information
without blowing it
5
About raw les
6
Saving the raw le
6
Saving the JPG le
6
Pros and cons
7
Reasons to shoot JPG
7
Reasons to shoot raw
8
Raw converters
10 Reading histograms
10 The 21st century light meter
11 About color balance
12 Noise reduction
12 Sharpening
12 Its in the cards
13 A matter of black and white
13 Conclusion
Making the Transition

from Film to Digital
In the beginning
Photography became a reality in the 1840s. During this time, images were recorded on
lm that used particles of silver salts embedded in a physical substrate, such as acetate
or gelatin. The grains of silver turned dark when exposed to light, and then a chemical
xer made that change more or less permanent. Cameras remained pretty much the
same over the years with features such as a lens, a light-tight chamber to hold the lm,
and an aperture and shutter mechanism to control exposure.
But the early 1990s brought a dramatic change. Instead of using grains of silver
embedded in gelatin, digital photography uses silicon to record images as numbers.
Computers process the images, rather than optical enlargers and tanks of often toxic
chemicals. Chemically-developed wet printing processes have given way to prints made
with inkjet printers, which squirt microscopic droplets of ink onto paper to create
photographs.
Snafellnesjokull Glacier Remnant. Iceland, June 2004
This image was taken with a Contax 645 camera with Kodak DCS Proback and
Zeiss Sonnar 210mm F/4 lens. With appropriate processing, silicon imaging chips
can capture even the extremes of tonality and shading.
While digital technology has revolutionalized photography, the process of adopting
these new technologies is an ongoing one. In these early years of the 21st century, both
professionals and amateurs use traditional and digital photographic technologies. This
coexistence of technologies is comfortable, yet contentious. Silicon is gradually replac-
ing lm, and many photographers are still uneasy with the transition. At the same
time, there are passionate ne art photographers, as well as professional photographers,
for whom traditional chemical-based photography is unfamiliar. Digital photography,
digital image processing, and inkjet printing are the technologies they know, and no
historical paradigms exist as points of reference.
But for many photographers who are versed in the ways of lm, the transition is a work
in process. Therefore, this article is intended for both professional as well as amateur
photographers. Making the Transition
from Film to Digital
2
Making the transition
Photographers who learned their craft by using lm-based cameras can make the transition
to shooting digitally with ease. But to fully master this new medium, they must ascend some
portion of the learning curve. For starters, photographers must understand how silicon diers
from lm. In some ways, silicon responds to light similarly to the way that transparency lm
does, but there are important dierences. Likewise, the dierences between so-called raw les
and camera-generated JPGs and TIFFs need to be understood, as does the role of color balance
settings.
The image quality equation, when seen as a function of chip size and pixel count, is complex. The
size of print that can be made (and also the amount of cropping that can be done) depends on the
number of pixels that an image contains. Yet smaller pixels (such as those found on digicams) are
generally noisier than the larger pixels generated by Digital Single Lens Reex (DSLR) cameras
that use larger chips. This statement doesnt always hold true though, particularly when making
moderate-sized (A4 or smaller) prints. Each photographers level of critical acceptance also plays
a role in deciding whats acceptable in terms of image quality.
Grain versus pixels
Individual pixels, or photo sites, are incredibly small. Depending on the particular chip, these
pixels can be as small as 2 or 3 microns (thousandths of a millimeter) in size. A pixel this size is
smaller than a single particle of silver on ne-grain lm. And, unlike lm grain or dye particles,
which are spaced irregularly, pixels are organized in neat rows and columns.
As long as the pixels arent too small, they can capture enough photons to produce extremely
clean and noise-free images. The tiny pixels found in digicams begin to get noisy as the ISO
setting increases. The signal-to-noise ratio (an image-forming signal and a nonimage forming
noise) increases when the gain on the chips analog output stage increases. Chips have an inher-
ent sensitivity, usually its lowest published speed. You can achieve higher ISOs simply by increas-
ing the ISO setting on your camera.
