Just beginning with photography.

Consider this image:

enter image description here

(This photo was taken by Harold "Doc" Edgerton in 1964 using his Rapatronic camera.)

I am looking for a camera that can capture such an image.
I am not looking for more fps. less fps is fine, but it must be able to capture such an instance.

What features should I look in such a camera? How fast should be the shutter speed? What other things I should look for?


5 Answers 5


As Michael said, shutter speed is largely irrelevant – flash duration and timing relative to bullet passage is what counts. Shutter speed can be as SLOW as is helpful – maybe even "bulb".

Assuming that the photos is not doctored:
Bullet speed should be arranged to be as slow as possible without adversely affecting the affect on the target. Bullet needs to not move appreciably during exposure. So, how long (or short) does the exposure need to be?

Say bullet was travelling at 100 feet per second.
Bullet is ~~~ 1 inch long
and has moved a maximum of 5% of its length – maybe less.
So time = 1 inch x 5% / 100 feet/sec
= 1/24,000 th of a second or
~= 40 μS.

At 1000 fps that would be 4 μS.
1 μS would be even better.

Will a Xenon flash tube do this?

Wikipedia - flashtube

  • Discharge durations for common flashtubes range from 1 microsecond to tens of milliseconds, and can have repetition rates of hundreds of hertz. Flash duration can be carefully controlled with the use of an inductor.

See reference to PerkinElmer catalogue below.

Will an LED Based "flash" do this?

Achieving this sort of result with LED lighting would require significant power levels.
eg assume 1 lux, 250mm x 200mm illumination area, 10 μS exposure, 100 ISO, f 1/1 (trust me)
EV = t = 10-5
For real world results if we want say EV = 100 at f/1.4 we need illumination of
Ev_100/EV_1 x 1.4² / 10-5 = 2 x 107 lux (!)
Over an area of 250mm x 200mm = 0.05 m² that's
2 x 107 x 0.05 = 106 lumen
Leading edge Top LEDs manage around 200 lumen/Watt so power =
106/200 = 5,000 Watt of LED illumination (!!!).
In practice this is only required for 10 μS so actual power is a fraction of a Watt BUT the LEDs MUST be able to produce the peak power level required   and modern white LEDs have a peak: continuous ratio of typically less than 2:1.

So – "not really", so far.
Whereas, a suitably designed Xenon flashtube can be capable of producing these very high levels of power for extremely short periods.

Exceedingly nice PerkinElmer technical guide High Performance Flash and Arc Lamps


enter image description here

Figure R, page 13


As Paul said, the photo was produced by the late Harold Edgerton and is one of the ones he is most known for. You need to study his website re methods and equipment.

Alas, the flash duration that I calculated above by rule of thumb was just about right   and my "nice" value was spot on   one millionth of a second. Have a look at the PerkinElmer catalog for what you'll need.

This famous photo of Harold Edgerton's is specifically identified as having a 1/1,000,000th second exposure time.

enter image description here


See also:

Harold "Doc" Edgerton Collection website

HE - stroboscopic method

Bullet through banana

Many here

E&OE - it's late and there is still work to be done. I may very well have dropped or added a power of 10 or few above or done something really silly - by all means do point out any errors.
Be kind :-)

  • " Bullet speed should be arranged to be as slow as possible" , how to arrange this. Can you elaborate on this point? I did not get it.
    – gpuguy
    Aug 13, 2014 at 12:52
  • 3
    @gpuguy Bullet speed depends on load used. People accustomed to reloading bullets will be able to produce a "load" with a muzzle velocity well below what is normal for a given calibre firearm. Most real world weapons achieve MVs in 1000-4000 fps range. That's above my upper v of 1000 fps. A very light load indeed is liable to be in order - the apple will probably still die . chuckhawks.com/rifle_ballistics_table.htm Aug 13, 2014 at 13:26
  • 1
    @gpuguy If you don't know anyone who reloads and is willing to spend a fair amount of time in trial/error work to figure out how low the gun you're using can go while still working well (abnormally low velocities can cause reliability problems; mostly for semi-auto designs); looking for sub-sonic ammunition will get you on the bottom edge of what's available commercially (mostly this will be for older low power handguns). Aug 13, 2014 at 21:04
  • We can't reduce it to 100 feet per second though, that's slower than paint balls - the answer is off by at least ten times. All my handguns are over 1000 feet per second, the rifles in the 3000-3500 range. Just add a zero to every number in the answer. Also, please add some safety information - NEVER place yourself in front of a live firing line. Put your equipment out there and risk it getting shot, but don't you be anywhere in front of firearms.
    – Jasmine
    Aug 13, 2014 at 23:42
  • 2
    @Jasmine Doesn't getting the muzzle velocities wrong by a factor of 10 indicate that Russell isn't really qualified to give that information? Also, I think it's implicit that people should take appropriate safety precautions whatever they're doing. For example, advice on landscape photography isn't peppered with "Don't fall off cliffs. Be careful about dangerous animals. Bring adequate supplies if you're in the wilderness. Check the weather forecast." and so on. Aug 14, 2014 at 11:03

