Lenses for low-light (long)

Discussion in 'Open Discussion' started by Mellow, Nov 16, 2011.

  1. Mellow

    Mellow Mu-43 All-Pro

    Aug 27, 2010
    Florida or Idaho
    Tom
    I've hesitated posting this for some time, but since I find it useful I thought I'd put it out there. If you hate math you might not want to read the entire thing.

    Like many people, I like to take photos in low light, and rarely have a tripod with me. To be useful under these conditions a lens must: (a) have a large aperture, obviously; and (b) have a short focal length. The short focal length is important because hand shake is magnified with longer focal lengths, which is the origin of the oft-quoted "1/FL" shutter speed limit for hand-held shots. The focal-length (FL) in this equation is the full-frame equivalent and has to be doubled on an m43 camera; i.e. you'd want a shutter speed no slower than 1/40 for a 20mm lens. It's important to note that this relationship is more than a "rule-of-thumb"; it can be justified theoretically (Shutter speed rule of thumb). Although some people have steadier hands than others, the minimum 'steady' shutter speed is always proportional to 1/FL.

    In order to compare different lenses with different focal lengths and different maximum apertures, I've created a little algorithm. Here's the math:

    • The light-gathering capability of a lens is proportional to 1/A^2 (A = aperture). In other words, a lens at f/2 will gather four times as much light as a lens at f/4 [1/4 vs. 1/16].
    • The slowest shutter speed you can safely use with a lens is proportional to 1/FL.
    • Combining these effects, the 'low-light-ability' of a lens is proportional to 1/(A^2*FL).

    Since the numbers produced are small, I've normalized them against the common Panasonic 20mm f/1.7 lens. The low-light-ability of this lens is: 1/(1.7^2*2) = 1/60.

    So, for example, the Olympus 12mm f/2 lens has a low-light-ability of 1/(2^2*12) = 1/48, which is slightly better than the 20mm Panasonic. You can compute just how much better by the formula: LLA = 60/(A^2*FL); for the Olympus 12mm, it works out to be about 30% better.

    So here's a table of lenses, sorted by their 'low-light-ability' .

    [​IMG]

    A couple of clarifications might be needed. First, note that I've assumed "f/1.7" is actually f/sqrt(3), "f/1.4" is actually f/(sqrt(2), etc. I've also tried to account for the fact that many C-mount lenses vignette, sometimes badly, by redefining an 'adjusted' focal length that has the same angle-of-view as the bright central part of the image. So, for example, the 25mm f/0.95 Navitar (which I own) vignettes pretty badly, forcing you to crop out the central part of the image, which essentially increases the effective focal length to 30mm. It's the adjusted focal length I use in the calculation of LLA.

    Well, what's the purpose of all this? (Some might argue: NONE!, lol). For me it helps me identify lenses that might be useful for shooting at night, or indoors, or under other difficult conditions. It explains the popularity of the C-mounts, which although they often vignette, also often have relatively large apertures. And it explains why the new 12mm f/1.6 Noktor might be a really cool lens.
     
    • Like Like x 12
  2. Nice post, and a useful table for comparison, as many users will be aware of what they can do with the 20/1.7. You've also given me another reason to give my wife as tomwhy the CV 25/.95 is "needed": "Honey, it's 2.7 times better in low light, and you know already how good the 20/1.7 makes you look!".
    I think one of the growing strengths of the format is how well it can do in low light, with certain lenses. Perhaps the high ISO is weaker than larger sensor formats, but small fast lenses more than compensate for that. A m43 body with the CV25/.95 isn't an imposing camera to take into a restaurant or club. Anyone like to research how big you need to go in APSC or FF to get similar ability in low light?
     
  3. Bokeaji

    Bokeaji Gonzo's Dad O.*

    Aug 6, 2011
    Austin, TX
    *brain exploderizing*
     
    • Like Like x 1
  4. RnR

    RnR Mu-43 All-Pro

    Sep 25, 2011
    Brisbane, Australia
    Hasse
    Interesting comparison. If you are shooting moving people ie your friends being their lovely selves infront of cameras, you would be trying to be in the 1/125-1/250 region, making 1/FL considerations moot. Yes?

