R U Thinking f/1.4? F-Stop Blues......

[bhop]

Once again...we've got people making jokes about scientific testing, and we've also got really advanced, critical workers who have noticed that the linear exposure response we see at medium to small apertures does not actually exist at the widest apertures, and now DxO Mark is telling us that the camera makers are messing with the ISO gain control to either "overcome" or to "hide" this lack of linear exposure at wide apertures; those big, fast f/1.4 lenses sell for tons of money...The big camera makers can't have the rabble realizing that another $1,000 spent on an f/1.4 lens really does NOT GIVE THE USER a TRUE two f/stop gain in light gathering compared to what a modestly-priced f/2.8 lens does....

The reason you're getting jokes is because, as Sw1tchFX alluded to, not everybody, and i'd guess most people these days, aren't buying big aperture lenses for light gathering abilities anymore thanks to modern sensors' high iso abilities. They're most likely buying them for the super bokeh fad that remains popular, so in the actual really real world, these scientific tests aren't as meaningful as the gear heads and 'scientists' want you to think.. to most people anyway..

 

[djacobox372]

Ss I guess DOF has nothing to do with getting faster lenses?

Beat me to it, in this day of clean, sharp images as iso 6400 and higher, lens "speed" is not the primary reason to opt for wider apertures--its all about dof.

 

[Garbz]

It also specifically states digital sensors, what about film?

Take a sheet of film and look at it front on. Then look at it from the side. Both cases the results are quite similar as film has a matte surface. The chemicals in the film are equally reactive to light. Now since I'm assuming you don't have a sensor in your hand to play with, go to a window and look at it head on, now look at it from an angle and note the difference in the amount of reflected light you see.

This is a problem with photonics or opto-electronics. The interface between the air and the silicon has some kind of coating which often has a high refractive index. I.e. A photon that is incident on a point of the sensor at 45 degrees has a much higher chance of being reflected away than a photon hitting the lens dead on. You can see this quite well when you use an IR filter on the camera with a lens that doesn't have an anti-reflective coating on the back element, at wide apertures the light bounces back from the sensor, and onto the back element creating a hotspot.

At wider apertures the photons hitting the sensor come at more extreme angles. There are many ways around this problem. Nikon and Canon went for the microlens option, a small lens on each photosite which gives a photon a higher chance of being absorbed if it comes in at an angle. The 4/3rds format was also designed to have telecentric lenses for this reason and as such typically has less vignetting than other formats.

I do think some of this has to do with the lens as well.... that wasn't a variable that was much explored in this article. When I was reading tests of the Sigma 20mm f/1.8, corner vignetting wide open was so intense that the _average_ brightness across the frame was the same at 1.8 as it was at 2.5, but the center was still brighter at 1.8.

Yes as above. Telecentric lens designs, or lenses which have light rays which are more parallel to the optical axis have a better absorption rate. This affects both vignetting and total quantum efficiency. (Can't remember now if quantum efficiency is the right word for this. It's the right word for the other way round for sure, such as a photon escaping an LED)...

 

[Helen B]

Image-space telecentric lenses address a different problem to the one being discussed by DxO here, I believe. They reduce the angle of incidence at the image plane caused by the distance of the image point from the optical axis. The exit pupil is at infinity or at least very far away. The exit pupil should still have to subtend the angle that applies to whatever aperture is set.

 

[Garbz]

Isn't that part of the point? If you reduce the angle of incidence your photon capture rate increases. I think the same phenomenon which causes more vignetting on digital cameras than film cameras is at play here with wide apertures. Though am I right in saying telecentricity only affects the former and not the latter?