cgardner
Offline
• • • • •
Upload & Sell: Off
|
 p.1 #15 · Speedlight Tesing - Help! | |
For many it seems their first experience with external flash is to put one in the hot shoe in ETTL mode. Their expectation is that it will be just like shooting with ambient lighting. When they find it is not that simple they blame the equipment and abandon ETTL for the more fool-proof manual method instead of trying to understand the equipment. But if you bought a violin, would you expect to play Mozart flawlessly the first time you put it under your chin?
Ironically when using manual flash one quickly realizes the biggest difficulty with using flash: for any given flash output level the exposure is only correct at one distance from the flash. Once that basic physical reality is realized and applied to ETTL use, ETTL is much simpler conceptually: its a matter of helping the camera automatically identify the distance of the object you want correctly exposed based solely on the light reflected from the scene.
That is exactly what one is forced to to when using manual flash, either with a meter, or by testing or trail and error. Back when I used my manual Vivitars I benchmark tested them at various shooting distances with and without diffuser so I knew EXACTLY what the correct exposure would be at shooting distance 6, 8, 11 and 16 feet from a subject with either a single flash or two flashes in a 3:1 H:S ratio configuration. Why those shooting distances? Like the f/stop numbers each increase in distance (e.g. from 8 to 11 ft.) changes exposure by one f/stop (more or less). When using two flashes placing the off camera light at 4ft when shooting at 6ft. results in a 3:1 ratio. When moving back to 8ft. the camera aperture gets opened a stop an the off camera light moved out to 6ft. to maintain the ratio. Moving out to 11ft. for a wide shot requires opening the lens another stop to maintain correct exposure and the off camera light must move to 8ft. to maintain the same 3:1 ratio.
ETTL ratios , like any tool, is ideal for some jobs and doesn't perform very well for others. For example when shooting portraits in a fixed setting with multiple people because ETTL adjusts exposure based on reflectance of the scene, as judged from the light hitting the viewfinder, exposure will change from subject-to-subject resulting in the need to find the ideal level of FEC for each. Manual flash, using the method above is simpler and more consistent. I put my pair of identically diffused 580ex on 1/2 manual power (for faster recycling) and place the center if the off camera flash diffuser exactly 5.5ft from the subject's nose at an angle of 45 degrees. My arm span happens to be 5.5ft, so that measurement is consistent. I arrange the pose, feet-up, which takes all of about 30 sec., then take four measured steps backwards to 8ft. to shoot. 8ft is distance with renders a face with flattering perspective and it also puts my Master / fill flash where it needs to be (8ft vs 5.5ft) to create a perfect 3:1 lighting ratio. Because this set-up is consistent with regard to power and distance of the lights the exposure is always the same also, f/8. There is no need to meter or chimp. It is a total no-brainer technically, but I always take a test shot to make sure the idiot behind the camera hasn't done something stupid like leaving the camera on ISO 1600...
http://super.nova.org/TP/TowelGary.jpg
It doesn't matter what the person in front of the camera is wearing, or what complexion their skin is, the same exposure, f/8, is perfect because exposure is based on the brightest highlights (not the average of the scene) and all faces have the same tone eyes and teeth (more or less) which need to be rendered correctly.
So yes, manual is a much better mode that ETTL in any situation where lighting must be consistent, shot-to-shot. But it also requires what is most important to be closest to the flash. For example if a person is posed in a white shirt with their shoulder closer to the key light than the face either the detail in the shirt will need to be blown to correctly expose the face further away, because if the shoulder is correctly exposed (i.e. to preserve detail) the face will become under-exposed.
As mentioned earlier, perceptually we will by eye accept a loss of detail in areas like the shoulder if the face is correctly exposed because our brains expect the face to be correctly exposed and will tune out the shoulder immediately. But the "brain" in the camera in evaluative ETTL mode is programmed differently and will try to identify the brightest area of detail and expose it correctly. It does this very predictably and consistently because its a computerized machine. To the extend it would be perceived to fail when it correctly exposes the white shoulder at the expense of under-exposing the face is more a case of the user having unrealistic expectations based on their pre-conceived perceptual notion that "correct" exposure is when the face looks right. If the face isn't correctly exposed in their view ETTL has failed.
Making ETTL work in the situation described above, person wearing a white shirt, is actually quite simple. Change the pose. If the pose is changed so the white shirt is the same distance as the face the ETTL metering will again "nail" the exposure of the shirt, but this time also nail the face because it is the same distance to the flash. In other words to make ETTL work effectively you first need to understand how it interprets a scene and then change your methods to enable it to interpret the scene correctly. Stop being like a 2 year-old who throws a tantrum when something doesn't work "my way" and learn instead how to make it work.
Regardless of how flash is controlled, manually or ETTL, the key factor for correct exposure is composition: the need to place everything requiring correct exposure both closest to, and the same distance from the flash.
ETTL metering is most useful in dynamic situations where subject - flash distance is constantly changing or un-predictable. The problem with respect to getting correct ETTL exposures is the fact that in those situations it is nearly impossible to control what is in the foreground of the scene.
Let's say you are shooting two people on an empty dance floor 10ft away. With manual flash, assuming you've benchmarked, it would be simply a matter of dialing in the aperture needed for correct exposure at 10ft. If the dance floor is clear and ETTL metering would also be able to correctly meter the dancers 10 ft away. But now fill the dance floor with other people closer to the camera. Manual will still correctly expose the couple 10ft away, but at the expense of blowing out the people in the foreground. ETTL, relying on reflected light to make its decisions will correctly expose the foreground and under expose the couple.
Such are the dilemmas faced by photographers in nearly every candid (i.e. non static) flash situation? Should ETTL be indicted for failing to get the exposure correct? No, not any more than manual should be for blowing out the foreground. The solution in that situation is for the photographer, cognizant of the laws of physics and light, to realize the better strategy in that situation would be to bounce most of the light off the ceiling to evenly light the entire dance floor then place a second flash behind the dancers out of camera view to separate them with back-rim lighting.
The advantage of the Canon approach, vs using manual flash with radio triggers, is the ability to switch back and forth from ETTL to full manual remote control simply by pressing the MODE button on the master. ETTL is far more convenient in dynamic situations because it reacts and compensates to changing subject flash distances. When the main center of interest is closest to the flash and centered in the viewfinder ETTL does a good job of getting exposure correct. Once it is realized how un-important foreground detail skews ETTL exposure it is possible to either eliminate it with cropping, or anticipate how it will skew exposure and compensate with FEC. In situations where I find ETTL isn't producing the desired results I just press MODE, switch to M and dial in the power levels manually using the camera overexposure warning as a guide to when the highlights are correctly exposed.
Part of craft is knowing which tool is best for each job and its a poor craftsman who blames his tools for a job poorly executed. I have no problem at all using manual flash when it is the best tool for the job. But to completely dismiss the use of ETTL methods like the "Stobist" does on "ethical" grounds that manual is somehow superior in all situations is just myopic and foolish in my view. I have manual flashes and radio triggers but prefer using the Canon system because in it I have a single tool which can perform the same task two different ways, controlled remotely with the push of a button from the master flash on the camera; vs the manual only "Strobist" approach which locks one into working one way which, while more consistent, is far less convenient to use in most candid situations.
|