    Or am I missing something yet again? :biggrin:
     
  5. Ray Sachs

    Ray Sachs Super Moderator

    Apr 17, 2010
    Near Philadephila
    I don't like math - too objective. :cool:

    -Ray
     
  6. avidone

    avidone Mu-43 Top Veteran

    520
    Jun 24, 2011
    Rome, Italy
    very interesting analysis... and I feel myself getting GASsy... oh no!
     
  7. Mellow

    Mellow Mu-43 All-Pro

    Aug 27, 2010
    Florida or Idaho
    Tom
    No, you're 100% right. I should have made a point that this analysis was JUST about light-gathering, and nothing to do with other factors that are often important. If the subject's moving you'll need a faster shutter speed, obviously. And if you want a deeper DOF then that will affect your aperture. And some lenses, like the 25mm f/0.95 Voigt, are pretty darn soft wide-open so that would influence your choice too.
     
    • Like Like x 2
  8. Orientator

    Orientator Mu-43 Regular

    48
    Dec 7, 2010
    Partially similar approach...

    Cool,

    there is someone in this universe who does similar calculations like I do! :smile:

    Please see my excel sheet for the same purpose as yours:

    LowLightAbilityCalculation.

    The calculation in column D is somehow like yours before normalization, just made a bit more readable by multiplying with 100000. Of cause this value is absolutely artificial.

    In column D I took the logarithm to the basis 2. Finally I got something like the "relative exposure value". I prefer this approach because in photography we often talk about "three stops" difference instead of "8 times the amount of light". So it seemed to be more natural to me.

    Looking forward to your opinions!

    Stefan
     
  9. RobWatson

    RobWatson Mu-43 Hall of Famer

    Yeah, but, my 25mm F0.95 cmount is so tiny compared to the CV 25mm F0.95 that it works 'better' in some situations. You know the CV 25mm F0.95 is actually a pretty sizable beasty?
     
  10. phrenic

    phrenic Mu-43 All-Pro

    Sep 13, 2010
    Cool post, thanks for sharing your calculations!
     
  11. Gwendal

    Gwendal Mu-43 Veteran

    300
    Jun 6, 2010
    One thing though : I feel that the 1/FL rule works less well with native lenses. Somehow, I'm able to achieve longer exposures without visible camera shake when using native lenses - 1/15 with the Oly 45mm for instance. With an adapted lens of the similar FL, I probably would have to stick to 1/45 to attain the same result. This might have to do also with the body, maybe IBIS works better with native lenses...
     
  12. meyerweb

    meyerweb Mu-43 Hall of Famer

    Sep 5, 2011
    Two other things to keep in mind: field of view and cropping, and IS (especially on Panasonic bodies).

    The 1/fl rule assumes full frame shots. Cropping means greater magnification of both the image and camera shake.

    Yes, a 12mm lens, following the 1/2*FL rule, can be hand held at a slower shutter speed than a 20mm lens. But if you have to crop the 12mm image to the same FOV as the 20mm lens to get the framing you want, then you're effectively shooting with a 20mm lens, and your handheld shutter speed limit is 1/40th, not 1/24th.

    And, of course IS changes things, too. The 14-45 at 14mm can be hand-held at slower speeds than a 12mm without IS.
     
  13. MajorMagee

    MajorMagee Mu-43 All-Pro

    Feb 24, 2011
    Dayton, OH
    I built a similar table of my lenses, but I have two other ways to do comparisons:
    Focal Length / min f-stop as an indication of Effective Lens Size (bigger number tends to be a bigger lens)
    Front Lens Diameter / Focal Length as an indication of Light Gathering Ability (bigger numbers are brighter)

    Combining them to determine relative capability was by dividing the entrance lens diameter by the focal length divided by the aperture size squared. Not surprisingly the PL 20mm f1.7 gets the top score, followed by my old Zuiko 50mm, f1.4, and a little more unexpectedly by the 9-18mm at the 9mm, f4 end of it's zoom.
     
  14. woody123

    woody123 New to Mu-43

    1
    Jul 7, 2012
    missing image

    hello mellow,

    this first post is missing the jpg image.

    can you re-upload it?

    thanks,

    woody
     
  15. kevwilfoto

    kevwilfoto Mu-43 Veteran

    294
    Sep 23, 2011
    Colorado
    can't see the chart

    I can't see the most crucial part of the OP's post - the data! Anyone know what happened?