Light from f/2.8 lenses

A few days ago I noticed my Sigma 17-50mm 2.8 achieve different exposure at 50mm between f/4 and f/2.8. At f2.8 the lens was a third stop darker than expected.

It wasn't happening at all between 17mm - 35mm. I asked another forum member to test it on their Canon 17-55mm, and they also saw a slight LOSS of light on the long end wide open.

It leads me to believe that even on the same lens, f2.8 gathers a different amount of light throughout the range.
(I'm open to being proved wrong, it's driving me crazy thinking my lens might have a fault!)

Probably Sigma cut a corner a bit and the diaphragm at its maximum opening isn't quite enough to provide f/2.8 @ 50 mm. A slight fault, it just proves the lens is human

SASman wrote in post #14490474
It leads me to believe that even on the same lens, f2.8 gathers a different amount of light.
(I'm open to being proved wrong, it's driving me crazy thinking my lens might have a fault!)

I'm not surprised, especially 1/3 stop and especially with a relatively fast zoom.

Even for primes, I doubt that my sig 1.4 is actually quite 1.4 wide open, I think sigma got a bit generous with that.

In my experience no lens if perfect, they all have some 'flaws' or perhaps more accurately some 'character' to them. That's one reason even one good/experienced photographer will prefer one lens and another will prefer another, similar lens, often without necessarily knowing why.

tonylong wrote in post #14485936
Unfortuanately, this seems to break down with Macro lenses, at least with Canon Macros!

From what I've read, this is true of most macros (no sure about Zeiss). I think at the MFD the sigma 150 2.8 is actually something closer to f5 - 5.6.

I think Elie had the answer, that's at least how I understood it.

SkipD wrote in post #14485485
I have no idea what you're trying to say with the term "ratio".

Any lens set at a particular f-stop will illuminate the film plane or sensor surface with the same intensity of light. That's not saying that there are the same total number of photons going through the lens, as the field (angle) of view changes with different focal lengths. However with the same ISO value (or film speed), shutter speed, and f-stop setting the exposure will be the same (within manufacturing tolerances of the lens' f-stop settings).

In simplistic terms, the brightness of the image on the film will be the same for any focal length lens which has its aperture control set to the same f-stop value. If this were not true, light meters and the exposure settings we use would be useless.

tonylong wrote in post #14485936
Unfortuanately, this seems to break down with Macro lenses, at least with Canon Macros!

tzalman wrote in post #14490440
If I Remember Correctly (not at all guaranteed) that is because focal length is calculated with the lens focused to infinity. As the lens is focused closer the effective FL changes, thus changing the f-stop ratio, although for a lens with a normal focus range the shift is small enough to be within tolerances (we can, after all, only set exposure to 1/3 stop precision). However, at the close distances of macro lenses the relationship breaks down.

It occurs to me that in the same way that in a constant maximum f-stop zoom lens the aperture diaphragm is cammed to the zoom ring, it should be possible to make a macro lens with the diaphragm cammed to the focusing barrel and I expect that is the case with good quality lenses - as opposed to when you use extension rings and lose light.

ejenner wrote in post #14491636
From what I've read, this is true of most macros (no sure about Zeiss). I think at the MFD the sigma 150 2.8 is actually something closer to f5 - 5.6.

I think Elie had the answer, that's at least how I understood it.

I hope this little "side topic" is not confusing people!

As was said, when you are in a "normal" focusing range, then the differences are pretty negligeable.

But at macro MFD differences, it actually adds up to a stop (or more) of usable light.

So, with my Canon macro lens if I set it at the MFD it automatically slows the shutter speed down by I believe one stop or if in Manual the meter will show a stop underexposed, and this is no matter whether I'm actually metering/focusing on something at the MFD, so it's easy to test.

However, the camera will still show the same f-stop reading...it's a bit misleading, since we actually are not getting the same working f-stop!