One of the most important advantages of shooting in raw consists in the fact that
is possibile to compensate exposure during editing. As if you had a time machine you can,
just moving a virtual cursor, obtain exactly the same result you would have obtained
by a lower or a higher exposure. The only difference will be at the level of noise, which
in the first case will be even lower than usually
(
Noise reduction through overexposition and correction during development of the raw file),
higher in the latter case. But this capability has limits. Upwards (increase of the apparent
exposure during editing) the limit is subjective, given by the maximum amount of noise you
consider acceptable. On the contrary, downwards (reduction of the apparent exposure during editing) the limit is
objective, since it is given by the saturation point of the pixels. Once a pixel has reached saturation it
is no more possible to recreate the correct exposure, and attempting it during postprocessing will
bring to innatural colors in zone often with sharp edges.
The ability to recover overexposures ir very important because the more recovery space you have, the larger
will be the allowed error range in exposure.
In this page we compare the behavior of the Fujifilm s100fs with that of the PANASONIC Lumix dmc-G2K
(from now on simply Panasonic G2). Theoretically you should expect a much larger recovery capability in
the Panasonic, due to the fact that the surface of its sensor is four time larger than that of the Fuji
(4/3 sensor against 2/3 sensor),
while it has a number of pixels only slightly larger (12 Mp against 11 Mp). The single pixels should
therfore be much larger, with a proportionally higher saturation point.
In fact a pixels can be imagined as a bucket which collects light instead of water. The image is recreated measuring
the light collected by each bucket. In the case of a real bucket, once the bucket is full, it is not
possible any more to find out the real quantity of water fallen on it, because the exceeding water
gets lost without leaving trace of itself. The larger the bucket the more water can fall on it and measured properly before the saturation
point is reached. With a pixel it is similar, with light instead of water. This is the reason why small sensor camera produce easily burnt out pictures: if on the same
area there are more buckets each bucket has to be smaller and it will more rapidly become full (apart from
collecting less water anyway).
The expreiment is very easy: with each of the two cameras I have shot the same scene several times, each time
increasing the exposure of on stop. The overesposition has been later compensated during postprocessing.
Theoretically, if there wasn't the problem of saturation the final result should look the same all
the time for each given camera.
Note: Since the first picture was in reality underexposed, the exposure compensation starts from the third
one of each series.
Crop Fuji s100fs
|
Crop Panasonic G2
|
|
|
ISO: 100; 1/400; 5.6; compensation 0
|
ISO: 100; 1/400; 5.8; compensation 0
|
|
|
ISO: 100; 1/200; 5.6; compensation 0
|
ISO: 100; 1/200; 5.8; compensation 0
|
|
|
ISO: 100; 1/100; 5.6; compensation -1
|
ISO: 100; 1/100; 5.8; compensation -1
|
|
|
ISO: 100; 1/50; 5.6; compensation -2
|
ISO: 100; 1/50; 5.8; compensation -2
|
|
|
ISO: 100; 1/25; 5.6; compensation -3
|
ISO: 100; 1/25; 5.8; compensation -3
|
It is possible to note that, contrary to what one might expect, it is the Fuji s100fs
which has the sensor with the larger resilience. In the case of the Panasonic G2 already
the picture overexposed by one stop has problems: the colors are faded and there are burnt areas with
sharp edges. The picture is practically unusable! The homologous picture made with the Fuji s100fs
on the contrary is still almost perfect. It also has some small areas which are probably saturated, but
the edges are faded, so that you almost don't notice them. Furthermore the colors are still correct.
In the case of the Fuji s100fs even the picture with -2 compensation is still usable, at least
as emergency solution. In some areas it is saturated, but the colors are still not too bad. On the
other side the Panasonic G2 picture with -2 compensation is totally unusable!
The reason for this result is unknown to us. This experiments proves without doubt the high
quality of the Fuji sensor, which in spite of its small dimensions offers a much larger error margin
than the couterpart by Panasonic. This last one practically offers only a half stop margin, against the almost two
of the Fuji. Really
not much! It means that when shooting with the Panasonic you have to pay much more attention
to exposition, even in raw, than with the Fuji s100fs, because the sensor reaches the saturation point
earlier and in a less progressive way.
In fact, when shooting with the external flash mounted during parties I had to eliminate more than
one picture when shooting with the Panasonic G2, whereas this almost never happens when I use the Fuji s100fs.