understanding "fastness"

[geekpower]

Hi all,

I am relatively new to DSLR. I understand the relationship of aperture, iso and shutter speed, but when comparing a couple of lenses I've run into something that doesn't quite add up.

First a bit of background. I have a 6d and my only lens so far is a 50L. I know that this is a "challenging" lens but I chose it deliberately because I thought it would be good to learn on before branching out into other lenses. I figured the 50mm length would force be to learn good composition, while the 1.2 aperture would allow me to get "artsy" with my dof. As a side benefit, I've come to quite enjoy the ability to shoot hand held in low light with this combo.

My buddy lent me his 70-200 2.8L to try out. My impressions are that it is good and sharp, but I was having to use much, much slower shutter speeds and/or higher isos than I expected, even with the benefit of IS.

I decided to do a scientific comparison, so I set up a tripod, a focusing target, and some lights. Of course comparing 1.2 to 2.8 is not fair, so I tested both lenses at 2.8, and at 4.0 (since 2.8 stopped down vs 2.8 wide open might still not be a fair comparison, given the influence of the aperture blades). The end result was nonetheless consistent, the 70-200 was metering a full 2/3 of a stop less light than the 50 at the same settings.

What gives? The 70-200 has a bigger front element. How can it transmit less light at the same aperture?

This also raises the question, if I'm shopping for lenses that will be good in low light, what qualities other than aperture do I need to consider?

 

[Lee Jay]

Several ways - vignetting, and transmission losses are the most obvious, but since you don't have the same focal length you don't have the same scene either, and that could make in-camera metering inconsistent between them.

Vignetting will likely be higher on the wide open 70-200 than that stopped down 50L. Transmission losses will as well because the zoom has many more elements than the prime.

Look up the difference between T-stop and F-stop.

Lee Jay

 

[nc0b]

That is why T stops were invented, but unfortunately only used in the cine world. You didn't say which 70-200mm you had, but the current IS II has 23 elements in 19 groups. Your 50mm has 8 elements in 6 groups, so its light loss is going to be less. I own the 70-200mm IS II, and it is a fantastic lens in many ways. It is relatively large and heavy, but is a wonderful portrait lens, among its other uses. The only 50mm I own is on the other end of the spectrum, the f/2.5 macro lens. Both you and I own a 6D, and with its high ISO capability, the T stop comparison should be a non-issue. Use the appropriate lens for a given shoot, and let the camera sort out the F stop vs. T stop differences. (T stop lists the actual Transmission of the lens, not the physical aperture.)

 

[Don Haines]

The is the primes VS zooms thing....

A prime is a much simpler design with fewer elements. Less elements means more light. You loose a bit of light going through each element and a tiny bit reflects off of each surface. It all adds up, and in the end a F2.8 prime passes more light through it than a F2.8 zoom.

 

[geekpower]

Thanks, that is very helpful. I knew the 70-200 2.8 ii had far more elements, but I guess I falsely assumed that each element would transmit at least 99.9% and that the cumulative effect wouldn't be that great.

Even though the 6d has comparatively good high iso performance, between my fondness of hand-holding and inexperience at post-processing, I am still trying to use the lowest iso possible to allow me maximum room for error in over-sharpening without introducing a bunch of noise.

I can see why manufacturers would not want to publish their T numbers but I've found the info on dxomark, so that will be handy going forward.

70-200 2.8L ii = 3.6T
50 1.2L = 1.4T

Groovy.

 

[Deleted member 91053]

Interesting thread that explains what I had observed in the past but had no explanation for. The differences that I noted were less but still there.

 

[Sporgon]

I don't think 'T' stops fully explain what is happening. For instance the 24-105L has a T stop of 5.1 ( as opposed to f/4) and the 24-70 L IS has a T stop of 4, yet both these lenses will under expose by about one to two thirds of a stop compared with fairly simple primes such as the 50 /1.4 and the 85/1.8 for the same exposure setting.

