When light passes from air to water at an angle,
it is refracted or bent at the interface. That can be either the
water's surface or, in our particular situation, the glass or perspex
of our underwater housing.
Imagine yourself resting on a raft floating on clear blue water
in a tropical lagoon. Feels good doesn't it. When you look directly
down in the water at a fish you are observing light that is passing
directly through the surface. If you reach down with a stick you
can easily give it a poke. At least you can if the fish stays still.
If the fish was to move away from you the light that you are observing
is bent at the surface giving you a false sense of perspective.
The stick now appears to be broken or bent so that you have to adjust
the angle of the stick to touch the fish.
A lens that has an angle of acceptance similar to our eye or longer
will work just fine behind a flat lens port. The light that it "accepts"
or takes in is not bent at the central axis, it passes directly
through and is not distorted.
However a wider lens starts to run into problems. Light passing
through a flat port directly in front of the lens axis is not bent.
However a wide angle lens begins to take in light further away from
the axis. As in our example above, the image becomes distorted at
the edges, stretched and warped by the light bending through the
flat port window.
Also, light is made up of different wave lengths (chroma or colours)
which are bent or distorted at different rates according to their
wave lengths. So red light at one end of the visible spectrum is
bent differently to blue light at the other extreme. The result
is colour fringing, the appearance of a colour halo around objects,
called Chromatic Aberration. It is this inherent quality of light
that enables a prism to break light up into it's colour components,
producing a rainbow-like spectrum of colours.
Next time you are viewing an aquarium, put your face close to
the glass. Look directly ahead. Everything looks fine. Now, without
moving your head, move your eyes to the extreme right or left. You
will see that everything becomes distorted, blurred and has colour
fringing.
The solution to this problem is to use a dome port, theoretically
a perspex or glass sphere cut in half. Because the surface is curved,
light passing through the dome interface always passes through perpendicular
to the dome surface, avoiding any distortion and chromatic aberration.

And that is why wide angle lenses should be used behind a dome
port. An added benefit is that the lens avoids the magnification
that flat lens ports produce. A 60mm macro lens becomes effectively
"longer" behind a flat port. A coin photographed above
water to fill the frame will need to taken from a greater distance
underwater to also fill the frame. The same lens used behind a dome
port (and yes I often use a 60mm macro behind a dome) avoids that
advantage/disadvantage (depending on your point of view).
It seems that, at least in the past, the people who designed camera
housings were not serious underwater photographers. In fact sometimes
I think that they never dived at all. I say this because with some
lenses, and in fact almost all zoom lenses, dome ports did not work
properly. While the centre of the photo was sharp, the edges were
soft, blurred and distorted with chromatic aberration. What was
happening? Had the laws of physics missed that day in class when
dome ports were discussed? Not at all. The problem was a little
theoretical point in space called a nodal point
Nodal points exist only in theory, sort of like those "points"
used in physics or algebra. "Point "a" is x-1 from
point "b" so x = a - b -1 (or something like that)".
Put simply, the front nodal point in a lens is where in theory all
of the light is focused by the front element(s). A dome port also
has a nodal point. In theory it is that point that is the absolute
centre of the sphere from which the dome was cut. If the lens' front
nodal point is not aligned with the dome's nodal point then you
will begin to get an out of focus effect, more noticeable on the
edges of your shots.
In almost all cases of blurry edges caused by dome ports (lets not
even think about blurry edges caused by cheap, nasty and badly designed
lenses which get worse when trying to work underwater) the problem
is the dome nodal point being behind the lens nodal point caused
by the lens extending into the dome.

My 60mm macro lens which I use behind a dome extends way, way
into the dome, almost touching the front surface at full extension.
However, because it is using only a small area of the dome, the
edge distortion is not noticed; it is not "accepting"
or taking in light far from the lens' axis.
Put a 20mm lens right up next to the dome surface and the photo
will be useless.
Subal make very nice housings. (Hello Mr Subal. How about sending
me a free housing.) However, their dome ports do not work. At least
not as you are directed by them. Back in the bad old days when I
bought my first housing (An Aquatica. Hello Mr Aquatica. How about
sending me a free housing.) I decided that I wanted to shoot stuff
that no one else was shooting, namely big animals such as whales
and white sharks, and I needed a really wide lens to get good clean
results. Being a commercial photographer I had a 14mm rectilinear
lens that I used to photograph very tall, skinny buildings. My first
white shark trip was coming up so I took it along.
All of the photos that I took with that lens had soft edges. Fortunately
most of the photos did not have anything of interest in the corners,
just empty water. The problem was that the 14mm lens has a very
long body. It is about twice as long as a 20mm lens. The front nodal
point of a 20mm lens was close to the nodal point of the dome so
gave good results. The 14mm was way out of the ball park.
My salvation was Chris Newbert, a photographer who was, at that
time, based in Hawaii and who specialises in coral reef scenes.
His subjects filled the entire frame so the fuzzy edges showed up.
He had a whole trip load of shots from the Rowley Shoals (I think
that was the destination) that suffered from fuzzy edges and were
all subsequently stored in land fill somewhere near Kona. He had
taken up the problem with Canon and had slowly figured it all out.
The answer was to move the dome out from the camera to align the
lens' nodal point with that of the dome. Over dinner (a great meal
of tacos at that Mexican restaurant in Kona right on the waterfront)
he explained how he worked out that the 8" dome was not an
8" dome at all and that using figures obtained from Canon in
Japan he was able to work out the length of a spacer to place between
the housing and the dome.
I used the same calculation procedure (my dome was a totally different
diameter as well) and had a spacer made up. It worked like a charm.
If you buy an Aquatica housing, they will provide you with a range
of spacers to suit particular lenses, zoom lenses being the main
culprit as they are much longer than fixed lenses of similar focal
lengths.
As a general rule of thumb, check your dome using this quick and
nasty method. Assemble your housing, lens and dome. Look directly
down (or across) the edge of the dome where it meets the housing.
If the front of the lens extends into the dome by any great amount
(over 1cm) then you will need a spacer.
|