Steve:

Here is what you need to know:

Sped is measured at a standard degree of development. This standard
uses a 'speed point' which is a point on the H&D curve at which the
curve achieves a defined slope. The particular slope is chosen because
a minimum CONTRAST is necessary to achieve detail. That slope-point is
used for all films. You need such a point to be so defined to be able
to measure film speed, because the point moves a little with changes in
the degree of development. So, all films are developed the same degree
for speed testing. Is that clear?

Changing the degree of development makes very little actual difference
in the usable speed (maybe 1/3 stop, despite what you may have heard.
As development time is increased, contrast increases, and if
development is prolonged by 100%, contrast becomes very much higher
than normal. Grain also rapidly increases.

Just forget about pushing and N+1. It will not increase your speed at
all.

   They are not magic numbers and are only rules of thumb for _some
films_. When you develop for N+1 you are developing the film for one
paper grade more contrast than "normal" development, and a word about
what that is in a minute. The aim is to increase the contrast of the
straight line portion of the film characteristic, where most of the
image is recorded, so that a low contrast scene will print on "normal"
grade paper. The same thing can be accomplished by  printing on paper
that is one grade harder.
  When pushing the idea is to increase the contrast of the toe area of
the curve without much attention to what happens to the straight line
portion. This is because the "underexposed" image is all or mostly on
the toe. By increasing the contrast the image is easier to bring out.
Again, harder paper can be used but this tends to exagerate any
blemishes, etc, on the film.
  When development is increased both contrast and density are
increased, however, the increase in density is proportional to the
exposure so more exposed areas increase faster than low density areas,
that is why the contrast increases. If it is desired to maintain a
given maximum density then the exposure must be reduced. This is why
the "speed" is increased when developing for N+1. It isn't really
necessary to decrease the exposure if somewhat dense negatives can be
tollerated. For most films the difference in the speed index for N+1 or
N-1 is around 3/4 stop. For most conventional emulsion films the amount
of time differnce is around plus or minus 33% from the "normal" time.
For pushing somewhat more contrast increase is usually desired so the
film is usually developed around 50% more to get a full stop of speed
change.
  Now, what is "normal" contrast? To some extent its arbitary but it
has been chosen to so that the range of exposure in a average scene
will be recorded within a range of densities on the film where the
maximum density is around log 1.0. This results in film which will
print in a reasonable time and where the densities are within the range
where neither grain or loss of sharpness will become excessive. Both
grain and bluring of edges due to scattering of light in the emulsion
vary with the density becoming greater with increasing density. As far
as tone rendition most films have relatively long range of densities
over which the result will be good, provided the exposure is greater
than the minimum which will avoid recording to much of the image on the
toe. Lloyd Jones of Kodak Labs found that the minimum was where the
contrast of the toe was about 1/3rd that of the straight line portion
of the curve. For exposures of this amount the shadow detail is
adequate but many films will give good tone rendition with as much as
10 stops more exposure than this although they will become hard to
print.
  Paper grades are adjusted to match the "average" contrast of film
which meets the criterial above. The contrast used in the ISO speed
standard for B&W negative film (for still cameras) is about that which
will match a "normal" negative to Grade-2 ("normal") paper when exposed
using a diffused source. B&W films,which have an image composed of
particles of metallic silver, are subject to scattering of light by the
particles. As a result the _effective_ density depends on the type of
illuminant used in viewing or printing. The coarser the grain the
greater will be the scattering. When a specular source is used the
light comes from only one direction. Any light scattered by a particle
does not travel to the lens (or sensor in the case of a densitometer).
When a diffuse source is used the light strikes the image from all
angles so some light coming from an angle is scattered in a way that it
_does_ travel toward the lens. As a result the image looks denser when
illuminated by a specular source than when a  diffused source is used.
This is known as the Callier Effect or "Q" factor, after the man who
discovered it. Callier effect is important in predicting the actual
image contrast available in a printer or projector. Generally, diffuse
density is used in photography. This is partly because diffused sources
are most common and partly because they do not vary much from one to
another. True specular sources are very rarely used but sources which
are somewhere in between are quite common. The usual condenser enlarger
is such a source. It uses a source which is at least partly collimated
(all light coming from one direction) but the lamp is large and has a
diffused surface, so the resulting light is somewhere between a true
specular and true diffuse source. For a given  contrast on the
baseboard, that is a given ratio of measured density, a negative for a
diffusion enlarger needs to be about one paper grade more constrasty
than one for a condenser enlarger. The effect is linear so it can be
compensated for by changing the paper grade was well as the negative
contrast, the results are identical either way.
  The ISO speed method defines a range of log exposure and a range of
log density values to be produced by it. So, while it does not specify
a contrast directly it does so indirectly by specifying the two
quantities use for defining contrast, namely the rate of change of
density with exposure.
  If the film is developed to another contrast, for instance to match
a condenser enlarger, the speed will also change. Since the only
"official" speed is that measured for the contrast in the standard the
speeds for other degrees of development are called Exposure Index or
EI, these can be plugged into a light meter calculator but are not
officially ISO speeds.
   The amount of change in development for a given change in contrast
must be determined experimentally. Again, for conventional films it is
about 33% for one paper grade change but for tabular grain films it is
only about 25%. This, BTW, is one reason some people have problems with
tabular grain films; they need better control of time and temperature
and better uniformity of agitation.
  UC is right when he says that pushing film really does not increase
speed. A genuine increase in speed would result in a uniform increase
in all densities for a given exposure with no change in contrast. For
instance, compare an ISO 100 and ISO 400 film, both are developed to
exactly the same contrast but the ISO 400 film records a given exposure
with greater density.
  When film is underexposed the desired part of the image is probably
recorded mostly on the toe or low contrast part of the curve, so
pushing is only a way of incresing the contrst of this area to make the
image more printable. If there is any desired image further up the
curve it will become excessively contrasty.
 At some point, if the exposure is small enough, no image will be
recorded. There just will not be enough light to produce the changes in
the silver halide crystal to make it developable, so pushing is very
limited in what it can do.
  The ISO speed has virtually no safety factor. While most films have
enormous overexposure tollerance (latitude) they have no more than 2
stops at most of underexposure tollerance. even one stop will result in
some loss of shadow detail and 2 stops in a noticable degradation of
tone rendition, regardless of increased development.
  I didn't really mean to write a treatise in sensitometry but seem to
have done it.

