Lens test: Cooke S4 + Alexa

Directly behind the model's head is a small 300w Fresnel. When the model is in their normal relaxed position this light is blocked. When she leans to one side, we get a "direct" flare in the lens. Another source (also 300w) is placed out of frame aimed at the front element of the lens to look at "indirect flare "or veiling glare.

 

First I record a few seconds of the model still, then they lean to the side and we record a few seconds of direct flare( the second image) then they lean back and we turn on the side light( the third image ) and record a few seconds of indirect flare then turn that off and cut the camera. It all happens in one take .

Testing each lens at every stop is crucial. Not only are "aberrations" (chromatic or spherical)effected, but contrast reproduction and sharpness are also effected and diffraction starts to play a role as you stop down.

 

While changing lens aperture, changing as few parameters as possible that might effect the results. I prefer to leave the lighting alone. So I light to 5.6 at the camera's native ISO (for epic is 800). This way we can achieve exposure compensation for every stop on the lens merely through shutter angle and frame rate changes.

​

The reason for compensating exposure through the fps/shutter combo is that adding neutral density filters affects MTF ; Once you add any glass, you should multiply their MTF value correspondingly(the value is definitely <1). Thus, adding any other glass will affect contrast/sharpness, the test should had one variant at once(which is the Aperture of course) , one should maintain other factors as possible. Changing ISO is not ideal, take epic for example, it affects the dynamic range distribution(above or below middle grey), this may be misleading while

evaluating at post. You want to actually increase or decrease the light coming into the lens, not just adjusting nominal ISO. 

​

That's the quick look-up chart here. We're not looking at motion in the test, so those changes do not adversely affect the results. Notice that 5.6 is set at 180 degree shutter and 24fps. When opening up to a 4, exposure compensation is achieved by closing the shutter angle 90 degrees (1 stop). The sequence continues with closing down shutter angle and then increasing frame rate (if the camera is incapable of the extremely tight angle necessary for a T1 exposure - assuming the lens is that fast). When the lens is stopped down to an 8, we open the shutter up to maximum and beyond that we slow down the frame rate. Not all lenses (extremely few, actually) open to a T1 or F1, so it's rare to go to that extreme. Also not all lenses close to T/f22, some stop at T/f16 or even T/f11.