Does size matter?
“What is the chip size?” The most basic digital camera question.
Chip size – or more correctly sensor size – directly and dramatically alters the way images look.
If you forget for a moment about useablity, physcial size of the camera and cost, most DPs will go for cameras with larger sized sensors – sensors the same size as a 35mm motion picture frame.
But who can forget about useablity, physcial size and cost when you are shooting television?
So it’s not surprising that 35mm sized sensor cameras like Red, Panavision Genesis (pictured above), Sony F35, and Arri D21 and Arri Alexas are used for shooting television dramas, high end commercials and feature films.
Cameras with 35mm-sized sensors offer the same shallow depth of field as 35mm motion picture cameras which means – all other things being equal – the image will have that big screen look.
The wild card at the moment is digital single reflex (DSLR) cameras. DSLRs are still photography cameras which in some cases will shoot HD video. The fuss is caused because cameras like the Canon 5D or 7D have slightly larger sensors than 35mm motion picture cameras and hold out the promnise of shooting with a large sensor but in the small size and comparativley low cost of a stills camera.
House snaps out from the background thanks to a shallow depth of field
With larger sensors focus becomes a creative tool for separating the subject from the background.
Small sensor cameras like 1/2″ or 1/3″ have a much deeper depth of field meaning that it everything, except perhaps the immediate foreground, tends to appear in focus.
The most common professional broadcast shoulder mount television camera sensor size is 2/3″ and they are somewhere in the middle; with very long lenses and wide open apertures you can use focus to separate your talent from the background somewhat, but nothing like the larger cameras.
There are some lens adaptors that allow 35mm lenses fit smaller sensor cameras and emulate the sensor size and look of large camera but they aren’t anywhere as good as the real thing.
Hi Peter,
I agree to all that you have said… in the case of 3D, if you believe the hype, then we’ll see heaps more of it, then I believe smaller chip size will be king. Depth of field is a MUST when shooting for 3d. The difference between Avatar and Alice in Wonderland is a prime example. Avatar, with its huge depth of field made for an excellent 3d experience… however Alice was shot as one would for 2D (minimal DOF) and the result was unpleasant… the viewer needs the depth to enjoy the “real” experience. I suspect that the 3D was an after thought in this film and most of the 3D elements placed in later. Sad as the viewer paying extra for poorly made films won’t be fooled forever!
Good point, ‘Yogi Bear’ was shot on 2/3″ sensor Sony F23s too. An irony perhaps that to employ the binocular vision depth cue – i.e. stereoscopic/3D shooting, you need to pretty much drop the shallow depth of field depth cue.
Thanks Peter, very interesting explanation of the differences between systems. I’m curious about the (physical/optical) reasons of the different DOF for different sensor sizes. Any clues?
At http://www.panavision.com/tools.php you will find a bunch of lens calculators which I wrote for fun on my laptop while watching TV in the evenings. I was working for Panavision at the time – some years ago.
From that site you can find out that a 50mm focal length full frame DSLR lens = a 15mm lens on a 2/3″ 16:9 video camera based on the width of the frame.
Take a subject distance of 5m and aperture of f2.8 on those lenses:
The 50mm DSLR gives: DoF near 4.3m, DoF far 6m, total DoF 1.73m
The 15mm 2/3″ gives: DoF near 2.15m DoF far infinity.
The truth is that the formula for calculating depth of field doesn’t contain image sensor size. It contains circle of confusion (CoC), focal length, f/stop, distance.
For two different format cameras shooting the same frame as in the example above DoF calculations depend on what CoC is selected and that is usually a nominal number appropriate to the image sensor size and dependant on the viewing image size or screen, viewing distance and average visual acuity of the viewers.
Most formats CoC seems to be about 1/1000 of frame width. (Somebody’s law I think – I can’t off hand remember who)
Apparent DoF is influenced by other factors including overall sharpness, and magnification and other stuff.
What I do know for sure – it works in practice.
To J Ubeda…
It is simple optics…
Hope it is clear enough…
In 35mm still cameras, normal lens is 50mm (aprox). As the sensor gets smaller the normal lens gets shorter too. So 2/3′ normal lens is 11mm.(You can also see this.. http://simple.wikipedia.org/wiki/Normal_lens). And because DOF is dependent of two factors (lens mm and iris), the larger the sensor, the longest the lens, the shortest the DOF…