Image Processing Reference
to compress it. Currently we compress integer representations of pixel values. Floating-point
pixel values will require some rethinking of how the compression algorithms operate.
Some scientists hope to go beyond high-dynamic range imaging by developing remote
sensing systems. This work is called hyper spectral imaging and is designed to collect light
from ultraviolet to the infrared end of the spectrum. While this isn't yet producing mov-
ing images, it will one day in the future.
Other scientists are looking at building imaging systems that operate at very high
resolutions or at improbably rapid frame rates. The Cordin company in Utah manufac-
tures a camera that can take 200 million frames a second. This can photograph an explo-
sion as it happens.
So far no one is attempting to combine all of these into a supercamera—unless you
know better, of course.
Summary: All Roads Lead to Rome
In this chapter we have looked at digital-imaging systems. Ultimately, whether our source
footage originated on film or video, it is going to end up in a digital-imaging system on
the way to being compressed.
Shooting some movie film, migrating it through a telecine box to a video format, and
then ingesting it into a digital-imaging system is probably going to introduce artifacts
from all three disciplines. Eliminating some of those stages by scanning the film digitally
is obviously beneficial.
Film will have contributed grain; dust, scratches, and gate weave; and faded colors.
You will then have lost significant amounts of detail and resolution as it was converted to
video, and then as it became digitized, quantization and sampling artifacts came into play.
It may seem unlikely that we end up with anything worth watching, because we
have not finished perpetrating all manner of punishment and depredations on that
footage. Oh no, not by a long way. If you thought video caused some heinous damage to
your images, just wait until you see what a badly set-up compression system is capable of.
But don't worry. It is possible to achieve some amazing (almost miraculous) results
by taking care and minimizing the artifacts at each step. You might even make the result
look better than the original if you care to apply some restorative efforts.
Before we leave the area of source-content formats, we'll spend a chapter looking at
(or should that be listening to?) audio. While it might not seem obvious at first, many of
the problems with compression affect audio in the same way as video. Because we use a
different sense to experience audio, they seem to manifest themselves in a different way.