Telephoto stereo-photography

Alec uses two digital, Olympus, E 620, SLR cameras on a home-made mount. A narrow mount is illustrated here, but he uses up to 1 meter stereo base when required.
The cameras are triggered with a single radio transmitter, firing radio receivers on each camera. This gives accurate synchronisation, which can stop rapid movement. With SLR cameras there is much less trouble getting them to fire simultaneously than with a pair of small digital cameras, like Sony V3. A shepherd electronic synchronisation, to overcome un-synchronised clock cycles, is not needed on the SLR.
For telephoto stereo-photography, two 300mm zoom lenses are fitted. Since the 4/3 system uses half-size receptors, the lens focal length, in 35mm terms, is doubled to 600mm. Stereo base calculations for the fully zoomed lenses have to be done for 600mm focal length, since the formulae on this web site are for 35mm equivalent focal lengths.

Photography from a hide: Alec Kennedy

The bait for the wild harriers was dead possum. Australian possums are a pest in New Zealand, doing severe damage to native trees and farmers are encouraged to eradicate them. Traps and poison are used, but I prefer shooting them at night, using a powerful air-rifle fitted with a spot-light and telescopic sights. This provides food for the dog and bait for the hawks.

It only took a few weeks to get the hawks accustomed to feeding on possum on the lawn outside the kitchen window, even as close as a few meters.

My first photos had too much background disparity so I moved the bait closer to the nasturtium bank.
Shooting through the open kitchen window gave a subject distance of 16 meters. The hawks seem to easily see movement, even through the window glass, so it's difficult photographing from inside the kitchen.

Shooting through one (open) window I could use a separation of 300mm but the stereo image was "cardboardy."

I have 3 adjacent windows but only the two outer ones open, so I had to increase the separation to 1 meter to shoot through them. This gave a bit of hyperstereo, and excess background disparity even with the harrier close to the nasturtiums.

To reduce the stereo base to 600mm I then set up a hide outside the kitchen window giving a subject distance of 14 meters to the nasturtium bank. I also found that moving around inside the hide did not disturb the harrier. Double bonus.

The hide is a plastic cover draped over a stock crate on the back of a trailer. There is quite a lot of room inside for me and the cameras.

I have only triggered the cameras from within the hide, not remotely from the kitchen. Since I'm using spot focus for each shot I need to continually adjust the camera aim.

Over summer the hawks have been too well fed and there haven't been good photo opportunities. Also the bait site has been in the shade of a plum tree. It's now autumn so the tree has dropped its leaves, giving better lighting again. Also the hawks are hungrier. Time to try for some more 600mm photos.

The original results are in full colour and you can see some of them on the "3dpan" Flickr site, in cross-eye format.

On this page the hawks are anaglyphs, which Alec and many other stereo-photographers hate. John only tolerates anaglyphs because they are adequate as a demonstration format for people who cannot free-view stereo pairs on a computer. (The majority of people prefer anaglyphs, judging by the viewing numbers of various stereo formats on Flickr.) Also, colour coding of right and left channels makes it easier to measure stereoscopic disparity, which is tedious on side-by-side pairs.

A true colour stereo version is available here too, at decent size, with directions on how to see it in comfort.

Images:

1000 mm base

500 mm base

cross-eye versions

Use red/cyan anaglyph goggles to see in 3D.

Computing telephoto stereo base: John Wattie

Roundness

First decide how the images are to be seen and what the viewing distance will be. This determines the stereoscopic roundness.

For computer viewing at 1 meter, a base of n/15 is good.
For projection, 3D TV or a lens stereoscope, n/30 is required.
For movie screens the base may drop to n/100 or even less and then requires a mirror rig.

n/30 = 14000/30 = 466mm

n/15 = 933mm or nearly 1 meter

Maximum Acceptable Deviation (MAD)

This is set from the distance between the nearest object (n) and the most distant object (m). It can be computed accurately from the Bercovitz formula, or near enough from the pin-hole formula.

For use here, the formula is better re-arranged to find the maximum distance, m, since we already know the nearest distance, n, is 14 meters from the hide.

n = nearest distance: 14 meters
P = Parallax generally accepted for 35mm stereo photography: 1.2
B = Stereo Base: 600 mm
F = Focal length, 35mm equivalent : 600 mm
m = maximum distance: ?

m = 1 / ( 1 / n - P / B F )

m = 1 / (1 / 14000 - 1.2 / 466 x 600)

= 14894mm = 14.9 meters

14.9 - 14 = 900mm is a very small working distance and the background here is 1000 mm beyond the closest object. So the 600mm stereo base and/or the 600mm focal length must be too big. The harrier needs 600mm telephoto to fill the frame, so the stereo base has to be reduced.

The choice is to make a lot of experiments, wasting time and energy, or to calculate the base once.

B = P / ( F (1/n - 1/m) )

B = 1.2 / ( 600 (1 / 14000 - 1 / 15000 ) )

B = 420 mm

(The full Bercovitz formula, which is complex to run on a calculator, gives a base of 403mm, but this is not an important difference from 420mm in practice.

Bercovitz formula also reveals the correct distance to focus on for maximum depth of field is 14.48meters. It is including hyperfocal distance computation which separates the Bercovitz result from the pin-hole formula. The pin-hole equation assumes the camera is focused on infinity, which is certainly not true after Alec has gone to the trouble of setting precise focus before every shot.)

Conclusion

A 400 to 420mm base should be excellent for projection on an amateur size screen ( 4 foot, 1.22m ) and gives a roundness when viewed from 2 meters of:

Roundness =
[Stereo base actually used] / [Base computed for perfect roundness]

R = 420 / 466 = 0.90 (Pin-Hole formula)
R = 403 / 466 = 0.86 (Bercovitz formula)

SO, my advice to Alec is to use a 400 mm stereo base, from 14 meters away and then the stereo pair, through his 300mm lens on 4/3 camera, should be suitable for projection. Maximum fractional deviation should then be around 1/30, which is the recognised value for amateur projected images, according to the ISU
(Quoted in New Zealand by Max Pow).

For computer viewing, Alec's original 600 mm (big screen) to 1000mm base (smaller, portable computers) seems fine to me.