Fuji S5 dominates dynamic range amongst DSLRs

An amazing performance, more than I would have guessed.

http://www.dpreview.com/reviews/FujifilmS5Pro/page18.asp

 

Good to see that Fuji are keeping up to date with their contributions
to DPReview.

;-)

 

But apparently not enough.
The S5 only managed a "recommended" rating, not the "highly recommended"
rating reserved for Nikon and Canon.
;-)

 

Isnt the S5 just a Nikon cloned camera that needs to use Nikon lens
too?

Whats of Fuji, the sensor?

M/P

 

yep - it is a nikon d200 body with fuji's sensor and a slower frame
rate.

 

Yes, and the rest of the electronics.

 

The main contribution of the Fuji S2, S3 and S5 cameras is the sensor
which has two sensors per pixel: a large "normal" and a much smaller
sensor for the highlights.

This extends the dynamic range in the highlights by a couple stops.

The drawback is that there is not a lot of real estate left over to
squeeze in higher pixel counts in the future.

Another poster laments the slow frame rate, but people concerned with
capturing the full dynamic range are more likely people interested in
overall image quality including taking the time for a serious composition.

For example, a product photographer dealing with deep shaddows, rich
colors and lighting and highlight relefections from the product would
take his time with each image ... frame rate is not a concern.

Cheers,
Alan.

 

Pretty stale news.

 

Maybe the basics on DR, but the review itself is brand new.

 

I looked over the page and there is a major flaw in the
test setup. The curves on the plots of EV on the horizontal
axis (e.g. 1st plot going from -6 to +5) illustrate
the problem well: The curves flatten out (becoming
horizontal) as you move to the left to darker
parts of the test strip. This is like the "toe"
in characteristic curves in film. The problem is, digital
camera sensors are linear and in the "standard curve"
applied by raw converters, no such "toe" is created.
Figure 8b shows a typical transfer curve for a digital
camera (in photographic stops, the blue points):
http://www.clarkvision.com/imagedetail/dynamicrange2
The test setup seems to have several stops of scattered
light in the system causing this problem. This is quite
common with setups like this (single test targets)
and depends not only on the target and sensor, but
the lens and even how much dust is in the air.

Roger

 

It's not reserved for Nikon & Canon, and not all of them get it
either.

 

DPReview's DR tests are of particular monolithic conversions, not of the
RAW data. While it is expected that the Fuji would have extra highlights
because of the extra, less sensitive sensor wells, this does nothing for
sensitivity on the shadow end. It is highly unlikely that it has 1.5 stops
more "footroom" in the shadows, compared to the 5D. The previous S3 had a
read noise of about 1.3 ADU at ISO 100, compared to the 5D's ~2 ADU, which
suggests a maximum differnce of about 1/2 stop in favor of the Fuji (the 1-
series Canons have about the same read noise as the Fuji S3 at ISO 100).
The D200 can't be the same, either; the D200 has a read noise of about 3.0
ADU at ISO 100.

I really wish that more testers would become RAW-literate and test the
cameras, not the conversions of choice, complicated by their styles.

--

 

I'll agree with that. I found dpreview is very reliable when it comes
to how a camera handles, its features etc, but when it comes to image
quality, it's not so reliable (and neither is any other site, at least
judging from other images I've found on the web and also from personal
experience with some of these cameras). But I guess a mean must be
found.

eg if I were to do reviews, then most people here would probably find
my "optimal" high-ISO shots very noisy indeed, as I much prefer high
frequency luminance noise to lower (by measured variance) but lower-
frequency (coarser) noise; in fact, I'm positively allergic to coarse
noise, whereas I actually like high-frequency noise. Also, if you
introduce raw conversion into the process, you run into all sorts of
problems: some converters have much better resolution of high-
frequency details (I mean Y details), others are bad at this but are
very good at colour adjustment, etc. Which one is more important
depends on what you do with the resulting image, how much time you're
willing to spend processing in photoshop (or whatever you're using),
your skill at colour manipulation, the hardware at your disposal, your
priorities etc. Impossible to satisfy in a review. Best to do your own
tests eg using raw files, if you ask me.

As far as the Fuji is concerned, I just read the review, and I also
find it rather hard to understand the extra range in the bottom (not
to mention the rather unsubtle noise reduction applied). It's probably
flare, as Roger says.