Richard Knoppow
dickburk@ix.netcom.com

These "20%-33%" are rough estimates.  It also depends on your enlarger
type (condensor vs. diffuser), to some extent film technology ("classic"
vs. "t-grained" (Ilford Delta, Kodak T-Max).

Adequate shadow detail is Zone III; that is the first zone where you
have fully textured shadows.  Zone II in comparison, has *some* detail
in the shadows, but not a lot.

What defines "Zone III", "Zone II", etc., I've found to be purely
subjective.  Myself, I print very heavily, and my prints are very dark,
so my Zone V is really probably a Zone IV for other people.

When you get into zone's, your talking about the Zone System, and that
requires establishing your workflow from start to finish, so that you
can "previsualize" your images (when looking at a scene, you can tell
where your detailed shadow (Zone III) and detailed highlight (Zone VII)
are.  There, you can adjust for N +/- development.  E.g., if you have a
properly placed Zone III, but your most textured highlight is really in
Zone VI, then you have to perform N+1 development to "shift" your
textured highlight to Zone VII from Zone VI.  By contrast (no pun
intended), if you have your textured shadow in Zone III and textured
highlight in Zone VII, then you have N development.  Finally, if you're
shadow is in Zone III and highlight in Zone VIII, then you would do N-1
development.

Cheers, Patrick