Technique: Part 2 – The Light

Continued from previous post

I use fairly typical high-speed photography techniques:

I leave the shutter open for a relatively long time (in a darkened room) and use a flash to illuminate the splash. The flash needs to be of a fairly short duration to stop the motion well. I use something like a 50 microsecond flash. I get this by modifying the photo sensor circuit on Vivitar 285HV flash. describes the technique (as well as the theory of xenon flash units). I replaced the photo sensor with a variable 20K ohm potentiometer so I can set the duration at will. Here is a graph showing the relationship between resistance and flash duration:

285HV Flash Duration

This graph doesn’t tell the whole story, though. The flash does not turn on and off instantaneously. Rather, it takes a bit to start, then ramps up and ramps down. Here is what it looks like:

Typical flash output at low power

What would you call the duration? Typically, people measure the width of the curve at the half of its maximum power. Here, it is about 30 microseconds. Also, notice that it takes about 45 microseconds from the time the flash is fired until it reaches one half its maximum. Individual units have their own signature, too. The maximum output varies some between units, as does the shape of the curve.

It is a balancing act to get well-lit photographs. On the one hand, a short duration flash is needed to stop the motion. On the other hand, with the close focusing distance, a fairly small aperture is desired to get a good depth of field. But, that requires more light. This can be solved with careful placement of the flash and, sometimes, multiple flashes.

A bright, constant light source (photolamp) and a fast shutter speed can also be used to stop the motion. This works reasonably well if the shutter is electronic (that is, the sensor is turned on and off electronically) rather than a mechanical focal plane shutter. The main issue with a mechanical shutter is that it is difficult to get it to release at the precise moment you’d like. There is always a bit of dither: sometimes it releases 50 milliseconds after it is triggered, sometimes 60 milliseconds, sometimes 54, etc. Another problem with the focal plane shutter is that even though it may be a 1/2000th of a second or shorter, it achieves this by sliding a narrow slit across the film plane, so each part of the frame is exposed at a different time over a 1/200th of a second (which might be used to good effect, but is not generally desirable). Finally, most cameras have still relatively slow shutters: 1/4000th of a second is 250 microseconds – five times as long as what I like to use.

Next, Control

This entry was posted in Photography. Bookmark the permalink.

30 Responses to Technique: Part 2 – The Light

  1. Florian says:

    I found your blog via google by accident and have to admit that youve a really interesting blog 🙂
    Just saved your feed in my reader, have a nice day 🙂

  2. Andrew says:

    Extremely inspiring works!

  3. Dave Ackerman says:

    I really enjoyed viewing a slideshow of your work. As a physicist and amateur photographer, I am taken by your artistic flare and exacting timing and lighting. Must be lots of fun when things are working well! Best of luck.

  4. brenton says:

    Great pictures ,keep up the good work.
    have a Vivitar 283 and I raise the camera’s ISO but feel that there still is not enough light,
    What make of flash do you use?

  5. martinw says:


    I use the Vivitar 285HV. Sometimes I use multiple flashes, since, as you have experienced, it is hard to get enough light and still stop the motion. I also use a highly reflective box to trap as much light as I can.

  6. brenton says:

    I will try that ,many thanks.

  7. brenton says:

    Bingo, Iv’e got it I think. Got the 283 working well with lots of reflectors etc. 1 absolute peach out of 230, not a good ratio, but I am happy.

  8. martinw says:


    Excellent. I would consider one out of 230 to be a fine success rate. This is tricky stuff.

  9. Brenton says:

    Tricky stuff indeed, Now I am into solenoids etc..see what you’ve started!
    I am thinking of using 2 sources of water, is this what you do?

  10. Ben S. says:

    How difficult is it to use multiple flashes all at high speeds? I was about to buy a second 285 to get more light, but was concerned about exactly the issue you mention here—that even though they were triggered at the same time, they would flash at different times. I guess one could add different delays for the two flashes, but I’d rather not…

  11. martinw says:


    You raise a good question. My measurements show a distinct variation in response times between units (a maximum of about 20 microseconds). You can see an overlay of the timings of a few units at:

    Admittedly, the tests I did were done with well-used units, so it may not be particularly fair. I should repeat the tests with new units.

    That said, 20 microseconds is not a terribly long time for many types of shots.

