DxO and E-M1 lenses: less sharpness than the E-M5?

Discussion in 'Olympus Cameras' started by With_Eyes_Unclouded, Dec 2, 2013.

  1. With_Eyes_Unclouded

    With_Eyes_Unclouded Mu-43 All-Pro

    Apr 17, 2012
    DxO recently posted their "best lenses" roundup for the E-M1.

    Checking those, a somewhat disturbing pattern emerged. Although some lenses seem to excibit the same "sharpness" (the way they define it) as the E-M5, in some important cases the E-M5/lens combination seems sharper. Examples are the 75mm f/1.8, the Panasonic 20mm f/1.7 (first version), etc. In no case is the E-M1 sharper.

    That said, they seem to measure slightly better light transmission in those lenses on the E-M1!

    So, what gives? Is this some flaw or variation in the measurement? Or does indeed the E-M1 sensor doesn't use the lenses as "efficiently" as the older OM-D?
    • Like Like x 1
  2. tomO2013

    tomO2013 Mu-43 Top Veteran Subscribing Member

    Oct 28, 2013
    It's hard to know if this these are the exact same lenses as they used on the E-M5 so could likely simply be sample variations in the lens as the differences are quite small in light transmission and sharpness where they exist.

    Some lenses score slightly better overall on E-M1, some slightly less on E-M5 and vice versa. You would think E-M1 would show overall slightly better result as the removal of AA filter, but since the E-M5 already has a fairly weak filter perhaps the changes are minimal at best where they exist in DXO's lab testing methodology. Interestingly those small differences in overall sharpness do not reflect their overall score for the lens/body combo's...

    Overall scores E-M1 :

    Overall scores E-M5 :

    It could also be accountable by a change in their testing methodology...we know too little here about their testing methodology to comment really :)

    Of course a large 'weighting' in DXO's overall score is attributable to the resolution of the sensor; so given that there was no major bump in megapixel count I woudn't expect to see a huge bump in scores here. DXO will generally always score a 24Mp sensor higher than a 16Mp etc... It's a nice indication to give you a basic idea of resolving power but doesn't necessarily equate to real world overall IQ performance or what people would commonly agree is subjectively going to give you the better image. It's a pure lab test, no more , no less.
    E.g. take a look at the scores and weighting for the Leica M9 and D700.....
    • Like Like x 1
  3. mattia

    mattia Mu-43 Hall of Famer

    May 3, 2012
    The Netherlands
    The one thing missing from these lab tests, for me, is a good description of methodology and variance measures for their tests. For repeatable, reliable results you'd want several bodies of a given model, several lenses, and multiple measurements (on multiple days, with resetting of the test setup) for each combination to account for any error in the measurement system. i.e. I'd like to see some confidence intervals on those score ;)
    • Like Like x 1
  4. riverr02

    riverr02 Mu-43 Veteran

    May 2, 2011
    New York
    Perhaps it's related to the pixel elements that now serve the PDAF function instead of an imaging function. That'd be my bet for the slight reduction in sharpness- and would also explain why Olympus would remove the AA filter to mitigate it somewhat.

  5. b_rubenstein

    b_rubenstein Mu-43 All-Pro

    Mar 20, 2012
    Melbourne, FL
    DxO does not publish resolution numbers, so there's no way of knowing what the actual difference is between say 7P-Mpix and 6P-Mpix. Due to the coarse granularity of the scale there could be very little resolution difference between a low 7P-Mpix and high 6P-Mpix.
  6. goldenlight

    goldenlight Mu-43 All-Pro

    Jan 30, 2010
    The E-M1 was probably tested before the new firmware.
    • Like Like x 1
  7. flash

    flash Mu-43 Hall of Famer

    Apr 29, 2010
    1 hour from Sydney Australia.
    Every lens I own shows *slightly* more sharpness on the EM1, including the 75mm. I'll trust my eye over DXO anyday, for real world usage.

    • Like Like x 1
  8. MAubrey

    MAubrey Photographer

    Jul 9, 2012
    Bellingham, WA
    Mike Aubrey
    You're seeing sample variation, nothing more. DxO doesn't test a dozen copies.
  9. pdk42

    pdk42 One of the "Eh?" team

    Jan 11, 2013
    Leamington Spa, UK
    That would be my guess too.
  10. Kiwimac

    Kiwimac New to Mu-43

    Oct 9, 2013
    I know that whenever I see an image that takes my breath away, DXO is so far from my thought processes that they may as well wind up the company…!
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.