The camera doesn't matter. In high speed photography such as this photo it is all about the speed of the flash and being able to fire it at precisely the correct moment. The flash is usually fired with an electronic trigger that reacts, after a specific delay of several milliseconds, to the noise of the gun firing . The camera's shutter can remain open for several seconds before and after the exposure because other than the flash the room is totally dark.

  • For this image, what should be the flash/shutter speed ?
    – gpuguy
    Aug 13, 2014 at 12:33
  • 5
    "The camera's shutter can remain open for several seconds before and after the exposure" - the flash is providing all the light, so it doesn't matter.
    – Philip Kendall
    Aug 13, 2014 at 12:35
  • 1
    9x19 has velocity of 390 m/s and about 1cm in length if you can tolerate 5mm motion blur you would need 0.005/390 ~ 1/80000s flash speed.
    – Andrew
    Aug 13, 2014 at 12:38
  • 2
    @gpuguy - not those specs, no - they tell you how bright and what it connects to but not for how long. Flash duration or speed is often measured as the time the flash takes to get from full power to half power or 10%. Also called t/0.5 and t/0.1 times. photo.stackexchange.com/questions/1620/what-is-flash-duration would be the place to look, and maybe the answer there needs some more attention. Aug 13, 2014 at 13:34
  • 1
    @gpuguy - Questions are welcome - but they need to be based on the foundations provided by prior answers. You need to look at the material provided and understand what you are being told before you ask more questions. If you do not understand by all means ask about what has been written. The 1/24000th second was the time that the bullet would move 5% of its length given the assumptions I stated - including a VERY SLOW muzzle velocity. Read the portion I posted from the PE catalogue (it was there because it was relevant) then look at the catalog. No 'normal' camera flash will do this .... Aug 13, 2014 at 14:04

I took shots like this in Doc Edgerton's lab in the 80's. The setup was simple.

  • Basic film camera, nothing special

  • Rifle permanently mounted at the end of a long rail

  • Sliding support for target which can be adjusted along the rail

  • Microphone connected to the strobe (on or nearby the camera)

From experience we would guess roughly where to place the microphone between the rifle and the target. Can do quick calculations to get an estimate.

  • Turn off the lights. Stand back! Fire a practice shot and observe where the bullet is. The very short duration of the strobe makes the bullet stand out quite clearly.

  • Turn on the lights. Move the target/support to where the bullet appeared. (simpler than moving the microphone or adjusting the delay on an electronic trigger - just move your target to where the bullet is going to appear, based on the current location of the mic)

  • Repeat test shots to fine tune the location of the bullet, which was very repeatable and consistent.

  • Final shot: Place object on the support. Turn off the lights. Set camera to bulb mode. Fire.

enter image description here


This photo was taken by Harold "Doc" Edgerton in 1964 using his Rapatronic camera. He used a strobe speed of about 1/1,000,000th of a second.

The bullet was traveling at 2,800 feet/second.


If you move away from main stream xenon flashes & have a look at air gap flashes they can hit 1/1,000,000 of a second. Plans on the internet are DIY (REALLY dangerous stuff) but there may be commercial vendors as well. Search on air gap August 1974 issue of Scientific American.

And look at the Vela One http://www.vela.io a LED system (quite pricey though) & there are DIY of those LED flashes as well. But I'll agree with others & you never get the amount of light you want/need but they do have sharp attack & decay timing.

the book you want : Engineering and high speed photography, Willian G. Hyzer 1962 -of course its about film but most of the flash & shutter stuff has not changed.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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