  12. Jeff Myers says:

    Martin, first of all my hats off to your work. It is truely inspriational to say the least. I have been painstakenly been working water drops for the last year off and on and just ordered the HIVIZ trigger to narrow my variables a bit more. My objective is to capture the collisions on call or better than 1/1000 ratio lol. Thanks again for your dedicated work.

    Jeff M

  13. martinw says:

    Thanks, Jeff. It is very satisfying to think that my work inspires others.

    Best of luck with your endeavors.


  14. fotoopa says:

    Thanks Martin for the nice description. As I’m actief into this highspeed work, I can confirm all this explanation. I’ve worked also with the vivitar 283 flash in the past, this is realy a good flash but now I use the newer nikon versions SB-80DX and the SB800. I read that you use micro controllers, I use a FPGA from Altera but the resolutions look the same. Working with 100 usec resolution is almost sufficient. I use 4 digits for all the timers into this format: xxx.x msec. Multiple laser detections give me the inputs for all the timings.

    Just like you I’ve multiple power outputs and all the electronics are homemade. Camera settings looks the same ( near F16..F22 at 1/200 sec) My shutter delay on the D200 is about 52 msec so my flash delay into the controller is set at 53.5 msec just into the window of the camera.
    Best of luck.

  15. martinw says:


    Thanks for getting in touch. I have seen your wonderful work for a couple of years now. You have built some very nice equipment and get great results.

    And thanks for the confirmation of my explanation and approach. Since much of this work is experimental in nature, it’s nice to have an independent assessment.

    As for the flashes: I have begun experimenting with the Canon 580EX, and it does seem that at very low power it has a reasonably short duration (though I have yet to measure it). I may well be changing to use it since it seems to have superior light quality in terms of evenness.

    Keep up the good work, and let’s stay in touch.


  16. AJ says:

    why don’t we use a High Shutter speed in capturing water drops .. i noticed that some photos the shutter speed of them are 1 sec or 3 .. So may i know the reason =)

    thnx alot

  17. martinw says:


    Good question. Of course, we need a short exposure to stop the motion. And a fast shutter speed will do that. The problem with a mechanical shutter is that it doesn’t expose all of the film (or sensor) simultaneously, especially at high speeds. SLRs use a “focal-plane” shutter that is two curtains that form a slit that moves across the film area exposing each part of the film at a slightly different time. So, while the “shutter speed” may be set to 1/4000 sec, the complete exposure process actually takes 1/500 sec.

    So, the classic high-speed technique is to work in a darkened room, open the shutter for a long time during which the drop splashes and a flash fires. The flash can be made to have very short durations (in specialized flashes that can be 1/2,000,000 sec.)

    Some digital cameras now have electronic “shutters”. These don’t have a mechanical device, rather the electronics turns on and off. They can provide very short exposures without the sliding-slit problem of focal plane shutters.

    One final note: the sliding-slit “problem” can be exaggerated and used to create strange effects. Andrew Davidhazy at RIT has explored this some to create “peripheral portraits”, e.g.:

  18. First let me just say: BRAVO!!! I absolutely loooove your work. You´re the reason why i´ve started playing around a litle with HSP myself, even though i´m certain i will never reach half your level. Still if i do…i will be mighty pleased even with that. 🙂 Heck i even watch that short Discovery-´documentary as often as i can only to draw further inspiration and knowledge. 🙂

    Just like many others i too struggle with the lighting. I use a Vivitar 283 (un-modified) and then a Sigma EX580 to light the background when needed. Preferably i would only like to work with one flash of course. Then i read in this blog that you use a highly reflective box to try and trap as much light as possible. Could you perhaps explain what that box is made of? Is the far end made up of your base background? I read that you use gels on your flash to get different tones so i reckon the background is white or something? Yup…trying to get all your secrets ít seems….muaaahahaha. lol. Well i should probably stop this ramble…bla bla.

    Have a great weekend and good luck with all that you set out to do in the future. I know for sure i´ll keep an eye on your work. 😉

    / Andreas

  19. P.S. Forgive me for any typos or poor grammar….i am swedish after all…and it´s late and i´m tired…and if that wasn´t enough i´m writing this using wifey´s EeePC…you know that tiny 10″ laptop that no ordinary mans fingers are made for typing on. :-/

  20. Rose says:

    Hi Martin,

    I wrote you before, but have those answered. I just wanted to ask you what you use for the actual “water drop” set up.

