7.2 How Flash Works

Just as with the operation of the camera itself, it’s important to understand some basics of flash technology in order to better appreciate the capabilities and limitations of the flash unit when operating it in the field.  Although some of the material in this section may at first appear a bit esoteric or arcane, its importance will become more apparent in later sections of this chapter, and will help you to better understand why particular techniques work (or don’t work) in particular scenarios.
    Although it should be obvious to most camera owners, it bears mentioning at the start that flash light is both very bright and very brief.  Its brevity, in particular, is worth keeping firmly in mind both as you read the following sections and as you operate the camera in the field.  Because the flash’s pulse of light is typically much shorter in duration than the time that the camera’s shutter is open, the effect of shutter speed on overall exposure no longer follows the rules of reciprocity (introduced in section 6.1) when flash is in use.  Methods for dealing effectively with this failure of reciprocity in flash photography will be discussed in section 7.5.
    In contrast to the brevity of a flash pulse, ambient light—from the sun—is generally continuous in duration.  Note that ambient light needn’t come directly from the sun.  Photons of light arriving here from the sun will bounce off of the first object they encounter, thus being re-routed in a new direction.  They may bounce off of any number of earthly objects before eventually reaching the bird you’re watching.  After bouncing off of the bird, some light rays (beams of photons) will, just by happenstance, be reflected in the direction of your camera’s imaging sensor.  These photons, which originated at the sun and may have followed a rather contorted path before reflecting off the bird and being registered by your camera’s sensor, comprise what is known as ambient light.  Because photons are continually being emitted by the sun, ambient light is available during the entire interval that the shutter is open. 
    Flash light, on the other hand, has a very limited duration—typically around 1/20000 sec, which is significantly faster than the maximum shutter speed of even today’s pro DSLR cameras (excluding those with electronic, rather than mechanical, shutters).  Because flash is (typically) so much faster than the camera’s shutter, changes in shutter speed often don’t have any effect whatsoever on the amount of flash light that is collected.  The figure below illustrates this via a left-to-right timeline (i.e., the horizontal axis represents time).

Fig. 7.2.1 : Flash, ambient light, and shutter speed.  The horizontal axis represents
time.  Since the flash pulse is typically far shorter than the exposure duration, changes
in shutter speed usually don't affect the amount of flash light received.  Doubling the
shutter speed halves the ambient light received, but has no effect on flash.  As a
result, the ratio of flash to ambient light changes.

In this figure (above), the flash duration is seen to be far shorter than the amount of time that the shutter is open.  When the shutter speed is doubled from 1/100 sec to 1/200 sec, the amount of ambient light collected by the sensor obviously halves, but since the flash emission is still wholely contained within the exposure interval, the amount of flash light that is collected remains unchanged.  In this case we say that the flash ratio—the ratio of flash light collected, versus ambient light collected—has doubled.  Changing the flash ratio can have profound effects on the captured image, most especially on the relation between the lighting of the subject (the bird) and the background.  Later in this chapter we’ll see some examples of different flash ratios, and consider methods for choosing the right ratio.
    Differences in lighting between the subject and the background obviously depend very strongly on their relative distances to the camera.  If the subject is much closer to the camera than the
background (i.e., any trees or bushes that are visible behind the bird), then the flash will have a disproportionately greater effect on the subject than the background.  This is due to a fundamental law of physics regarding electromagnetic radiation (of which light is one form), referred to as an inverse power law:

For those not fond of mathematical notation, the above formula indicates that if the distance (between the camera and any reflective surface such as a feather or leaf) decreases at a certain rate, then the gain in illumination due to flash will increase much faster.  There are several important consequences of this basic law of light.  The first is that in dim environments, any reduction in the distance-to-subject will tend to result in an appreciable increase in illumination of the subject due to flash.  If the images you’re getting are too dim even with flash at full power and exposure settings at their practical limits, and you think you can get a few steps closer to the bird without scaring it away, then it’s usually a good gamble to try it.
    A second consequence is the differential effect that flash can have on a subject and its background, if the two are sufficiently separated.  The figure below illustrates the situation schematically.

