19

When we say a light source emits 1500 lumens of light, it does not really matter if the light source is incandescent, CFL, diode etc. Similarly, when I say f/1.8 can I assume that all lenses pass the same amount of light at that f stop? If I am incorrect in this, is there a term that tells me across all lenses the amount of light they will pass through to the sensor?

mattdm
  • 143,140
  • 52
  • 417
  • 741
Regmi
  • 1,640
  • 3
  • 18
  • 35
  • Are you also concerned about the format size as well, or simply from lens to lens at (for example) f1.8? If I am incorrect in this, is there a term that tells me across all lenses the amount of light they will pass through to the sensor? No, as the amount of light from your source can change, so will the amount of light on the sensor even at the same aperture. When we say a light source emits 1500 lumens of light, it does not really matter if the light source is incandescent, CFL, diode etc. If you're not worried about colour/white balance, no. It's the same amount. – BBking Feb 08 '13 at 01:59
  • @BBking I meant any lens regardless of format size. – Regmi Feb 20 '13 at 07:20

2 Answers2

17

No, this is not the case. Aperture F stops are calculated on pupil size and focal length of the lens.

From wikipedia

In optics, the f-number (sometimes called focal ratio, f-ratio, f-stop, or relative aperture1) of an optical system is the ratio of the lens's focal length to the diameter of the entrance pupil.2 It is a dimensionless number that is a quantitative measure of lens speed, and an important concept in photography.

Whereas a T-Stop is a measured unit and lenses set at the same T-Stops will make the same exposure.

More, from wiki

A T-stop (for Transmission-stops) is an f-number adjusted to account for light transmission efficiency. A lens with a T-stop of N projects an image of the same brightness as an ideal lens with 100% transmission and an f-number of N. For example, an f/2.0 lens with light transmission efficiency of 75% has a T-stop of 2.3. Since real lenses have transmission efficiencies of less than 100%, a lens's T-stop is always greater than its f-number.[6] Lens light transmission efficiencies of 60%-90% are typical,[7] so T-stops are sometimes used instead of f-numbers to more accurately determine exposure.[8] T-stops are often used in cinematography, where many images are seen in rapid succession and even small changes in exposure will be noticeable. Cinema camera lenses are typically calibrated in T-stops instead of F-numbers. In still photography, without the need for rigorous consistency of all lenses and cameras used, slight differences in exposure are less important.

camflan
  • 2,071
  • 14
  • 27
  • You're generally right, but... Assuming all else is equal, lenses of differing focal lengths will let in the same amount of light at the same aperture, so the OP is basically correct. Lenses, however, are not always otherwise equal and some are better than others, and that is the T-stop case. It's your opening sentence that makes me hesitate to give a +1. – Joanne C Feb 08 '13 at 04:05
  • 2
    You are welcome to edit it ;) but I don't see how my statement is wrong though possibly overly pedantic :). If a variance in exposure is allowable, then yes f stop to f stop will give you the same exposure. But if accuracy counts, f stop isn't good enough. An F Stop is much more like a theoretical max. Lower quality will be much darker at the same f stop. Them there is also the issue of low quality lenses that actually vignette so much that they get dimmer as you open them up. – camflan Feb 08 '13 at 04:48
  • For practical purposes of exposure calculation, the f-stop is matched to the shutter speed and ISO to produce the exposure and I doubt most human eyes would really see much of a difference amongst lenses. It's your answer though, but I think your first statement is wrong, we generally don't use T-stop in our exposure calculations for photography. – Joanne C Feb 08 '13 at 04:57
  • 1
    JoanneC: Maybe not for still photography. Film and video tends to use T-stops. Check out B&H, all the lenses are listed as e.g. T1.5: http://www.bhphotovideo.com/c/buy/Lenses/ci/1884/N/4292338124 – Dietrich Epp Feb 08 '13 at 06:13
  • @DietrichEpp - We're not a video site... They don't do that for photographic lenses. – Joanne C Feb 08 '13 at 11:32
  • 1
    JoanneC: The line between photo and video equipment is getting blurry. Some of the mounts are the same. Some people are using video equipment for stills, or vice versa. It seems to be happening at both the low end and the high end, too. – Dietrich Epp Feb 08 '13 at 11:46
  • 1
    @DietrichEpp - You're missing my point, and it's backed up in the last sentence of the Wiki details copied into the answer... For stills photography (which this site is purposed to), f-stops are used because they're sufficient and do in fact basically describe the amount of light that can pass. At any rate, he's not going to pick a regular photographic lens (or P&S camera) and see t-stop on it, so why confuse the issue? – Joanne C Feb 08 '13 at 12:01
  • I actually tested this theory a couple of years back. There was a 1/3 to 2/3 stop variance between my lenses. The zooms at the long end of their range were the farthest off from the primes, or zooms at their shortest focal length. Maybe you don't care if shots are off by fractions of a stop - but that doesn't change the fact that same f stops do not equal same exposures. I am being pedantic, but I think that the more information people have, then they will be able to deduce what level of accuracy is sufficient. If I ask a question, I'd rather be over-informed than under. – camflan Feb 08 '13 at 13:20
  • 1
    For what it's worth, I don't use T-Stops in photography - but I would also never assume that 5.6 on one lens equals 5.6 on another. I'll swap lenses, re-meter, and then shoot. – camflan Feb 08 '13 at 13:23
  • @camflan the 1/3 to 2/3 stop variance you mention is much more likely to be due to the focal length differing due to focus distance, as well as the manufacturer stated fl and f/stop numbers being rounded very "optimistically", than differences in transmission between the lenses. [Good] cine lenses don't breathe and are probably closer to the specified fl making the t-stop slightly more worthwhile, though manufactures still tend to arbitrarily round to one decimal place. I've seen tons of t/1.3 spec lenses but never a t/1.345 – Matt Grum Feb 08 '13 at 13:47
  • 1
    I think the first sentence of the answer should read "In theory, yes, but in practice there is often a slight variance between lenses". –  Feb 08 '13 at 14:40
  • Just a single, but significant, case in point: the Sony 135mm f/2.8 [T4.5] STF. Shot wide-open, it is most definitely an f/2.8 lens, but if you expose for f/2.8 (rather then the T-stop of 4.5), you're going to miss by a mile. That was done deliberately by introducing an apodization filter to butter up the bokeh, but it also occurs, to a lesser extent, in other lenses due to coatings and imperfect transmission in various elements (especially non-glass crystals). I used to use external metering a lot, and 1/3 to 1/2 stops between lenses wasn't uncommon. –  Feb 12 '13 at 21:54
  • Based on one of the comments of @camflan,"..but I would also never assume that 5.6 on one lens equals 5.6 on another. I'll swap lenses, re-meter, and then shoot.", I choose this as answer as it presents me a very clear and concise view that is practical to follow. – Regmi Feb 20 '13 at 07:32
8