 

[Marsu42]

I am still trying to use the lowest iso possible to allow me maximum room for error in over-sharpening without introducing a bunch of noise.

Don't let crop sensor habits get in the way of full frame shooting - everything below and including iso 800 is a wash on the 6d, including the dynamic range. Choosing a higher shutter speed will get you more keepers and make you a happier shooter than pixel-peeping in post.

My buddy lent me his 70-200 2.8L to try out. My impressions are that it is good and sharp, but I was having to use much, much slower shutter speeds and/or higher isos than I expected, even with the benefit of IS.

The recent IS systems are good for about 3 stops, though the effect is statistical and not the same on every shot. So for static subjects and calculating in the difference in focal length 50->70, the f2.8 lens should be roughly equivalent concerning camera shake.

The end result was nonetheless consistent, the 70-200 was metering a full 2/3 of a stop less light than the 50 at the same settings.

One reason is the heavy vignetting on fast primes, so what do you meter for anyway - the bright center to prevent clipping, or an average across the whole frame including the dark corners? Another potential reason for strange metering results is that the f1.2 lens cannot make full use of the additional light from odd angles on current sensors, so Canon cheats by secretly raising the camera's iso setting.

Every camera's firmware seems to deal differently with this, but it's safe to say that Canon puts a whole lot of work to make their premium f2.8 zooms like the 24-70L2 and 70-200L2 shine - including the best possible metering even on the 6d. On the other hand, I wouldn't rely on as much testing with 50L+6d from Canon.

 

[ajfotofilmagem]

That depends on a few factors:

1 Light transmission in T stops (darker than F stop).
2 Vignettes in the image corners when the lens is used wide open.
3 Different frameworks (50mm and 70mm) cause discrepancy in the camera's metering.

 

[candc]

I like your thinking with the 50l. If you can shoot that, you can shoot anything. You figured out something that took me a while: the lenses max aperture is just that, a physical relationship to the focal length. That relationship determines the depth of field but the light transmission also must factor in the quality and number of elements as well as the coatings.

Roger at lens rentals wrote a good article about it.

http://www.lensrentals.com/blog/2011/12/reflections-on-reflections-the-most-important-part-of-your-lens

 

[mps]

quite possible i now write total crap, but hey - thats what i think is right and if i am proofen wrong i learned something too so here i go

its about the FOV - a 50 has a greater FOV so there is naturally more light coming through the lens (or better - to the front element).

"The 24mm f/2 lens collects light from a comparably wider field of view than the 100mm f/2. Since they’re both f/2, they both capture light at the same “speed”. So for equal shutter speeds, they should provide the same illuminance at the sensor. So in terms of exposure value, the 24mm lens will produce equivalent brightness images for any given ISO and shutter speed because it’s pulling light from more of the scene than the narrower 100mm lens, hence the identical f/number rating" -> http://petapixel.com/2014/01/29/picking-great-lens-milky-way-photography/

as it comes for faster shutter speeds its basically the same thing - a rule of thumbs i once learned (which again can be totally wrong) is to at least double your focal lenght. for a 50 that would be 1/100, for a 200 thats 1/400... every1 feel free to correct me if i am wrong

 

[gregorywood]

Nothing to add here except to say that I'm enjoying this thread and I'm learning something. Thanks to all the contributors.

 

[geekpower]

quite possible i now write total crap, but hey - thats what i think is right and if i am proofen wrong i learned something too so here i go

its about the FOV - a 50 has a greater FOV so there is naturally more light coming through the lens (or better - to the front element).

if that's true then zooming from 70 to 200 should cause the exposure to be reduced, correct? i'll try this out later.

 

[ajfotofilmagem]

quite possible i now write total crap, but hey - thats what i think is right and if i am proofen wrong i learned something too so here i go

its about the FOV - a 50 has a greater FOV so there is naturally more light coming through the lens (or better - to the front element).

