PDA

View Full Version : Inherent DOF characteristics tied to sensor size



jeebus
04-09-2013, 08:44 AM
Hey everyone,

I was wondering if someone could help me understand the complexities of dof as it relates to sensor size. A little bit of background:

I've been shooting projects for around 3 years now. I've always understood that dof at a given stop, distance from subject, and focal length, would be shallower on a larger sensor as opposed to a smaller sensor. All of my real world results, researching, and dof apps seemed to have confirmed this. I'm currently 1/2 way through a cinematography degree at Columbia College Chicago and had all of this understanding thrown out the window during class last night.

We were discussing lens characteristics (coatings, mtf, abberations, hyperfocal distances, etc...) when the conversation slightly turned to the aptly named circle of confusion. The point in which the conversation turned was when we were discussing "a 50 is a 50 is a 50" no matter what format. I completely understand that all of the spatial relationships between objects in a frame do not change when switching a lens from one sized sensor to another. However, my teacher said that the dof characteristics do not change either. When I contested that in fact dof does change, she brought up circle of confusion.

The definition I was given for COC is as follows: The largest size a point of light can be on the film plane (or sensor) and still be considered in acceptable focus. For 16mm film it is .0006" and for 35mm film it is .001". I know this has to do with intended projection size and whatnot. If you are planning on blowing 16 or s16 up to 35 or imax, you'd be wise in having a smaller circle of confusion. This makes sense to me. I also believe a larger circle of confusion (or larger projected display than initially intended) will create shallower dof due to the near and far focus points being pulled closer together as those points of light have become unacceptably soft.

However, that is the edge of my understanding. How COC applies in digital becomes hairy, for me. Is it simply that as sensors get smaller, camera manufacturers implement a smaller COC in order to maintain image quality when blowing up, adversely affecting dof in the process? Can anyone offer some insight?


Thanks!

morgan_moore
04-09-2013, 12:31 PM
Overall there is no theoretical reasons why one sensor size cannot match another sensor size completely - just open up to a wider Fstop when using a smaller sensor (and of course choose a wider lens)

I think to compare you need to consider additional 'variables' to see where the 50 is a 50 'falls over'

'Field of View' and 'distance to subject'

To see why large sensors are considered to have less DOF

A smaller sensor for a given lens and fstop will force you to back of to maintain FOV - and you will therefore be focussed closer to infinity and get more DOF

Conversly if you keep the camera in the same position and select a wider lens (to get the FOV) you will get more DOF

Going back to 'just open up to a wider Fstop' (to make a smaller sensor match a larger one) - in the real world that may not be so easy as the desired Fstop may not be available in an affordable lens

For example a 24/2 combo on a 'fullfrfame' sensor cannot easily be matched on a sensor half the size as you would (roughly) need a 12/1.4.. such a lens may not exist..

S

jeebus
04-09-2013, 01:23 PM
Conversly if you keep the camera in the same position and select a wider lens (to get the FOV) you will get more DOF


S

Thanks for replying!

I had not mentioned FOV because I was conceding different fields of view due to differences in sensor sizes.

Let me put my question into practical terms. If I have a C300 and a BMCC EF next to each other both using the 24-70 @ 70mm focused at say 10', the dof should be shallower on the C300, correct? The BMCC would obviously have a different fov but it would also have a deeper dof, no?

I shot a project on the Sony FS700 a few weeks ago and we ended up using the metabones speed booster with the canon cinema primes. In reading about the adapter, I came to understand that along with a wider fov & +1stop of light, the dof changed as well. Seemingly supporting the "larger sensor = shallower dof" as the adapter essentially (obviously not physically) increases the sensor size.

Thanks for any input! Just trying to wrap my brain around this concept.

morgan_moore
04-10-2013, 05:31 AM
Let me put my question into practical terms. If I have a C300 and a BMCC EF next to each other both using the 24-70 @ 70mm focused at say 10', the dof should be shallower on the C300, correct?

Nope the DOF will be identical - but the image will be more telephoto (cropped) on the BMC

BUT you move the BMC back until you get the Same FOV you will now be (around) 15 feet back and therefore focussed at 15 feet - which has more DOF than being focussed at 10feet!

If you keep the cameras at 10ft and zoom out with the BMC (to maybe 35) you will also get more DOF as you are now using a wider lens, but if you open up you can bring the DOF back making the images identical.. if you have the available stop on the lens

The speed booster adds a stop to a lens making it maybe 1.4 instead of 2, once you account for that and the markings now being wrong its just a 'normal lens' just a different one than before you applied it.

S

jeebus
04-10-2013, 09:12 AM
Nope the DOF will be identical

S

Exactly what I was looking for. Glad I could clear that up, finally. Thanks for your help!