With lm, you can achieve higher sensitivity and increased graininess by increasing the size of
the grain particles, for example, by using a high-speed (or high ISO) lm. Similarly, digital images
printed from a camera that is set to a high ISO are noisier.
But, the photographers ability to instantly change the cameras ISO setting provides digital
imaging with one of its greatest advantages. A photographer can be shooting images at ISO 64
one moment and then increase the camera to ISO 400, or even higher for the next image. With
lm, the photographer would need two separate cameras, or at least would need to change lm,
to take pictures at a higher ISO. Making the Transition
from Film to Digital
3
This image was taken with a Canon 20D at ISO 1600.
Noise reduction was not applied to the image.
Current DSLR cameras are able to shoot very clean images at what used to be considered high
ISO settings. Some DSLRs can shoot at ISO 800 and produce essentially noise-free images. Even
ISO 1600 is very usable. For some photographers, especially those who use slower zoom lenses
rather than fast primes, shooting at high ISO settings is a major advantage of DSLR cameras.
On the other hand, digicams with their smaller photo sites are prone to fairly high noise at ISO
settings above ISO 100. For this reason, digicams with fast zoom lenses are preferred over those
with smaller, though slower, optics.
Exposure considerations
Shooting with a digital camera is similar to shooting transparency lm in regards to exposure
because both are direct-positive processes. As with transparency lm, photographers need to
avoid overexposure. When a part of an image is overexposed on reversal (transparency) lm, it
becomes clear, but a clear piece of lm contains no useful information. Similarly, when a digital
image is completely overexposed, the numeric value becomes 255 (in 8-bit mode) and the exposure
contains no image data. Its simply pure white.
Shadow areas are another matter. Shadows on slide lm contain information to a certain point.
Of course black is black, so if theres nothing on the slide lm, theres nothing to be extracted.
But it is often possible to pull useful information out of the shadow areas of lm.
Finding useful information in the shadow areas is also possible with digital images, but to a
lesser extent. Shadow areas contain less information than midtones and highlights, and less
information means a lower signal-to-noise ratio. In other words, the shadow areas can be noisier
than elsewhere in the image.
To understand this comparison better, and to gain some insight into how image data is recorded
as numbers, we need to look at the issue of bit depth. Making the Transition
from Film to Digital
4
This wont hurt a bit
When photographers who are new to the concepts of digital image processing encounter the
phrase bit depth, their eyes glaze over. They know that it has something to do with image
quality, and that more bits are somehow better, but thats about all they know. Lets look at this
relatively simple, but often misunderstood topic and see if we can make sense of it.
We should get some basics out of the way rst. A bit is the smallest unit of data. It can be 1 or 0,
black or white, and on or o. Eight bits make up a byte. A byte (or eight bits) can therefore
represent 256 dierent states, or 2
8
.
Most of the digital world operates with 8-bit images. Inkjet printers and monitors typically use
eight bits of information to create most images that you see.
8 bits or 24 bits; which is it?
At times, conversations become confusing when we speak of an 8-bit image because were
usually talking about an image that consists of three colors, which is also referred to as a 24-bit
image (3 � 8).
High-bit images
Instead of using just 8 bits to represent a single color, we can use 12 or 16 bits. A 16-bit image can
handle 65,536 discrete levels of information, instead of the 256 levels that an 8-bit image can.
The increased degree of subtlety that these higher bits make available is dramatic. And, just as
an 8-bit image is actually a 24-bit image when were dealing with color, a 16-bit image is actually
called a 48-bit image (16 � 3) in terms of a three-color composite. A 48-bit image is capable of
billions of colors.
All imaging chips in digital cameras are capable of producing 24-bit images (8 � 3). Some digital
cameras can capture 36-bit images (12 � 3) capture, and some high-end scanners and cameras
can even capture 48-bit images (16 � 3) images.
Why would you want to use a high-bit image?
The key advantage of using a high-bit image is that when you apply a Levels or Curves adjust-
ment in Adobe� Photoshop� software, you adjust much more data than if you adjusted a low-bit
image. If you change the tonal range of a high-bit le that has 65,536 levels (compared to a low-bit
le w