While I am trying to figure out what I need for split-underwater photography, an interesting fact came to light:
A lens focuses at different distances in air than underwater.
How is the focus different? If the lens is set to focus at a certain distance, can one estimate at what distance it will focus underwater? What about DOF, does it stay the same or is it affected too?
Does this depend on the type of underwater casing? What about the lens port or dome? Do these change how focus is affected?
Some physics applied here. This question has been similarly asked in some websites like Yahoo Answers!, Wiki Answers and questions like If a spherical mirror is immersed in water, does its focal length change?.
Well, the first question is How does the lens focuses on light? It's back to our old physics days when we studied the general properties of light. Light travels in faster in some materials and slower in other materials. For example, in air light moves at almost two hundred thousand miles per second, while in glass it moves at about two-thirds that speed. Same thing, light travels faster through air and slower through water. The focal point moves farther away from the lens and objects do not appear as big through the submerged lens.
In my first link, the answer involves computations and formulas, but it was concluded that focal lens of the lens increases when it is underwater. This was also called "The Art and Science of the Over/Under" by Stephen Frink
Here's one portion of the article:
Focus Factors
The dome port offers no special optical properties above water, functioning essentially as a clear window. This means that a subject above water, 10 feet away, will be in sharp focus when the lens's distance scale is set to 10 feet. Lenses are calibrated to work in air, unless they are amphibious water-contact lenses like the Nikonos 15mm or 13mm (for the RS), in which case they don't work well above water at all. In a housing, the topside portion will have air in front of the lens and air in front of the port, so there is no differential medium.
For the portion of the dome port that's in the water, however, the rules change. Because the underwater portion has air in front of the lens, but water in front of the port, the dome port acts essentially as another lens element. A "virtual image" of that portion of the photo is created inches in front of the lens.
The point of focus for the virtual image is dependent on the size of the dome. In ballpark terms, the virtual image exists at twice the diameter of the dome away from the lens. For a six-inch dome, the virtual image is about 12 inches away, and for a nine-inch dome, the focus point can be assumed to be 18 inches away.
Of course this means that a lens that can't focus as near as the virtual image won't be able to focus at all. Some wide-to-telephoto zoom lenses may have a minimum focus of nearly two feet, so to use them behind a six-inch dome requires the use of a supplementary close-up lens known as a diopter. This will change the range of focus, accentuating the near focus capability and therefore allowing focus on the virtual image.
and his Pratical Solution
The Practical Solution
A more forgiving means to shoot over/unders is with a full-frame fisheye lens (16mm Nikkor or 15mm Canon). Recently, I have had good luck using the Nikon 12-24mm zoom on my digital SLR cameras as well. The depth of field with these wide lenses is fantastic, enough so that with an aperture of f-8 or smaller it is possible to hold focus on both the topside and underwater scene. I usually focus on the foreground part of the scene, which tends to be the underwater section, and let the background go a little soft if necessary.
UPSIDE: The photographer has the choice of vertical or horizontal compositions because the water/air interface can be anywhere in the frame. Also, this optic is more forgiving of surface chop as there is no fixed intersection of diopter and neutral density to contend with.
DOWNSIDE: The downsides include the extreme distortion typical of fisheyes, and the lack of exposure correction. If you're shooting a shallow reef with lots of reflecting sand, the f-stop difference may be very slight. In this case, you should expose for the highlights (topside) and let the underwater portion record slightly dark. Local controls in Photoshop can bring up the levels in the underwater scene.
Basically the water/air transition in a dome port acts as an additional lens element which results in magnification of the image. I believe the exact amount depends on the lens and curvature of the dome (though I'm not certain on that). I don't know that there is a particular way to quantify the difference. Normal autofocus should still work though since it is effectively focusing on the image being produced by the lensing effect of the dome port and the water/air transition.
Since the transition should have a magnifying effect, it has the effect of increasing the focal length of the lens. With a flat port, it's about a 25% increase in focal length. I'm unsure what portion of this is corrected by a dome lens, but it should still be an increase in focal length.
If an object is half in the water and half above it and pretty much parallel to the film/sensor plane the portion under water will appear closer than the portion above water, due to the higher density of water than air.
Many split underwater shots I see take advantage of this fact to feature an underwater object that is closer to the camera and a larger above water object in the background. As with any shot when a wide depth of field is required, use of a narrower aperture allows a greater possibility that everything will appear sharply in focus at an acceptable display size.
