Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: Photo Image out of focus

In my last message the words "left" and "right" were
in wrong places. The inconsequential correction
(using capital letters) is shown below.
Ludwik Kowalski

In browsing through the article suggested by Brian (his 3rd
URL below) I concluded that the H-D curve is the "darkness
versus the exposure" relation which is the most important
characteristic of any photographic emulsion. Is the loss of
information due to the inappropriate film equivalent to the
loss of information due to defocusing? I think so.

Consider emulsions A (high contrast) and B (low contrast).
Let me idealize their H-D curves as shown below.

Black ********
^ *
| *
| *
| *
| *
| *
Transparent ******** ---------------> exposure

A - records minimum (transparent on the negative) when the
exposure is below 35 units and records maximum (black
on the negative) when the exposure is above 45 units. The
transition from transparent to black (between 35 and 45) is
gradual (for example, straight line).

B - records minimum (transparent on the negative) when the
exposure is below 5 units and records maximum (black
on the negative) when the exposure is above 75 units. The
transition from transparent to black (between 5 and 75) is
gradual (for example, straight line). In other words the
film B has a much lower contrast than the film A.

The photographed object is a screen displaying the gray
scale. The screen is white along the left edge and black
along the rigth edge. The darkness of that screen changes
gradually from left to right. The screen is photographed
using the high contrast film A. The image shows a broad
RIGHT margin (uniformly transparent) and a broad LEFT
margin (uniformly black). These margins did not exist in
the object. The transition from transparent to black is
gradual but only along a narrow central strip of the film.

Suppose transparent letters were present on the original
object. Some of these letters would not be seen in the photo
taken with the film A but they would be seen if the photo
were taken (under identical conditions) with the film B.

My point is that the information lost due to the limitation
of the film is not recoverable; it is like information lost due
to very poor focusing. A photo of the first film using the
second film will not help us to recover lost letters. The
"all of the king's horses and all of the king's man"
comparison is applicable. What is lost is lost.
Ludwik Kowalski

brian whatcott wrote:

At 09:42 6/24/01 -0400, Ludwik wrote:

I can think of some helpful conditions - an emulsion latitude
capable of capturing the intensity contributions of several
source contributions - I seem to recall the appropriate measure
of this property is the Hurter-Driffield curve. The fact that
it *is* a curve is unhelpful.

1) What is the Hurter-Driffield curve?
2) The first sentence would be more clear to me if the word
"capturing" were replaced by "distinguishing". Is it not true
that a camera film always captures contributions from several
sources? I must be missing something important; what is it?
Ludwik Kowalski

Here are some Google hits on Hurter-Driffield, to give a feel for the
contexts in which the H-D curve is mentioned. The third URL is
a product description written from a workmanlike physics viewpoint,
and includes a reference to Hurter-Driffield's original paper.

Remote Sensing Tutorial Page 10-2

CS563: Recovering High Dynamic Range Radiance Maps From Photographs

photolithography paper in SPIE

To answer Ludwik's second point - the emulsion misses intensity
contributions if they are overlapped too numerously to place the
resulting emulsion density (or other detector response) in the
linear part of the H-D curve.

brian whatcott <> Altus OK