"The 24mm f/2 lens collects light from a comparably wider field of view than the 100mm f/2. Since they’re both f/2, they both capture light at the same “speed”. So for equal shutter speeds, they should provide the same illuminance at the sensor. So in terms of exposure value, the 24mm lens will produce equivalent brightness images for any given ISO and shutter speed because it’s pulling light from more of the scene than the narrower 100mm lens, hence the identical f/number rating" -> http://petapixel.com/2014/01/29/picking-great-lens-milky-way-photography/

as it comes for faster shutter speeds its basically the same thing - a rule of thumbs i once learned (which again can be totally wrong) is to at least double your focal lenght. for a 50 that would be 1/100, for a 200 thats 1/400... every1 feel free to correct me if i am wrong

Let's simplify things:

Imagine two different lenses are used in F5.6 to minimize vignetting and transmission discrepancy (T stop versus F stop), with the same framework (at different distances). In this case, the lens and 50mm to 100mm will have exactly the same exposure.

You could spend time theorizing about the pupil diameter and size of the front element of the lens, but none of that changes the fact that the exhibition will be exactly the same.

 

[e17paul]

I like your thinking with the 50l. If you can shoot that, you can shoot anything. You figured out something that took me a while: the lenses max aperture is just that, a physical relationship to the focal length. That relationship determines the depth of field but the light transmission also must factor in the quality and number of elements as well as the coatings.

Roger at lens rentals wrote a good article about it.

http://www.lensrentals.com/blog/2011/12/reflections-on-reflections-the-most-important-part-of-your-lens

+1
I learned SLR photography using a 50/1.4 mainly at 1.4 to make full use of available (usually indoor) light. I had to learn how to make the best of it, I could only afford ISO 100 film in my student days. Its great that you are taking the 50/1.2 as a challenge, it's also good for disciplined framing.

Now wide apertures are a creative choice instead of a technical necessity, and I try to keep ISO at 800 or below on my 6D. It's great to still get reasonable results beyond 800, especially for low res web output.

 

[chromophore]

The size of the front element of a lens is only an upper limit on the light-gathering ability of a lens. What really matters is the diameter of the entrance pupil.

Take a lens and remove both caps. Set it to infinity focus and place it on a table on its side, or mount it on a tripod if it is heavy. Now stand next to the lens and look through the front element along the optical axis--that is, look at it straight on, right through the center. Observe the apparent size of the circle of light that is created within the lens. This is the entrance pupil.

Depending on which type of lens you are using, the entrance pupil may look like it takes up most of the front element's size, or it may not. For example, an 85/1.2L will look almost completely "empty" inside. But a 14/2.8L, even when viewed with the aperture wide open, will show an entrance pupil that is nowhere near as large as the entire front element.

Furthermore, the distance at which you are viewing the entrance pupil influences its apparent size in relation to the lens: if you are viewing it close up, you will perceive it to be smaller than it actually is. This effect is more pronounced with telephoto lenses. For example, take a 300/2.8L IS and look at the entrance pupil from a distance of a few feet, then from several yards. The size of the entrance pupil will grow until it appears to occupy almost the entire front element.

The entrance pupil's apparent diameter as viewed from a distance infinitely far away, is part of what determines the light-gathering ability of a lens. The other part is the focal length of the lens. The ratio of focal length to entrance pupil diameter is called the f-number of the lens.

This explains why a lens with a very long focal length requires a large front element in order to gather the same amount of light as a short focal length lens. A 50mm lens that achieves f/2 has at that aperture an entrance pupil diameter of 25mm. But a 300mm lens, in order to achieve f/2, would need a 150mm entrance pupil diameter; and since the entrance pupil cannot be larger than the front element, the latter must be made at least as large as 150mm.

