1. Reminder: Please user our affiliate links to get to your favorite stores for holiday shopping!

12-50mm light-loss

Discussion in 'Native Lenses' started by silversx80, Jun 20, 2012.

  1. silversx80

    silversx80 Mu-43 Veteran

    229
    Apr 27, 2012
    North Carolina
    Hey all, time for another unscientific test... actually, it's quite scientific, just not as rigorous as other tests (make observation, form hypothesis, test hypothesis, observe results, rinse & repeat).

    Anyway, I've had a sneaking suspicion that the 12-50 kit lens has some minor light-loss that does not correspond to the aperture. Meaning, I can get the right depth of field for a given aperture, but my exposure will be just slightly under exposed. It's not a huge deal, and I'd guess it to be less than 1/3 stop. Also, I suspect it's a little worse at the wide end.

    Here are some samples, all in full manual. Also, lighting conditions were pretty equal for each focal-length comparison. Natural lighting in my office, not a cloud in the sky, each comp shot was less than a minute difference.

    The following two images are with the 17mm and the 12-50 set to 17mm
    E-M5 @ ISO 1600, 1/40s shutter, f/5.6 for both lenses (LN jpeg).

    1) The 17mm prime
    [​IMG]

    2) The 12-50 @ 17mm
    [​IMG]


    Next, the Oly 45 and the 12-50 set to 45mm.
    E-M5 @ ISO 1600, 1/20s shutter, f/8 for both lenses.

    3) The 45mm
    [​IMG]

    4) The 12-50 @ 45mm
    [​IMG]



    As you can see, the difference is slight, and more pronounced at 17mm. If you want full-res, let me know.
     
  2. dre_tech

    dre_tech Mu-43 Veteran

    314
    Jan 31, 2012
    You do notice that the field of view is not actually the same on the 17mm and the "17mm" reported by the 12-50 zoom?

    I'd be more concerned about that first, then light-loss. :wink:
     
  3. kevinparis

    kevinparis Cantankerous Scotsman

    Feb 12, 2010
    Gent, Belgium
    sorry... I see no difference in the pictures.....maybe you see the difference and maybe that difference will make a photo significantly better for you ... but i somehow doubt it

    K
     
  4. meyerweb

    meyerweb Mu-43 Hall of Famer

    Sep 5, 2011
    EVERY zoom has more light loss than a good prime at an equivalent focal length. Every element both reflects and refracts light, and even with modern coatings that adds up. The more elements, the more light loss, and modern zooms tend to be very complex designs with many elements.
     
  5. silversx80

    silversx80 Mu-43 Veteran

    229
    Apr 27, 2012
    North Carolina
    It's slightly off because I was not using a tripod, and that the light starts to bend at the front element, which is a couple inches further out on the zoom. This will give the effect of the camera being slightly closer to the subject. Also, my guess is that the camera has an error of +/- 0.5mm when displaying focal length.

    The difference is ever-so-slight, but noticeable to me, and certainly shows in the histogram. I'm not presenting this as a problem, however, but as an observation.
     
  6. heli-mech

    heli-mech Mu-43 Top Veteran

    959
    Mar 9, 2012
    Vancouver Island, Canada
    Andrew
    Every lens has a little different "tranmission" values, depends on how much and the quality of glass the light has to travel through. I believe DXOmark actually tests for this. Google T-stop vs F-stop and you should find some more info.
     
    • Like Like x 1
  7. kevinparis

    kevinparis Cantankerous Scotsman

    Feb 12, 2010
    Gent, Belgium
    stop looking at the histogram and look at the image would be my take...

    each their own

    K
     
  8. RobWatson

    RobWatson Mu-43 Hall of Famer

    Be prepared to be amazed as some lenses with very impressive F numbers can have very sad T numbers ... very hard to find T numbers for almost any lens except maybe cine lenses. Need F number to know DOF and T number to adjust lighting/exposure levels.

    Mostly technical stuff and not terribly relevent to actual photographic demands except in some niche applications.

    Ceratinly one should expect zooms to transmit light less efficiently as primes given that there are fewer glass/air interfaces in a prime than in a zoom lens (usually). Even crappy anti-reflective coating can be 98% transmissive so the differences can be slight - one stop would be huge and hard to find except maybe in some uncoated glass.
     
  9. David A

    David A Mu-43 All-Pro

    Sep 30, 2011
    Brisbane, Australia
    Yes, the 12-50 shots are just noticeably darker to my eyes. I also imported the first 2 into Aperture and tried using the dropper tool to measure the colour and luminance values for 2 spots on each shot, trying to measure the identical point on each. There was a difference with the 17mm shot being showing luminance values of 81 and 54 compared to 68 and 42 respectively for the 12-50.

    When I applied .35 stop negative exposure compensation to the 17mm shot, the luminance values for that shot dropped to 72 and 64 respectively but while those values are brighter than those for the zoom shot, the 17mm shot actually looked darker in tone than the zoom shot. Based on appearance I'd assess the difference between the shots in your post as less than 1/3 stop despite what the numbers seem to say.

    Bottom line: your minimum exposure adjustment is in increments of 1/3 stop. The difference looks to be less than that. There will be a difference in transmission between lenses as help-mech has pointed out. Given that, I don't think you'll ever get a shot from one of your primes and an equivalent shot from the 12-50 ever looking, or measuring, identically but the difference is very slight and less than the difference you're likely to introduce if you change exposure by 1/3 EV either in camera or in processing. Since both shots in each comparison look to be correctly exposed, ignore it. Just shoot and process to your liking. In real life you'll never tell the difference because you'll never be doing this sort of direct comparison and the zoom shots look fine.
     
    • Like Like x 1
  10. silversx80

    silversx80 Mu-43 Veteran

    229
    Apr 27, 2012
    North Carolina
    Indeed. I'm hardly one to measurebate, but sometimes I can't suppress the engineering instincts and the need to investigate the data.

    ... reminds me of a joke one of my engineering professors told me in college. How can you tell if someone is an engineer? They'll tell you.