's an important quote (emphasis mine):

"During the 1998 Leonid shower, I had six cameras going concurrently with 16mm, 18mm, 24mm, 35mm, 50mm, and 85mm lenses. I shot 5 - 10 minute exposures consecutively for about 5 hours. I ended up with about 30 frames per camera, for a total of 180 frames. I recorded 3 meteors. One was not even a Leonid..."

So shooting 90 frames and getting 2 meteors seems about the right hit rate.

I think the main problem is the long shutter speed, meteors are very short lived, the extra shutter time is not going to make them any brighter. Had you shot at ISO1600 for 15 seconds, the meteor would have been 4 times brighter.

It's similar in principle to balancing flash with ambient, the meteor is the flash, and it's brightness depends on aperture/ISO, the stars are the ambient so their brightness depends on shutter-time/aperture/ISO therefore you can control the balance between the two using the shutter speed (I'm counting longer trails as appearing "brighter").

share|improve this answer

If you look at the background of the second link you can see a lot of sensor noise. This means that it was probably a longish exposure or that the ISO was cranked up to some high value (as the article states). "High value" depends on the camera, but 1600 to 6400 is a good range to experiment with. If you take your original and do some Photoshoppery to enhance the trail (crank up the exposure) you'll see a slightly better trail. I have no doubt that the images you've linked to have had their contrast and exposures optimized to show the trails. Here's your JPG with exposure and contrast adjusted...

PhotoShopped example of Paul Cezanne's image

share|improve this answer
    
Fortunately, I shoot in RAW so I'll play around with it. I was just surprised at how bright it was to my eye yet dim on the image. –  Paul Cezanne Dec 14 '12 at 18:29
1  
@PaulCezanne I think it makes sense if you consider that each pixel/sensel is exposed independently. If the stars move say 10 pixels in 60 seconds they get an effective exposure time of 6 sec per pixel, while the meteor may move say 200 pixels in two seconds and get an effective exposure time of just 1/100th of a second per pixel. As per Matt Grum's answer, the only in-camera fix (that I know of) is to increase ISO or aperture. –  j-g-faustus Dec 14 '12 at 22:42
    
@j-g-faustus: Technically speaking, just increasing ISO is not actually a fix. As you stated, over a lengthy exposure of 60 seconds, the meteor's time-on-pixel is a fraction of that for each star. Increasing ISO does not actually change the sensor's sensitivity in any way, it just changes the point at which each pixel becomes fully "saturated". Assuming no change in exposure time, that wouldn't actually fix the issue at hand...short time on pixels for the meteor with long time on pixels for the stars. The best fix is to get more light on the sensor, and only a wider aperture can do that... –  jrista Dec 15 '12 at 17:10
    
However, reducing the exposure time along with an increase in ISO changes the time-on-pixel ratio while maintaining or improving exposure. In my shots, I used a 4s exposure, which is 15x faster than Paul's. That means the time-on-pixel ratio between meteors and stars is much lower, possibly even equal (in the case of a longer-lived meteor that started as soon as the shutter opened, the ratio approaches 1:1.) A higher ISO does nothing other than maximize saturation before readout so the image looks better, but it does not actually improve the sensitivity of the sensor. –  jrista Dec 15 '12 at 17:13
1  
@j-g-faustus: I totally agree. I think the concept is sound. I had sort of had that idea in the back of my mind...although it was more centered on avoiding star trailing and minimizing star exposure. The way you termed the whole idea is much better, though...time on pixels. Critical factor, IMO, in getting good meteor shots. I also agree that once startrails start forming, there is no benefit to additional exposure. The rule of 600 is probably useful here, and the notion that using one of the top 2-3 native ISO settings. You may not always need/want to use the highest, but very high is good! –  jrista Dec 16 '12 at 0:21

Even if the meteor was about as bright as the stars you need to keep in mind that the meteor was only illuminating your sensor for at most few seconds, and typically less than one, while the stars where for the entire length of the exposure and their light is spread out over a much larger area on your sensor than the stars were.

If you're new to astrophotography, you generally need to play around with light curves significantly in order to get the dimmer objects in your image to appear clearly because bright stars will typically saturate the pixels they're illuminating to your sensors max while almost everything else is much fainter.

share|improve this answer
    
correction: the meteor would illuminate the sensor for a FRACTION of a second. –  Darkcat Studios Dec 14 '12 at 21:41
    
@DarkcatStudios Meteors can last for several seconds; this is most likely to happen when the radiant is low on the horizon because the meteor rate of penetration into thicker layers of the atmosphere is slower. amsmeteors.org/mcleod/mcleod4.html –  Dan Neely Dec 14 '12 at 22:06
    
I was referring mostly to the type expected this week - small and very fast - burn up very quickly –  Darkcat Studios Dec 14 '12 at 22:37
    
And I was making a general answer... –  Dan Neely Dec 14 '12 at 23:35
    
@DarkcatStudios: Having observed the Geminids last night myself, I remember quite a few that lasted for several seconds, were particularly bright green in color, and had large tails. The majority of the Geminids are shorter, but there are plenty that last for long enough periods of time. –  jrista Dec 15 '12 at 0:19

It's a numbers game. The longer your exposure, the brighter the stars compared to the meteors. The wider your lens, the greater percentage of the sky you cover. The more shots you take, the more likely you will catch a nice meteor.

I also shot the Geminids this December with an EF 17-40mm f/4L on an EOS 5DII, the same gear as you used. Most of my exposures were 30 seconds or less at ISO 1000 and f/4. I manually focused using Live View x10 on a bright star, then refined on a dimmer one. I used a cheap timer that attaches to the cable release port and set it up to take continuous shots. I had to pause frequently to clean the frost off the UV filter on the front of the lens (placed there due to the temp/dew point). Over several hours I wound up with almost 400 exposures. I got one great shot and 3-4 other confirmed Geminids. I also got a couple more that I thought were meteors until discovering their paths and appearance times matched known satellites in orbit!

The scaling makes the meteor's line look uneven, but it is smooth on the full resolution version. It was probably the brightest meteor I saw all night, and I was lucky to have the camera pointed the right direction at the time. I hope I didn't wake the neighbors when I saw it and realized the camera was pointed in that direction and the shutter was open! This was cropped from 5616X3744 to 3924X2616 and then resized to 1536X1024 for web use. It is best viewed on a dark background.

Meteor

Here are the settings I used in Canon's DPP to convert the RAW file. Noise reduction was set to 2(lum) and 3(chrom). For the two following pictures I used the same settings except WB was set to 3700K. Settings

This was a more typical capture. Full 5616X3744 resized to 1536X1024 for web use. Notice the way it flared just before dimming. To the naked eye at the time that flare at the end lit up half the sky!

Meteor 2

This one has an airplane in the lower left. What appears to be a meteor in the upper right is actually CZ-2D, a rocket booster launched 11/20/2011 from the Peoples Republic of China. It is a spent rocket booster and is probably tumbling, which explains the variation in brightness. Cropped to 3968X2645 and resized to 1536X1024.

Plane+Satellite

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.