Olympus-OM
[Top] [All Lists]

[OM] More Dynamic Range

Subject: [OM] More Dynamic Range
From: Winsor Crosby <wincros@xxxxxxxxxxx>
Date: Wed, 28 Sep 2005 10:52:19 -0700
Since this was a recent topic I thought you might like what Thom  
Hogan had to say about it. He does reviews, writes user manuals for  
Nikon mount cameras and is a techie photographer in his own right.

 > Most of what I read was confusing depending on the sources, some
 > would pretend that a good DSLR (Canon or Nikon) has at best 3.5 EV
 > of DR, others say around 6.5 to 7.1 where for film photography (not
 > slides), the values would vary from 4.0 to 8.9 EV depending on the
 > source you search or read. Slides would rotate pretty much around
 > the same values as a Good DSLR.
 >
 > I would like to be able to get a decent answer to this.

Ah, into the crevasse with you! ;~)

DR has such a range of issues that it becomes a deep crevasse for  
most people (i.e., you either fall in unintentionally or you just  
stare down into it wondering what's down there).

First, an outdoor scene can have more DR than you can see, you can  
see more DR than you can capture, and you can capture more DR than  
you can print. This is the essence of Ansel Adams' work with the Zone  
System--trying to place values you see into values you can achieve in  
a print. It took Adams a lifetime and and many books to acheive some  
sense of completion on this. No post on a forum is going to do that  
justice.

Your question deals mostly with the capture stage. For film, we have  
Kodak, Adams, and many others who did a great deal of research and  
experimentation to come up with methodology to assess what film could  
capture. And even then there's minor to modest disagreement.  
Moreover, film's response has toe and shoulder slopes (e.g., non- 
linear capture), and there's disagreement over where in the toe and  
shoulder you hit the law of diminishing returns. That said, the usual  
values that get tossed about are 5-6 stops for slide film, 8+ stops  
for negative film, and sometimes as much as 10+ stops for BW negative  
film.

Digital is another story. We have several key problems that have no  
official methodology or substantive published research to go on. The  
first is the bit problem (mentioned elsewhere in these threads).  
Shadows just don't get as much differentiation of subtle changes as  
do highlights. Thus, you have to come to some conclusion about  
whether a wrong rounded value is actually "capturing the scene" or  
not. Related to this is noise. Noise also lives in the shadows, and  
varies with a heck of a lot of variables (temperature, sensor, ADC,  
ISO amplification, length of exposure, etc.). And we don't have a  
standard way of measuring it, either!

So, the questions quickly become: is a value of 16, 32, 29 actually  
an accurate "capture" of a value in a scene or an approximation? And  
is DR defined as "accurate capture" or just "capture?" Let me put it  
another way: if I encounter a scene that has four stops of DR in it  
and push my exposure so that those four stops are captured by the top  
four stops my digital camera is capable of grabbing (e.g., expose for  
the highlights), I have great confidence that the tonal values I  
record are accurate representations of the tonal values in the scene.  
If I put that same scene in the bottom four stops of my digital  
capture, I no longer have that confidence, as noise and bit reduction  
take their toll on accuracy.

So what's the DR number for digital? Personally, I run tests on every  
camera I own looking for two things: noise levels above a certain  
threshold, and visual compression of tonal ramps in shadows (what I  
call "shadow mush"). The first can be measured objectively, the  
second is purely subjective. For example, my D1x measured a bit over  
7 stops in the noise test at ISO 125, but visually, I often saw mush  
in the bottom two stops. So is it a 5-stop or 7-stop camera? (Or  
neither. You may have a different methodology.)





Winsor
Long Beach, California, USA






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

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