In conventional photography exposure times are usually quoted in hundredths of a second, as 1/200 sec or 1/500 sec. As one moves into the realm of high-speed photography, with exposures times measured in thousandths and millionths of a second, this fractional notation becomes increasingly cumbersome. It is more convenient to quote exposures in milliseconds (msec) or microseconds (msec). Thus an exposure of 1/20,000 second would be given as 50 microseconds. At TC Nature, for example, we construct our own variable power flash units with fixed pulse-widths from 1/10,000 to 1/1,000,000 second, or, to use the easier terminology, from 100 msec to 1 msec.

Strobe lighting was developed by Dr. Harold Edgerton of MIT during the 1930s, but did not become widely available to photographers until after World War II. The main appeal initially lay in strobe's ability to stop movement through very short flashes of high-intensity light. In time, however, the emphasis shifted away from speed and towards power, as studio photographers revelled in the creative opportunities presented by unlimited illumination free of the uncertainties of sunlight or cumbersome floodlights. Today, big studio systems often operate at relatively slow speeds of around 1/200 sec, whereas most hand-held strobes typically operate at between 0.5 and 1 msec, many possessing the ability go down to 50 msec or less.

So far as we are aware, nobody is currently manufacturing flash equipment that provides short flash duration together with high power output. To get short duration flashes one must either commission the construction of special custom units or rely on small portable strobes. The former are prohibitively expensive while latter have limited power output.

Without going into technical details (and strobe theory is very technical!), strobes specifically designed for high-speed use differ in several important respects from conventional strobes. Among other things, they operate at far higher voltages (necessary for short flash duration but also potentially lethal), they require special hard-to-find high-energy storage capacitors and they need custom tubes (flashlamps).

Those who feel an irresistible urge to build their own high-speed units should consider very carefully before embarking on such an undertaking. Relatively few electrical engineers have experience in the field, which has many unexpected pitfalls and advice is almost unobtainable. Information and components are hard to find and there is a constant risk of lethal shock. One careless mistake can kill - and has killed! Particularly dangerous are units with separate flash heads connected to a central power supply. The cables inevitably receive rough usage, get wet and get trodden on. It is only a matter of time before the insulation breaks down and massive amounts of stored energy, at several thousand volts, are explosively discharged with the possibility of serious damage or injury.

Most people, therefore, are forced to rely on small, mass-produced thyristor-controlled strobes, which are relatively inexpensive. These can easily be adapted to produce very short light pulses, but this comes at the expense of power output because the thyristor cuts of the flash before the storage capacitor has fully discharged. The shorter the flash, the less light is produced. It becomes necessary, therefore, to link several together in order to generate reasonable light levels. Unfortunately, even this approach soon becomes impractical as the numbers needed rise exponentially - first 1, then 2, then 4, then 8, 16, and so on, doubling for each stop gained. Moreover, the effective pulse width begins to increase due to individual variation between units and difficulty in ensuring simultaneous firings, so one gradually starts to lose the ability to stop very fast movement as light levels rise and the flash duration expands.

You should be aware that most strobe manufacturers are overly optimistic when quoting performance data for their products. Moreover, the figures can be misleading. First, not all manufacturers conform to the same convention when assessing flash duration. Some measure at 1/3 peak, some at 1/2 and some at 1/10, making comparisons difficult.

It has become customary for the 'power' of a strobe (= "desirability" in popular imagination) to be represented in terms of its stored energy, measured in joules or watt-seconds. Here too the unwary can be seriously mislead, for this measure of stored energy - the product of voltage and capacitance - can bear scant relationship to the light energy actually produced by the flashlamp.

Whereas the oil-filled capacitors used in high-speed flash units typically dissipate over 90 percent of their stored energy, the cheaper electrolytic capacitors used in conventional strobes may, in extreme cases, actually retain up to 80 percent of their energy on discharge. It is, therefore, theoretically possible for two units, rated respectively at 180 watt-seconds and 800 watt-seconds, to put out exactly the same amount of light, the former converting energy at 40 times the rate of the latter. Caveat emptor!

High-speed photographers are plagued by shortage of light and nowhere is that more apparent than when considering backgrounds. If the background is the same distance behind the subject as the strobe light is in front, it will be 2 stops darker (and if twice the distance, 4 stops darker) - the results of the notorious inverse square law. Lighting the subject adequately is hard enough - lighting the background is that much more demanding, which is why many high-speed images have no background visible. High-speed images showing blue sky must be enhanced in some way, for in this case the background is infinitely distant and hence infinitely dark!

It is, of course, essential that the background light has the same duration as the subject light and be fully synchronized. If conventional flash is used on the background in conjunction with high-speed flash all sorts of unacceptable shadows and ghost images are generated. This means that in most circumstances high-speed images are shot against either pure black or a substitute photographic print background. There is also the possibility of adding a background subsequently, either digitally or in the darkroom.

People often ask whether it is possible to get accurate readings on a flash meter if the pulses are very short. We have only tested the later models of Minolta flashmeters, and find that they are accurate to within 0.2 of a stop at a flash duration of 25 microseconds - an acceptable value. We have no information on other makes.

Just as electrical energy is converted into radiant energy in a flashlamp, so within a film emulsion light energy is converted into chemical energy. Over the range of normal exposure times light energy bears a straight-line relationship to its effects on the emulsion. This is the basis for the so-called reciprocity law, which allows incremental changes in exposure time to be matched by reciprocal changes in aperture. With very short or very long exposure times this relationship breaks down, necessitating additional exposure time.

Each film emulsion reacts differently, but typically films rated at 64 or 100 ISO will require at least one extra stop at 50 microseconds and 2 extra stops at 25 microseconds to receive proper exposure. This must be added to any value your flash meter may give.

The spectral composition of light produced in a xenon flashlamp varies with flash duration, the shorter the pulse the bluer the light because more ultraviolet is generated. The effect will vary with different makes of strobe as some incorporate ultraviolet filtration and compensation. We find that at 25 microseconds a color temperature shift of about +81 mired is required, meaning that a Wrattan 85c filter will produce an appropriate correction. This, incidentally will add a further 1/3 stop correction to the exposure. If color balance is critical - for example involving skin tones in an advertising shot - it is reassuring to know that the Minolta Color Temperature meter gives accurate readings with short duration flashes.

 Filmstock

Many photographers are obliged to use relatively fast films because of having to work at low light levels. If possible, it is best to work with ISO 64 or ISO 100 film because the fine grain size makes it easier to produce crisp enlargements. At TC Nature we work with both Kodak and Fuji fine-grain reversal films.

Because so much trial and error is involved in creating a successful high-speed image, many photographers make extensive use of Polaroid instant films. At TC Nature we get through large amounts of Polacolor Pro 100 on every shoot and are profoundly grateful to Polaroid Corporation for help they have given us in this regard.

Polaroid backs are available for many high-end cameras. These allow each frame to be processed and viewed immediately, which greatly simplifies and speeds setting up. The alternative is to use 24 exposure Polaroid 35mm roll film, which must be developed in a special manual processor. This is less convenient because, unless you sacrifice many frames on each roll, you cannot monitor modifications step by step.