Olympus-OM
[Top] [All Lists]

[OM] Re: Resolution cameras are able to deliver

Subject: [OM] Re: Resolution cameras are able to deliver
From: AG Schnozz <agschnozz@xxxxxxxxx>
Date: Tue, 10 May 2005 10:46:51 -0700 (PDT)
> I will admit that it's probably a stretch to equate number of
> bits per pixel with actual capability.  However, any camera
> that delivers a raw mode image is delivering a range that is
> at least one and possibly two 
> stop beyond what can be represented in a JPEG.  I call that
> 9-10 stops.

Well. Yes and no.  Let's talk film for a second. Ilford XP2 and
Kodak T400CN are chromogenic films with incredible exposure
latitude. It gets this latitude through little to no
straight-line section and mammoth toes and shoulders. The
shoulder of these films just go on and on and on.  Even at that,
you are pushing your luck to see 12 stops (or zones) of capture.
Several have claimed 14-16 stops with T400CN but I think that's
pretty much garbage.

Now, let's consider Velvia. Velvia has a nice straight-line
section with what appears to be a reverse-toe and a short
shoulder. 5-6 stops of dynamic range is what is usually
associated with Velvia.  Astia has a nice "normal" toe and adds
maybe a full stop or more to the bottom end. The mids and high
values remain about the same.

What about digital? Sensors are pretty much linear devices. 
Yes, there is a shoulder/toe to the response curves, but they
vary so much from manufacturer and type that you can't make any
generalizations. The in-camera processing chips take these known
response curves and make them linear before RAW encoding.

Unfortunately, brightness is not linear.  Therefore, when you
convert the data you crunch more bits into a small brightness
region. Shadows are represented with few bits, highlights a lot.
12-bits from the sensor don't go very far. The theoretical limit
of 12-bits may be around 12 stops, but you won't effectively get
it--even if the sensor was able to.

Next comes curve tweaking. Where film has toe/shoulder
characteristics which are part of the "native capture", these
have to be artifically created in the digital world AFTER
capture. This again requires a skewing of the limited number of
bits.

The end result is a theoretical "12-stop" latitude, but to make
it usable you'll effectively lose four stops due to adaption to
the analog equivelent.  JPEG is wrongfully accused of losing
dynamic range. It isn't JPEG per say, but how the camera or
encoder maps the bits during encoding.

> Not nearly so automatic but PW Pro has provided specific
> support for exposure composites for at least a couple of
> years.

And so has THE GIMP.  I use this functionality (in each) once in
a while.  It works, but it's very easy to overdue it.  But, hey,
nothing a bit of curves adjustment can't take care of.  And if
that doesn't fix it, convert it to sepia. "Old Photo" will fix
anything.

Back to dynamic range, I see the E-1 capture about the same
dynamic range as ASTIA.  During RAW conversion I can usually
milk a little more detail out of the highlights/shadows, but not
much.

AG


                
Discover Yahoo! 
Use Yahoo! to plan a weekend, have fun online and more. Check it out! 
http://discover.yahoo.com/

==============================================
List usage info:     http://www.zuikoholic.com
List nannies:        olympusadmin@xxxxxxxxxx
==============================================

<Prev in Thread] Current Thread [Next in Thread>
Sponsored by Tako
Impressum | Datenschutz