In an interview with Markus Stone, Head of Production at Inition, he explained shallow Depth of Field in 3-D movie scenes: “I would argue that shallow depth of field does in fact work. The test of course, is to shoot the same image with shallow depth of field and a deep depth of field and compare the differences; which I have done [see below] using a canon 5D Mk2. This camera has quite a large sensor (24 x 36mm) so creating a shallow depth of field should be easy on this camera.
I have reproduced the images below as anaglyph [monochrome – to alleviate any color based retinal rivalry] and as ‘freeeview’ side by side pairs [you can view these on a passive stereo monitor too, but you’ll need to swap the eyes – right goes on the left for freeview pairs].”

First we asked, what is Depth of Field: “When making an image with a lens, there is a distance at which the focus is set. The depth of field (DoF) is how far an object can be from the focal plane (in front or behind), while still remaining in focus.  It is determined by the size of the camera sensor, the focal length of the lens and the T stop.
As shallow depth of field is best achieved with:

• Large sensors
• Long Lenses
• Wide open T stop
• Focus set to an object reasonably close to camera

A shallow DoF is commonly employed in motion pictures as a tool to separate the subject from a busy background, making the frame feel less ‘cluttered’ than if the entire image were in focus. It guides the eye to where the cinematographer wants the audience to look.”

He continued saying, that shallow DOF is common in close-ups as the lenses are longer. Closeups tend not to be shot on shorter lenses as it distorts the features of the face.
He shot a simple scene with a parallax of about 2% of screen width, which would be suitable for screens up to about 50 inch, he said: “To really get a sense of the stereoscopic effect, try viewing the images full screen. The subject of our scene and the convergence point was set on the Chinese incense burner 0.4m away. The furthest object in the background is 2.46m, and the lens was about 45mm.”

In the first example, (taken at T22) the background is perfectly sharp. Markus Stone says about this picture: “That’s one of the problems with such a large sensor. Even at T22 it can be difficult to get a really deep depth of field – which is the opposite to the problem you have with a camera with a small chip. While not perfectly sharp, it’s sufficiently sharp for our purposes.”
click here to see image full screen

Next shot is the exact same sequence of images at T4:
click here to see image full screen

“Comparing these two sets of images, there is no greater eyestrain or discomfort with the version shot at T4 as compared to those shot at T22. Some people complain that it feels strange to be converging on objects  that are out of focus, as you can in the above images. This brings me back to the point we touched on above. The use of selective focus in film language is to guide the viewers attention. The only time you will see the subject of the shot go out of focus is if the cinematographer is directing your attention to another part of the frame, for example, the ‘monster over the shoulder’ shot. The irresistible force that draws your eye to the monster, rather than continuing to look at our blurry hapless hero in the foreground is the same force at work here. Of course you can force yourself to look at an out of focus section of the image, but the eye seeks resolution by finding the point of focus. It feels wrong in 2D or S3D.”

Markus Stone elaborates: “There is also a practical reason for the use of shallow DoF in 3D live action movies; 99% of all shots on a feature film are generally shot on a beamsplitter rig, where the incoming light path is divided by a half-silvered mirror. One camera sees through the back of the mirror, the other sees the reflection in the mirror. By placing the cameras at a perpendicular 90 degrees with the mirror at 45 degrees between them, it is possible to get the cameras closer together than you could otherwise do by simply putting them next to each other.”

He continues: “A mirror rig takes the light that would typically hit just one camera sensor, and divides it between two, creating a 50% reduction in brightness, or a loss of 1 stop per camera. When shooting on set, this translates to requiring more lights to deliver the same light levels to the camera sensors, and on a big set doubling the output of all your lamps is an expensive proposition for the lighting budget. S3D films where lighting is involved tend to be shot with the iris wider than would otherwise be possible for 2D – which tends naturally towards shallow depth of fields for obvious reasons.

Another common misconception is that if the background is sufficiently blurry you can get away with having more parallax. I took the same series of photographs with an interocular of 24mm, which would give us a background parallax of 6% of image size. First the T22 images;”
click here to see image full screen

Markus Stone: “You’ll notice that there is a lot more 3D depth here, you’ll start to feel the eyestrain that is caused by this image being ‘overcooked’ in terms of 3D. You will need to enlarge the image to experience this as it’s pretty hard to overcook an image in a browser sized image because of the small image size. The disparity necessary to induce parallax based eyestrain in such a small image would first suffer from an unfusable background due to the massive background cropping effect of such large interocular distances before the parallax itself became an issue.”

Markus Stone, shows us the same images at T4:
click here to see image full screen

He comments: “Again, in terms of eye comfort, I find the above images to be equivalent. Observing the out of focus elements is just as uncomfortable to me in 2D as in S3D, but not more so.
To be thorough, I decided to test the same series on a longer lens. Here the T22 image has a visibly soft background, which is why I shied away from using it to compare shallow and deep depth of fields above [as he previously mentioned, with a big sensor it’s difficult to get a deep depth of field – with a long lens, even more difficult]. I moved the camera back and zoomed in to 105mm. At T22 it looks like this;” pointing to the next image:
click here to see image full screen

Now the T4:
click here to see image full screen

Markus Stone: “Here we see that no matter how blurry the background gets, it yields the same result in terms of eye comfort.
In general terms, then we can treat the background the same – no matter how in, or out of focus the background is.  The only time you can disregard the background is if it is completely featureless and undifferentiated in the horizontal plane – like a white cyclorama. The same concepts apply to foreground objects, but the complication of edge violations adds another element of possible confusion. The same principles apply equally to foreground and background objects when we are talking about depth of field alone. Once you add in the other factors to do with negative (foreground) parallax you find that – blurry or not – you must generally use much less foreground parallax; but that opens up another can of worms so perhaps we’d best leave it here for now.”

Markus Stone
Head of Production

Inition
www.inition.com.au

Company Description:
Everything in 3D. Includes Stereoscopic and 3D content Production for film
and broadcast, Interactives, Augmented Reality, 3D scanning and printing,
3D equipment sales.