Fig. 7.2.2 : The relative distances to the bird and the background have an exaggerated
effect on the flash illumination of the two, due to the squared distance term in the
inverse power law.  The further separated the bird and background are, the greater
the difference in flash illumination (in a nonlinear manner), thereby helping to make
the bird stand out more.

If most of the background elements are positioned some distance behind the bird, then the bird can often be made noticeably more prominent in the image by increasing the flash ratio—i.e., by increasing the flash output and/or decreasing the amount of ambient light that is collected (by increasing the shutter speed).  If, for example, the distance from the camera/flash to the background is fully twice the distance of the camera/flash to the bird, then the amount of flash light that reaches the bird will be four times (not just two times) the amount that reaches the background.  This is simply due to the inverse power law.  Thus, if by repositioning the camera you can find an angle which places the background further from the bird, then the effect of the flash should accentuate that difference even more in the resulting image.
     Now let’s return to the issue of shutter speed versus flash duration.  An important fact which many people don’t realize is that the shutter doesn’t open instantaneously, nor does it close instantaneously.  There is an opening latency and a closing latency (which are typically equal)
that must be accounted for, as we explain next.  Let’s imagine what happens in slow motion.  Refer to Figure 7.2.3. (below) while considering the following chain of events. 

Fig. 7.2.3 : The phases of shutter operation.  At first, the shutter is still closed (A).  Then
the first curtain begins to open, exposing progressively more of the sensor to light (B).
When it reachees the far side, the shutter is fully open (C).  Eventually the second curtain
begins to close (D).  When it reaches the far side of the sensor, exposure is finished (E).

To begin with, the two curtains of the shutter are located together on one side of the sensor (part A of the figure).  When you push the shutter-release button, the front curtain of the shutter mechanism begins to open. As its trailing edge traverses the sensor window (part B), more and more of the sensor is exposed to the light coming in through the lens.  Eventually, the front curtain (also called the first curtain) of the shutter reaches the fully-open state.  Now the sensor is completely exposed (part C).  When it’s nearing time for the exposure to end, the rear curtain (or second curtain) begins to close.  As its leading edge traverses the sensor window, less and less of the sensor is exposed (part D).  Eventually, the rear curtain catches up to the front curtain (at the far edge of the sensor window) and then the sensor is completely covered and exposure has fully ended (part E). 
    Because the front and rear curtains do take some non-zero amount of time to travel from one edge of the sensor window to the other, the above scenario has to be slightly modified when using very short exposures.  When the shutter speed is set above a critical threshold called the maximum sync speed, the rear curtain has to begin is traversal before the front curtain has even reached the far edge of the sensor window—that is, the rear curtain chases the front curtain across the sensor window.  In the case of very high shutter speeds, the rear curtain may be only just barely behind the front curtain during their paired traversal, and in that case the two curtains form a narrow slit that slides across the sensor, allowing light to accumulate at each photosite while it’s exposed within the slit.  That exposure of individual photosites occurs at slightly different times, of course, depending on when the traveling slit reaches them.  This scheme works fine for ambient light, but for flash photography it complicates things, because a single flash pulse would illuminate some parts of the sensor and not others (depending on which parts are swept out by the chasing curtains during the flash pulse). 
    The clever solution that camera engineers have come up with is to replace the single, strong pulse of flash light with a near-continuous stream of weaker pulses that occur throughout the duration of the curtain chase.  If the pulses are fast enough (and indeed, the engineers have ensured that they are), then all photosites (i.e.,
pixels) will receive approximately the same total amount of flash light.  That’s obviously a good thing.  This latter fix has been dubbed high-speed sync (HSS) by the camera manufacturers, and is necessary whenever your shutter speed exceeds the maximum sync speed (MSS) of your camera.  Most cameras have an MSS of about 1/200, with higher-end bodies ranging up to 1/250 or 1/300. 