Ignoring transmission and other minor effects, two idealized lenses of different focal lengths will have the same light-gathering ability (in the sense of achieving the same exposure) if they are shot at the same f-number. So the aforementioned 50mm and 300mm lenses, if both were shot at f/2 and the same shutter speed and ISO, will achieve the same exposure, although the apparent field of view and magnification will be very different between the two. That is to say, over the region of the scene that is common to the two photographs, the brightness of that area will be the same. Loosely summarized, "f/2 is f/2 on any lens."

Thus far, we have discussed only those lenses with a single focal length--"primes" as they are commonly termed. If we look at "zooms"--i.e., varifocal lenses--we can see that changing the focal length of such a lens must also have consequences on the entrance pupil diameter.

Some zooms have a fixed fastest f-number across its focal length range; for example, the 70-200/2.8L IS design achieves a fastest f/2.8 at all focal lengths. Other zooms (usually cheaper ones) such as the 28-135/3.5-5.6 IS has a fastest aperture of f/3.5 at 28mm but only f/5.6 at 135mm.

If we take a lens like the 70-200/2.8, then you will see that as you change the focal length, the size of the entrance pupil will change in proportion to the the focal length, even though the front element stays the same. This shows you that for a typical zoom design, it is the desired f-number at the longest focal length that determines the minimum size of the front element.

 

[Lee Jay]

quite possible i now write total crap, but hey - thats what i think is right and if i am proofen wrong i learned something too so here i go

its about the FOV - a 50 has a greater FOV so there is naturally more light coming through the lens (or better - to the front element).

if that's true then zooming from 70 to 200 should cause the exposure to be reduced, correct? i'll try this out later.

Not correct unless the scene is non uniformly lit (which most are).

 

[Mt Spokane Photography]

Thanks, that is very helpful. I knew the 70-200 2.8 ii had far more elements, but I guess I falsely assumed that each element would transmit at least 99.9% and that the cumulative effect wouldn't be that great.

Even though the 6d has comparatively good high iso performance, between my fondness of hand-holding and inexperience at post-processing, I am still trying to use the lowest iso possible to allow me maximum room for error in over-sharpening without introducing a bunch of noise.

I can see why manufacturers would not want to publish their T numbers but I've found the info on dxomark, so that will be handy going forward.

70-200 2.8L ii = 3.6T
50 1.2L = 1.4T

Groovy.

Canon gives "T" stops on their Cinema lenses just like other lens makers. The definition of a aperture is a physical opening and does not tell you how much light is passing thru the lens. A T Stop does tell you.

If you read this section from Wikipedia, notice that the aperture is physical and does not specify how much light is passing thru the lens.

"The lens aperture is usually specified as an f-number, the ratio of focal length to effective aperture diameter. A lens typically has a set of marked "f-stops" that the f-number can be set to. A lower f-number denotes a greater aperture opening which allows more light to reach the film or image sensor. The photography term "one f-stop" refers to a factor of √2 (approx. 1.41) change in f-number, which in turn corresponds to a factor of 2 change in light intensity."

 

[Schmave]

If we take a lens like the 70-200/2.8, then you will see that as you change the focal length, the size of the entrance pupil will change in proportion to the the focal length, even though the front element stays the same. This shows you that for a typical zoom design, it is the desired f-number at the longest focal length that determines the minimum size of the front element.

I've wondered about this before. For the constant aperture zoom lenses, based on their physical design, could they also be made to be a variable aperture lens with a larger aperture on the short end of the range (or should I say, lower F number)? For example, with the 70-200 f2.8, at 200mm the entrance pupil has to be ~71.4 mm. If you took this same entrance pupil size with the focal length at 70mm, couldn't the lens max aperture be ~f1.0? So does the lens maker limit how much the entrance pupil opens based on the focal length? Or am I missing something (does it also have to do with the front element size?)?

 

[mackguyver]

I'm glad Mt Spokane Photography brought up T-stops as they are the actual measurement of the light transmitted through the lens. It's one of the more interesting things that DxOMark measures and if you look at the aperture vs. T-stop for many lenses, it's eye opening.