1. Reminder: Please use our affiliate links for holiday shopping!

Background blur comparison with different lenses and sensors

Discussion in 'Open Discussion' started by Klorenzo, Apr 5, 2014.

  1. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    Hi,
    this afternoon I did a quick comparison of the background blur using different lenses. I wanted to do just a quick check, to pick the right lenses for the next shooting, so it is not a rigorous test.
    But the result was quite unexpected so I'm going to share it. Sorry for the ugly pictures.

    First I started with the olympus lenses, on the E-M10, the 60mm and the 40-150:

    P4050835_60.

    P4050868_40-150.

    There is not much difference and not much blur. So I tried with a couple of manual focus Nikon lenses, same lenght but faster:

    P4050855_nikon_50b.

    P4050850_nikon_135.

    Well, not much difference here either. Maybe it's true: i need a bigger sensor :(

    So I picked up the Canon 5D and tried the only lenses available: 50mm 1.8 and 105 f/4.5

    IMG_7913_canon_105.

    IMG_7916_canon_50.

    And...there is not much difference here either!!

    I know, I know, the framing (and the focus distance) is not identical, I should use a fast long lens on the Canon, etc. but I used what I got and I wanted to do a real-life comparison, where you cannot always choose the perfect focus distance and aperture. The most direct comparison 102 vs 105 has a very different framing.

    The point here is that I tried SIX different options and I got very similar results. Maybe the subject was just too close to the wall (about three meters). There ARE differences, but I expected a lot more.

    So what now? Not much. Either I buy a 24-70 2.8 or a 85 1.4 for the Canon or I'm not going to see much difference.

    Maybe tomorrow I'll do a couple more pictures with the canon, but I do not want to do a "scientific test", we have way too much of those and we already know THE answer. But I usually do not shoot in a lab so those test can be misleading, giving wrong expectations.
     
  2. taz98spin

    taz98spin Mu-43 Top Veteran

    843
    May 13, 2011
    NYC
    Sorry but I'm pretty sure that the Nikon 135 f/2.8 image your focus was off :confused:

    So was the 1st Canon image.
     
  3. flash

    flash Mu-43 Hall of Famer

    Apr 29, 2010
    1 hour from Sydney Australia.
    Gordon
    You're close to your subject and background. A 2 stop change in aperture is going to be measured in inches, not feet at these distances. If your 50mm 1.8 didn't show a huge change buying a 2.8zoom wont either.

    The differences in changing formats and apertures are subtle. It's up to the individual whether those subtle differences are important. With a super clean background (say the sky) it can be impossible to see ANY difference. In a messy forrest the differences might be significant to the image.

    The real trick is to learn what you have and shoot within it's capabilities.

    Gordon
     
  4. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    In both photos there is a lot of motion blur, both were shot with quite long exposures.

    For the nikon I did a quick manual focus with magnify and peaking, maybe I was off of a few centimeters (and was not easy to keep the beast steady on the oly body).
    For the canon I checked the full size image and looks fine. I used a single focus point: if that missed it would have gone to the back wall.

    Anyway the motion blur should add even more blur to the background. What I'm noticing is that there is not much in any picture, in all of them the jackets details are clearly visible.
    I did not expect some creamy bokeh but at least to be able to easily tell them apart.
     
  5. OzRay

    OzRay Mu-43 Legend

    Jan 29, 2010
    South Gippsland, Australia
    Ray, not Oz
    One important consideration about background blur (bokeh) is nature of the blur that you get. Specular highlights especially can present well or poorly depending on how the lens renders that blur. I'm not sure whether you were simply trying to compare depth of field or bokeh. If the latter, then you need to do outside or interior scenes that are better lit, with highlights, to bring out the best and worst of bokeh.
     
  6. coffeecat

    coffeecat Mu-43 Top Veteran

    708
    Aug 4, 2012
    SW England
    Rob
    You can get a reasonable quantitative idea of the differences in blur achievable between different lenses (and across different sensor sizes) by a fairly simple mathematical formula:

    See here:

    http://www.dpreview.com/articles/6091822765/background-blur-and-its-relationship-to-sensor-size

    Obviously it's an approximation, and it doesn't tell you anything about the "quality" of bokeh etc.

    Rob

    PS: one of my favourite quotes - "essentially, all models are wrong, but some are useful" - Box, George E. P.; Norman R. Draper (1987). Empirical Model-Building and Response Surfaces, p. 424, Wiley
     
  7. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
    OzRay, my focus is on background blur/subject isolation more than on the qualities of the blur.

    flash is right saying that I'm close to the subject, 2 to 4 meters, and below 5 meters the difference in dof due to different sensor is extremely small. I shot close to the subject to have a small dof but in this way I cancelled the sensor size differences. Blur and dof are different things but are related and I think the same also applies to blur.

    coffeecat your link is exactly what I'm evaluating and I took a little time to study it.

    OM-D
    50mm f1.4: 6.0%
    60mm f2.8: 3.6%
    102mm f5: 3.4%
    135mm f2.8: 8.0%
    (150mm f5.6: 4.7%)

    Canon 5D
    50mm f1.8: 4.6%
    105mm f4.5: 3.9%

    And this somehow matches with my tests. All these values are in the range that he calls "strong blur" or "very strong".

    BTW, with his model a 50mm on a FF will have twice the "blur index" of a 50mm equivalent on a sensor with crop factor 2.

    What I do no like is that it is way too abstract. It assumes the exact same framing, and this is correct in the context of his experiment. But, in real life, it is not a small detail being 0.5 or 4 meters from the subject.

    I think I discovered a few simple things, those that you do no fully realize until you do some experiments:

    1. long lenses are better for isolating subjects but not due to dof/blur differences, but because they just see a smaller part of the background so it is easier to find an uncluttered area. Then even a little blur will give you some nice isolation.
    2. Technical aspects are relevant (f, f-length, sensor size) are relevant but only when they are THE main limiting factor. Otherwise these things often cancel each other out when you can not control all of them.
    3. always remember to set the correct IBIS focal lenght when using manual lenses: it changes a lot.