You should note however that a meter is only accurate if it has been compensated to the camera being used.
Even in manual mode the ISO of the camera controls its sensitivity. It is rare that camera ISO will match that of a meter. As a result if the f/8 indicted by the meter when correct metering technique is used for exposure (i.e. point dome at camera - fire all lights) the resulting file is not correctly exposed. Those not aware of the need to calibrate meter to camera will apply a \"Kentucky Windage\" solution to the problem, changing the way the meter is pointed until the direction the meter reading is taken (e.g. pointing it at the key light) produces accurate exposure.
The need to compensate meter to camera isn\'t unique to digital. The true speed of film varies from batch-to-batch and is seldom exactly what is marked on the box. One of the first tests for setting up the Zone System is to find the ISO speed which produces the desired shadow density on the negative. Since exposure of digital is most critical in the highlights the best way to compensate the meter is to shoot a bracketed series of exposures of a test target containing a textured highlight. Visual inspection of the texture in the towel on the PC screen at 100% in the RAW file will reveal very precisely the point at which overexposure begins to occur:
With my L-358 / Canon 20D combination I find that the meter reading from the L-358 produces a file in the camera which is about 1/3 stop overexposed. That occurs because the true ISO of the camera when set to ISO 100 is actually around 125. The bracketed exposure at f/6.3 shown above, produced a more accurate rendering of the test target that the nominal meter reading of f/5.6. On the L-358 pressing both ISO speed buttons allows a 1/3 stop Exposure Compensation Override to be entered, changing the display to f/6.3 (I set mine to 1/3 stops like the camera). Meter now is compensated to camera and meter readings, taken per manf. recommended technique (i.e. point meter at camera) are accurate over any range of lighting ratios. The compensation doesn\'t change the factory calibration of the meter, it just applies an override in the same way EC in Av mode overrides the default zero EC guess of the camera metering.
I know there are people who swear by any number of non standard metering techniques, and in practical terms just about any of them will work if used consistency. But I view a meter like I do a hammer: a tool designed in a certain way to perform specific ways which produces its best results when used as designed. The reason an incident meter has a dome instead of a flat surface is to average the light from all directions in front of the subject, simulating how it will reflect back into the camera. Sekonic who makes the meter tell us the correct technique is to point the meter at the lens to take the exposure reading. The reason the dome retracts is because that is how Sekonic designed the meter to produce accurate reading of light levels in absolute terms of foot-candles / lux / EV values. Best results with a hammer are obtained by hitting the nail with the heavy end. Best results with an incident meter are obtained over a range of lighting ratios by pointing it correctly at the camera to measure exposure AND compensating it to the true ISO of the recording medium.
It is also worth noting the portrait ratio convention of H:S where S is a constant value of 1 is based on the assumption even \"neutral\" fill is being used to illuminate both sides of the face the same. For example, two lights, equal in all respects, placed at the same distance with fill over the camera and key light off axis produce a 2:1 ratio per the portrait ratio convention:
H:S
1:1 even fill
1:0 equal but overlapping key light
===
2:1 H:S ratio -- 2x more light reflect from the shadows
If key light is 1-stop brighter (2x fill) the ratio becomes 3:1
H:S
1:1 even fill
2:0 overlapping key light 2x brighter than fill
===
3:1 H:S ratio -- 3x more light reflect from the shadows
So the portrait ratios inferred from making incident readings are actually expressing the amount of light the two lights, which are assumed to be key overlapping even fill, will reflect.
You should note however that a meter is only accurate if it has been compensated to the camera being used.
Even in manual mode the ISO of the camera controls its sensitivity. It is rare that camera ISO will match that of a meter. As a result if the f/8 indicted by the meter when correct metering technique is used for exposure (i.e. point dome at camera - fire all lights) the resulting file is not correctly exposed. Those not aware of the need to calibrate meter to camera will apply a \"Kentucky Windage\" solution to the problem, changing the way the meter is pointed until the direction of the meter is taken (e.g. pointing it at the key light) produces accurate exposure.
The need to compensate meter to camera isn\'t unique to digital. The true speed of film varies from batch-to-batch and is seldom exactly what is marked on the box. One of the first tests for setting up the Zone System is to find the ISO speed which produces the desired shadow density on the negative. Since exposure of digital is most critical in the highlights the best way to compensate the meter is to shoot a bracketed series of exposures of a test target containing a textured highlight. Visual inspection of the texture in the towel on the PC screen at 100% in the RAW file will reveal very precisely the point at which overexposure begins to occur:
With my L-358 / Canon 20D combination I find that the meter reading from the L-358 produces a file in the camera which is about 1/3 stop overexposed. That occurs because the true ISO of the camera when set to ISO 100 is actually around 125. The bracketed exposure at f/6.3 shown above, produced a more accurate rendering of the test target that the nominal meter reading of f/5.6. On the L-358 pressing both ISO speed buttons allows a 1/3 stop Exposure Compensation Override to be entered, changing the display to f/6.3 (I set mine to 1/3 stops like the camera). Meter now is compensated to camera and meter readings, taken per manf. recommended technique (i.e. point meter at camera) are accurate over any range of lighting ratios. The compensation doesn\'t change the factory calibration of the meter, it just applies an override in the same way EC in Av mode overrides the default zero EC guess of the camera metering.
I know there are people who swear by any number of non standard metering techniques, and in practical terms just about any of them will work if used consistency. But I view a meter like I do a hammer: a tool designed in a certain way to perform specific ways which produces its best results when used as designed. The reason an incident meter has a dome instead of a flat surface is to average the light from all directions in front of the subject, simulating how it will reflect back into the camera. Sekonic who makes the meter tell us the correct technique is to point the meter at the lens to take the exposure reading. The reason the dome retracts is because that is how Sekonic designed the meter to produce accurate reading of light levels in absolute terms of foot-candles / lux / EV values. Best results with a hammer are obtained by hitting the nail with the heavy end. Best results with an incident meter are obtained over a range of lighting ratios by pointing it correctly at the camera to measure exposure AND compensating it to the true ISO of the recording medium.
Nov 15, 2008 at 11:23 AM
Previous versions of cgardner's message #6384785 « who still uses a light meter? »