Basically, yes. From an optical perspective assuming you compose the shots similarly then different lenses with the same f/number will produce the same exposure. This is why the f/number is specified instead of the entrance pupil size (which is focal length divided by f-number).

To expand on this last point lets say we have a 100mm lens which has an entrance pupil of diameter 50mm. This means the maximum aperture opening appears to be 50mm wide when viewed from the front (the actual hole in the lens will be slightly smaller, as it is magnified by the lens elements, however this is not optically relevant - appearance is what matters).

Let's say we also have a 200mm lens which also has an entrance pupil of diameter 50mm. It would be tempting to thing that because both lens openings (apertures) are of equivalent size then these two lenses would produce the same exposure when photographing a uniformly illuminated white wall.

However the wider field of view 100mm lens views more of the wall (4 times more in fact) then the 200mm lens, whilst at the same time projecting the same image circle onto the sensor. 4 times more wall means 4 times more light onto the sensor, which equates to two stops difference in exposure. To get the same exposure we'd need an opening with 4x the area, i.e. double the diameter: 100mm.

It would be really convenient to have some value to quote for lenses that took this into account so you could know when exposure to expect from the lens. We've seen that a doubling of focal length requires a doubling of entrance pupil diameter to maintain exposure, so it's the ratio of these two values that determines exposure. Thus we can divide the focal length f by the entrance pupil diameter to create our new exposure determining value. Finally, to remind people of the meaning of this value we'll stick "f/" in front of it!

Thus our two lenses with equal size entrance pupils are actually f/2 and f/4!

This is all from an optical perspective, to address points raised in camflan's answer, in reality you might not get exactly the same exposure, due to different light transmission through the glass (a lens with the same f-number but different number of glass elements will absorb a different amount of light leading to a different exposure). The T-stop is a measure of how much light is actually transmitted, a throwback to when precise exposure matching was required when recording moving pictures on film.

However if you're going into this much detail then you also need to take into account the fact that the f/stop and t/stop numbers (as well as the focal length) are rounded arbitrarily by the manufacture anyway so there's no way to exactly predict exposure.

But two lenses with the same f-number will get you close enough for all practical purposes.

Matt Grum
  • 118,892
  • 5
  • 274
  • 436
  • 2
    It's also why a handheld light meter can give you a proper exposure value. – jaberg Feb 08 '13 at 00:32
  • 2
    @jaberg: As someone who uses a handheld light meter, you really ought to calibrate your film/camera/lens combination when exposure is critical, such as when shooting slide film. Different lenses make a difference in the exposure, in my experience. – Dietrich Epp Feb 08 '13 at 06:23
  • @DietrichEpp The key is the all practical purposes bit. You are correct, it's not an exact thing. Exposure will be effected by manufacturing tolerances, interaction between optics and body--how much light is lost between aperture and film/sensor, and emulsion chemistry. Quality of light may be a factor too. That being said, my Gossen Luna Pro generally works in the real world (sometimes better than in-camera metering) and, for *most* practical purposes, f8 is f8. – jaberg Feb 08 '13 at 14:22