    I am getting a Nikon D60 with the kit lens and the 70-300. I will also be getting a couple of tripods, as I want an extra one for when I order the High Speed Photography kit you mentioned here on a couple of pages. I am going for the $125 one and will have to make the actual flash from the throw away camera.

    I have the idea of how it all works now, so my question is really what the set up should be itself. Where the water should fall from as to height, and the basin you use to actually do the photography. Yours looks like a box that has been angled in the front with some kind of a set up inside. If you could show some close ups of the actual box you use, I can figure out how to set it up, but I can’t see it close enough with the photo on your home page.


  21. Jim Murphy says:

    This is wild stuff, and it’s the first time I’ve ever seen this field in detail. As an EE, I’m curious about how you captured the waveform of the flash. Phototransistor sensors react too slowly, but maybe photodiodes? I’ll soon need to do some work along these lines to measure the speed of some objects accurately, but all I have to work with is an old Tek scope and an HP Universal Counter. Two flashes in rapid (and known) succession should do the trick, but I’ll need to measure the latency of the flash units. Any specific hints on how? (I’m about 10% through scouring your Fantastic website! Thanks!!!)

  22. Jim Murphy says:

    Ach!! I should have said EE _student_. Apologies – I grad in 6 months. (I’m a father too, though. Hard work juggling an irresistable desire to tinker, 4 kids, and artistry – well, musician, in my case.)

  23. martinw says:


    Astute question. The graph shown was done with a phototransistor, and you’re right: its response really is too slow and should be done with a photodiode instead. I’ll get to that one of these days…

  24. martinw says:


    I use a variety of different setups. The most common is a tray about 10 x 24 x 2 inches. The height of the fall is your choice, depending on just what effect you want. It takes some playing with to get a feel for it. (I know, I should write a entry about my observations on this.) But, a typical range is 6 to 18 inches.

  25. Brad Steels says:

    Hi Martin,

    Your work is truly awesome. What I think makes it so great is your lighting techniques. (Not to mention your creativity with different ideas and that you’ve built all your own electronics…!)

    I am using a recently available commercial timing device – StopShot ( – and I find it allows me to focus my efforts on the photography vice learning to build circuits, as you mention. As I experiment with lighting, I am wondering how you get such a nice uniform lighting using small flashes? I am experimenting with flash gels bounced from white foam core, and shot through frosted acrylic into a tray of water.

    I saw your comment regarding a reflective box (saw a glimpse of it on Discovery Channel too!) and that you’re using gels and shooting from behind. Any tips on good diffusers? or placement of gels? Do you use sheets of gel? or cover flash heads? I regularly study your shots to try to reverse engineer your lighting. It is fantastic. I really love the ripples with gentle colour gradients that seem to stretch to infinity.

    Anyway, I know that these are the master’s secrets – so I’m asking a lot. Even a point in the right direction would be much appreciated!

    Cheers, Brad

  26. Pingback: Touch of Green |John R Jennings Photography

  27. Ivan says:

    Hi Martin,

    I have admired your work for months now and must say that my incessant desire for my personal reiteration of the same hasn’t still declined. As a matter of fact it keeps building up every time I pull up your links 🙂 Anyways I am an absolute plonker in these type of things and would highly appreciate if you could bring me up to speed on a very few quick points. I get the idea of a long exposure thing during which a flash fires at a small fraction of a sec, I also understand that you use an IR beam of light for a droplet to be able to cut through therein triggering a flash to capture that very moment but what I don’t understand is how come that you trigger a flash unit the very moment when the latter droplet splashes onto the former one? Isn’t the unit supposed to trigger the flash when the former droplet cuts through the IR beam which spoils the attempt of the later one to splash onto it? Please clarify. That’s my understanding of the whole concept, if I am missing something or you could fill me in on the blanks that would be highly appreciated!


  28. martinw says:


    My control unit waits for two drops to break the IR beam, and then delays for a specified time before triggering the flash. So, the sequence is:

    1. Wait for drop to break beam.
    2. Wait for drop to break beam.
    3. Delay xxx milliseconds.
    4. Trigger flash.

    Does that help clarify?


  29. Ivan says:

    Martin, that makes a perfect sense! Could you please point me to a device for a commercial use which could employ this technique? Thanks a million!

  30. martinw says:

    Ivan, have a look at They have all the necessary parts, I believe.

Leave a Reply

Your email address will not be published. Required fields are marked *