Objective - To
compare the D300 to the D200 using available tungsten lighting.
Original (but
down-sized) Photo
This chart compares the Nikon
D300
with the Nikon
D200. All the photographs were
taken at the same time under fairly
weak tungsten lights (four 60
watt recessed ceiling lights in my office). The following parameters
were strictly followed:
1- All used the
Nikon f2.8 17-35
at 35mm (53mm FOV as compared to a full 35mm format). This is an
excellent lens made for full frame photography which produces
outstanding results using a cropped sensor.
2- Carmagne carbon fiber
tripod with Kirk modified Giotto ball head - from the same exact
same position.
3- All shots were taken in
raw mode and converted to 16 bit
using Adobe Camera Raw version 4.3
and "as shot".
4- All shots were converted
with sharpening and noise reduction
OFF.
5- All shots had a gentle
unsharp mask applied as tiffs to reduce the gaussian blur caused by
the AA filters on the sensors. (I used
.4 Pixels at 150% and 0 Threshold.
6- The first thing we noticed
was how much better the default ACR settings worked with the D300
over the D200 (which looked flat with somewhat off color
balance). An easy and constant correction was then applied by using
PS CS3 Auto Levels on each and every photograph. This made a slight
improvement in the D300 photos and a large improvement on the D200
photos.
7- Exposures ran very, very
close between the two cameras. Less than 1/8 a stop in all cases.
That was surprising! The only exception was the D300 taking 4
seconds at -1 (supposedly ISO 100) verses 2.5 seconds with the D200.
This would put the D300's -1 pretty
close to an E.I. (Exposure Index) of 60.
8- The testing involved
both myself and a fellow pro, Nick Berezenko (of Arizona Highways
fame).
9- Click on each photo to get
a large but cropped 100% image. Please feel free to right click and
download. Also note, that because these files were taken from a
layered PS CS3 file, any attempt to pull EXIF will be futile as only
the base file will display the correct date (6400 data in the D300
case and 3200 data with the D200). I chose to use layers so I could
have precise control with cropping and levels.
Our findings:
1- It looks like the
D300 has about a 1.5 stop advantage both in
noise and rendered detail. We both
came to the same conclusion. By viewing precisely stacked photos and
using the levels "eye" in PS CS3 it was easy to instantly A-B back
and forth between any two layers and see differences. We looked into
shadows, various colors, fine print, etc. You can do this for
yourself by right clicking and saving - and then opening in
Photoshop. Make sure "snap" is enabled and drop in as aligned
layers.
Caveat: In order to compare
files from two different size formats we had to resample the D200
file up to the same size as the D300. Then, to be completly
objective, we did it the opposite way. We downsized the D300 file to
the D200 size. Both methods used PS CS3 resampling with bicubic
rendering. We could see NO difference using
either method.
We have some
outside shots using this same testing method (I will add these
later). The results are pretty much the same (in differences) except
that because more light was available we were able to get much
better prints - even with ISO 6400 using the D300. We did need to
adjust curves to get the best prints. A full size print on 8.5 x 11
inch luster looked VERY useable in outdoor lighting. Keep in mind,
that's with proper post processing!
