apersson850 wrote in post #15029131
It's actually so that cameras with E-TTL II (all newer digital cameras) measure the difference between ambient and flash light, when the preflash is fired. The different zones in the exposure meter are used for this, so cameras with 63 zone metering (7D) does this better than the previous 35 zone metering (5D Mark II).
If a zone peaks heavily when the flash is lit, then it's disregarded as probably being a mirror or something similar.
On the other hand, if there's no different with and without flash, the zone is disregarded as obviously being outside the range of the flash.
The the zones that are in between, i.e. made brighter by the flash, are evaluated, with emphasis on zone(s) that coincide with AF points that are in focus.
Ambient metering is also a bit adjusted, since if the camera's setting is such that the camera can achieve a standard exposure without the flash, then that standard exposure is reduced by about one stop, to allow for adding flash light without overexposing things that otherwise would be at standard exposure already before the flash fired. If you use P-mode indoors in the evening, it will not allow you to get up to standard exposure of the background (lest you have sky-high ISO), so this reduction isn't necessary. The flash will provide the brunt of the light anyway.
An eloquent description of flash logic. Canon flash does have programming called 'NEVEC' which does alter exposure settings a bit, but NEVEC does not always hold true. Most of the time (except for academic discussion) I ignore the effects of NEVEC. Unfortunately the text in blue cannot be proven to be true, as this series shows...
Test conditions:- Photos 1, 3, and 5 have no flash, as baseline exposures; 2, 4, 6 all have direct ETTL flash, to judge the effect of addition of flash to baseline.
- The camera was in evaluative ambient metering, with the focus point on top of the Colorchecker target, so ambient metering would be biased to that focus location.
- ETTL was set so that evaluative flash metering (factory standard) was set; CFn for Av shutter speed was set to force 1/250 for flash shots.
- Post processing settings were identical for all shots, with the exception of White Balance, due to some shots being made with incandescent overhead downlights; WB was post processed with eyedropper to render 4th grey sample as neutral.
Observations:- Eyedropper values for photos 1,3 (no flash) RGB=52%
Eydropper values for photos 2,4 (flash) RGB=42%
Eyedropper value for photo 5 (no flash) RGB=45%
Eyedropper value for photo 6 (flash) RGB=61%
- Photos 5 and 6 only differ in ambient settings by 1/3EV (f/3.5 vs. f/4), and that difference can be easily explained that P mode selects 1/60 f/4 as default.
We also see that photos 4 and 5 differ in ambient settings by -4EV (1/8 vs. 1/60) because P mode selects 1/60 as default flash shutter speed.
- Photos 5 and 6 differ in target brightness by about 0.78EV of exposure (flash shot 6 is 0.78EV brighter than no-flash shot 5)
Conclusions:In adding flash, (photo 1 to 2, and photo 3 to 4, and photo 5 to 6) there is
no apparent -1EV reduction in amount of exposure given for ambient, (contrary to blue text) to account for the addition of flash!
If I stand in a situation where 1/100 f/2.8 was needed in Av for ambient-only exposure (ISO1600), when I turn on the flash the camera says 1/160 f/2.8! The effects of NEVEC are in action.
If I stand in a situation where 1/400 f/2.8 was needed in Av for ambient-only exposure, when I turn on the flash the camera still says 1/250 f/2.8!