Fig. 7.2.4 : Normal flash mode versus high-speed sync (HSS).  In part A, the shutter speed is
below the maximum sync speed, so the flash can pulse just once.  In part B, the shutter speed
is faster than the max sync speed, so the first curtain doesn't have time to completely reach
the far side before the second curtain has to begin to close.  At no time is the entire sensor
exposed, so a single brief flash pulse won't work.  Instead, many mini-pulses are emitted,
mimicking the effect of a continuous light source.  Note that the two intervals are not
drawn to the same scale: the interval in part B should be much shorter than in part A.

Most name-brand flash units support a high-speed sync mode, which you’d typically activate by pressing one or more buttons on the external flash unit itself (though sometimes it can be set remotely through the camera).  There are some negative consequences of high-speed sync, however.  First, there’s the obvious drawback that using faster shutter speeds results in less exposure time for collecting light from the flash (as well as ambient light).  In addition, most flash units are limited to providing a significantly reduced power output in high-speed sync mode, due to the longer effective duration of the summed mini-pulses.  Because standard flash batteries (typically AA’s) can’t provide current fast enough, the power for a flash pulse (or a stream of mini-pulses) has to be stored in advance in a high-speed capacitor.  Because the sum of the many mini-pulses of a high-speed-sync (HSS) flash invocation can significantly exceed the power demands of an equivalent, non-HSS pulse (at slow shutter speeds), the overall power output typically has to be reduced in order to make the power from the capacitor last for the full HSS interval.  Keep in mind also that in HSS mode, much of the flash output is wasted, since many of its reflected rays that normally would strike the sensor (in a non-HSS exposure) are instead reflected away by some portion of the front or rear curtain as the narrow slit formed by the chasing curtains traverses the sensor window.  This isn’t just a loss of efficiency: since the maximum amount of flash power available (which is dictated by the flash unit’s capacitor rating) is limited, for higher shutter speeds it’s simply not possible to compensate for the HSS slit effect by increasing the flash’s output power.
    It’a also useful to know that for most flash units, the
power level that you set on the flash’s control panel is actually a duration setting.  In other words, most flash units produce light at the same (peak) intensity at all times, so what you’re really changing when your dial in a different power level is the flash duration: higher output levels are actually achieved by using the same (peak) emission intensity over a longer time interval.  As mentioned previously, these time intervals (in the non-HSS case) typically range from about 1/1000 sec to 1/35000 sec.  We’ll see later how this fact can be used to effectively freeze birds’ wings by choosing an appropriately brief flash duration.
    The foregoing discussion of shutter curtains has importance even outside of the high-speed-sync scenario.  Many flash units provide an additional mode allowing synchronization with either the front curtain or the rear curtain.  In part A of figure 7.2.4, we depicted the flash pulse occurring shortly after the opening of the shutter.  This is the default synchronization mode for most flash units, and is typically referred to as first curtain sync.  The other option is obviously second curtain sync.  A potential problem with first-curtain sync is that if there’s motion in the scene and the ambient light isn’t completely insignificant relative to the flash intensity, then objects in motion can appear to be moving backward in the resulting image.

Fig. 7.2.5 : First-curtain sync causes motion to appear backward.  The
snowflakes were frozen in position by the flash pulse, and then ambient
light traced out the following interval of their trajectory.  Using second-
curtain sync would have fixed this problem by emitted the flash pulse
at the end of the ambient interval rather than at the beginning.

As you can see in the figure above, the snowflakes appear to be moving up instead of down, because I made the mistake of using first-curtain sync instead of second-curtain sync.  In this case, the flash fired as soon as the first curtain had cleared the sensor window, freezing (no pun intended) the positions of the snowflakes at the beginning of their trajectories; the continuous accumulation of ambient light during the remainder of the exposure resulted in each snowflake having a downward tail.  Most people would interpret that downward tail as indicating that the snowflakes were moving upward, when in fact they were moving downward (as snowflakes normally do).  A similar effect can occur with the wingbeats of birds, so the issue of first/second curtain sync isn’t limited to just snowflakes and raindrops.