The brains. This is the part of the system that accepts data from the input device, decodes it, and stores it in memory, whether that be in our head or on a storage chip.
An SLR is a type of camera that employs an optical tunnel where light rays reflecting off the subject enter through the lense and are directed to the eye via a mirror. When the shutter release is activated, this mirror flips up, our view goes black (called blackout time), the shutter opens, and the light has a direct path to the imaging surface.
This imaging surface can be either film or a silicon chip. Common chips used in today’s digital cameras are derivatives of either a Charge Coupled Device (CCD) or a Complementary Metal Oxide Sensor (CMOS). Both of these types of sensing devices have their advantages and disadvantages. A CCD, for example, tends to be more sensitive to light than a CMOS. As a result, CCD imagers tend to have higher signal-to-noise ratios, masking background noise. On the other hand, CMOS chips are significantly less expensive to produce, resulting in an affordable camera.
Digicams (inexpensive, common digital point and shoot cameras), do not have a mirror, and therefore allow a live preview useful for framing a shot. In digicams, there are two separate light paths - one for the imaging sensor, and a separate optical tunnel used for framing. This optical tunnel, being a separate, detached system from the main lense, creates issues such as parallax errors on framing close subjects. Its view can be quite different from what the main lense sees.
With a DSLR, what is seen in the viewfinder is the exact same thing the imaging sensor will see. However, with the mirror in the way, a live preview is not possible. There is more discussion on the differences between various cameras in the “Which Camera Should I Buy?” section.
Almost all digicams use "Auto" metering, where a thryster (a photosensitive diode) measures the amount of light in the camera's general vicinity. Although this works most of the time, it can occasionally produce an inaccurate flash output. When the subject is close to the lense, or if the background is dark the flash typically overexposes the subject.
Through The Lense (TTL) metering was developed to counter this problem and can be found in more sophisticated cameras. This type of metering allows the camera to judge exposure from the actual scene framed.
However, with silicon based imaging sensors, the goal is to absorb light photons and convert them into a digital output. As a consequence, little to no light is reflected off the imaging surface. To work around this, camera manufacturers have come up with iTTL, dTTL, eTTL and the like.
Digital-TTL metering works slightly differently from traditional TTL metering. A quick preset pulse of light is sent out and the amount of light reflected back through the lense is measured. The camera then calculates if it is too much light or not enough and sends out the main strobe accordingly. With digicams, this can sometimes be seen as two discrete pulses of light. With DSLRs this all happens so quickly it is perceived as a single flash of light.
This is not to be confused with red-eye reduction, which sends out a series of pulses causing the pupils in the eye to constrict. These pulses must be timed long enough for the reflex reaction to occur. The result is a series of bright, blinding pulses that cause many subjects to blink or, for a few, to have an epileptic seizure. Since the red-eye reduction sequence is fired prior to the main flash, red-eye is effectively eliminated in the photograph and replaced by shut eyelids. My advice is to avoid red-eye reduction, as it can be dealt with in post processing.
A sync cord allows communication between the flash gun and the camera body. It greatly enhances the flexibility of the flash system and allows for creative lighting control.
Many newer systems have wireless TTL. Most of these work with infrared communications, where a series of binary pulses are sent out by the flash units that allow for communication between flashes. There needs to be at least two flash units for this type of communication - a master on the camera body itself (either the popup flash or a shoe mounted one), and a remote slave. The master can be configured so it does not emit an exposure flash, but will still send out a preflash sequence to communicate with its slaves. All this flashing occurs so quickly it is still perceived as a single flash.
With a corded TTL flash setup, we need only one off-camera flash unit. Additionally, since it is corded, the flash unit does not have to be able to see light pulses from a master flash. Wireless TTL requires the slave be able to see the strobe, either directly or reflected off of some surface, from the master.
Another form of flash control is via radio remote. These often do not support TTL, but have the advantage of being controlled through walls and other opaque surfaces. In my studio, for example, I use radio slaves to trigger my strobes. I measure the light output via a handheld incident light meter and adjust each strobe to fire a set amount of light. This way, there are no wires to trip over, and I do not have to worry if the angle of the triggering diode allows it to see the master fire.