On 26/03/2012 2:05 a.m., Robert Coe wrote:
On Sun, 25 Mar 2012 21:20:08 +1300, wrote:
: On 25/03/2012 8:01 p.m., Ray Fischer wrote:
: wrote:
: On 25/03/2012 4:36 p.m., Robert Coe wrote:
: On Sat, 24 Mar 2012 20:51:55 -0500, wrote:
: : Robert wrote in
: : :
: :
: : On Fri, 23 Mar 2012 14:10:14 -0700 (PDT),
: : wrote:
: :: Rendering of dark areas. This is a shot from Dpreview's new gallery
: :: of pre-production test images from the Canon 5DIII. 3200 ISO. I
: :: raised the illumination level 25% beyond theirs. Look at the black
: :: background. Film doesn't produce that ugly, mottled effect. It
: :: simply goes black, which means all the silver/dye simply washed away
: :: leaving the base of the film.
: :: (Typo corrected at no additional charge)
: :
: : Your favorite photo editor will let you blacken the shadows of your
: : digital images as much as your heart desires. And blow out the
: : highlights too, if that's what you want (e.g., if that's what you were
: : trying to achieve by jacking up the illumination level).
: :
: : Bob
: :
: :
: : The highlights held fine. I can pretty much guarantee the background
: : wasn't pitch black to human eyes in that photo shoot.
:
: Then how does that square with your assertion (see above) that film does a
: better job of rendering dark areas?
:
: Why use a Canon DSLR in order to make a proclamation about shadow detail
: recovery?
: Canon's latest FF camera has two stops less dynamic range than the
: competition at base ISO due to read noise from the sensor.
:
: And where did you get the supposed information? Did you hack into
: Canon's proprietary information? Did you disassemble a camera and
: measure the read noise from the sensor?
:
: http://home.comcast.net/~nikond70/Charts/PDR.htm

"Check on the camera model in list ..." What list? Is there some trick to
seeing it? Or something wrong with my browser?

: and
: http://www.sensorgen.info/
: I'm confident DXOMark will confirm the above in their own tests.

The text suggests that DXOMark doesn't accept the validity of the "figures of
merit" that underlie your interpretation of the data. Are you expecting them
to change their minds or something?

I'd say Ray is right to be skeptical.

That's not what the text suggests. It states that DXOMark "have decided
to present that information so as not to give the three major 'figures
of merit' which are commonly used by designers of imaging equipment".

Anyway, you can "re-order" DXOMark results, including selecting to
compare cameras of the same format - as it's pointless looking at MF
digital sensor performance if you want or need a u4/3 format camera.

The usual rebuttal of conclusions which may be drawn from the data is
"so what - if you expose correctly, then you never need to adjust to the
extent in post-processing that Canon's DR performance is an issue".
That might be true for some people, but doesn't exclude the reality that
some people want or need as much DR as possible, and that for sensor
technology, Canon is clearly falling behind.

I've seen arguments that Canon's sensor fab facilities are out of date,
and lack the precision to match Sony's column parallel AD converters at
FX sensor size. Whether this is true or not I don't know - nor whether
if it is true, the cost for a new sensor fab could be recovered by
future sales of FX sized sensors.
Perhaps that's why the 5D3 is so expensive for what you get.