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

EM1 vs EM5 ISO performance

Discussion in 'Olympus Cameras' started by kaitanium, May 12, 2015.

  1. kaitanium

    kaitanium Mu-43 Regular

    34
    Jul 23, 2012
  2. b_rubenstein

    b_rubenstein Mu-43 All-Pro

    Mar 20, 2012
    Melbourne, FL
    The difference is a fraction of a stop. The difference between scores have to be 2:1 for a whole stop difference.
     
  3. kaitanium

    kaitanium Mu-43 Regular

    34
    Jul 23, 2012
    Well that still means the EM1 lags a bit in that arena no?
     
  4. Wandering Aengus

    Wandering Aengus Mu-43 Regular

    105
    Mar 23, 2012
    Pretty sure they have different sensors. Not enough difference in that feature to sway me one way or the other.
     
  5. gryphon1911

    gryphon1911 Mu-43 Hall of Famer

    Mar 13, 2014
    Central Ohio, USA
    Andrew
    Yes, different sensors. One has Sony, the other Panasonic.
     
  6. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Yeah, a little bit more noise, a little bit better dynamic range...all those differences come out in the wash.

    The main reason for the different sensor is the on-sensor phase detection AF pixels. Without those, AF tracking would be worse, and 4/3 users wouldn't have a successor, so they're a big part of the camera's appeal.
     
    • Like Like x 1
  7. tkbslc

    tkbslc Mu-43 Hall of Famer

    You have to understand how DxO tests for their scores to make sense. And never trust the final score because they heavily weight dynamic range above everything else.

    The "Sports" score, you are likely focusing on is actually the ISO setting where DR drops below a threshold. It is correlated with noise, but it is not an actual measure of noise, and a camera that starts with a bit more DR will have an advantage here, even if the noise is similar.

    The real reason probably lies here in the "measured" ISO SNR test:

    http://www.dxomark.com/Cameras/Comp...1-versus-Olympus-OM-D-E-M5-Mark-II___909_1006

    The E-M5 II is less sensitive at the same ISO setting as the E-M1. Which means less noise, but also slower shutter speeds for same ISO/Aperture. But we are talking 1/3 a stop or less which within the margin of error for "correct" exposure.
     
    • Informative Informative x 1
  8. Klorenzo

    Klorenzo Mu-43 All-Pro

    Mar 10, 2014
    Lorenzo
  9. pdk42

    pdk42 One of the "Eh?" team

    Jan 11, 2013
    Leamington Spa, UK
    Frankly, I find it hard to believe anything DxO say. I saw a graph that showed the E-M1's DR bigger at ISO 100 than 200 - which is both counter to my and other people's observations as well as how ISO 100 is implemented. And don't even get me started on "perceptual Megapixels"!
     
    • Agree Agree x 2
  10. tyrphoto

    tyrphoto Mu-43 All-Pro

    May 25, 2014
    Seoul | NYC
    ㅇtㅈyㅅr
    I don't pay attention to DxO but my impressions after owning the EM1 and EM5 II are that the EM5 II seems to output ever so slightly cleaner files than the EM1. No idea about the EM5 since I never owned one.
     
  11. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Hmm, are you sure. I notice when you look at their measurements in the SNR 18% tab, the horizontal axis is the "Measured ISO" (as compared to the "Manufacturer ISO" axis in the ISO Sensitivity tab). So it seems like it should be an apples-to-apples comparison for the Sports score.

    I think the Sports score must be based on SNR, since the E-M1 beats the the E-M5 slightly in DR at every point, while trailing slightly in terms of SNR.

    Anyway, the two should almost definitely be considered equal and leave it at that. I doubt anyone can notice 1/6th of a stop of noise, even if they like to pretend they can.
     
  12. tkbslc

    tkbslc Mu-43 Hall of Famer

    According to their explanation page:

    "An SNR value of 30dB means excellent image quality. Thus low-light ISO is the highest ISO setting for a camera that allows it to achieve an SNR of 30dB while keeping a good dynamic range of 9 EVs and a color depth of 18bits."

    So I interpret that to mean that all of those conditions must be met. I am sure I could be wrong.

    I still have DxO aversion syndrome after shooting Canon for so long. THeir scoring method is not kind to Canon! :shakehead:
     
  13. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Hmm. The E-M1 has 9EV of DR @ 3200 ISO, whereas the E-M5 II is at ~2500 or so. The E-M1 has 18-bits at ~1250 ISO, and the E-M5 II maybe just trails it.

    So it must be the 30dB SNR that is the limiting reagent there, thus defining the score. It looks to line up pretty well with the charts.
     
    • Agree Agree x 1
  14. tkbslc

    tkbslc Mu-43 Hall of Famer

    Okay, so I was a little confused, then, thinking it was only the 9EV. And it appears that most cameras I have checked run out of SNR before DR, so 9/10 it probably is just going to be the 30dB SNR that ends up being the threshold.

    In my defense, I typed that up late last night! :)
     
  15. kaitanium

    kaitanium Mu-43 Regular

    34
    Jul 23, 2012
    I think for general info they are good. If I want to quickly see if a Canon 5D mk1 is better or worse in terms of low light performance than a OMD EM5 mk1 then dxo tells you right away. Anytime you get into the nitty gritty like we are with the em5 vs em1, then of course there is more room to argue.
     
  16. SkiHound

    SkiHound Mu-43 Veteran

    328
    Jan 28, 2012
    I think DXO measures something that correlates with sensor quality. I like their software. I'm a statistician and work in medical research. As I've gotten older and I think wiser I've grown increasingly, I'm not sure what word to use here so, skeptical of Western reductionist science. It's not that it's bad, I just think that gets so focused on tiny details that it misses the things that are really important. For example, instead of simply recognizing that broccoli is good for us, science want's to identify and isolate (and probably sell) those components that are good for us. Maybe we should just eat more broccoli, or Kale (hope I don't get sued for that). DXO is measuring things that are meaningful, but I'm not sure they're capturing all that determines sensor IQ with their ratings. Additionally, they acknowledge their ratings are basically +/- 1/3 stop and in terms of iso performance we basically have to double the iso rating to gain 1 stop. So the different ratings between the current 16mp m43 sensors I'd simply consider random noise. They are very small if taken as gospel and differences are within acknowledged measurement error. But we're just so ingrained to think that if 1000 is good, 1001 is better.
     
    • Agree Agree x 1
  17. Turbofrog

    Turbofrog Mu-43 Hall of Famer

    Mar 21, 2014
    Well said. Given the nature of sample variation and the "significant" differences that crop up in their testing of cameras that have identical sensors and image processors, I'm inclined to think that they would be much better off rounding their ISO values to the nearest 1/3 or 1/4 stop, and certainly more useful to consumers. The implied precision of 4 significant figures for their ISO values strikes me as a little bit...arrogant, and almost certainly wrong.

    But as you say, 1001 is better than 1000...and if I'm buying a new camera, I want the best, dogdarnit!
     
  18. b_rubenstein

    b_rubenstein Mu-43 All-Pro

    Mar 20, 2012
    Melbourne, FL
    If better measurable isn't better visible, then I don't care. If one can't correlate measured values to visually perceptible differences, then the numbers are meaningless. It all becomes a medieval Scholastic discussion of how many angles can dance on the head of a pin.
     
  19. I prefer the output from the E-M1 at higher ISOs (> 1600) to the E-M5, but I wouldn't necessarily say that it is because it produces less noise. Maybe it is because to me it produces better looking noise. My maximum Auto ISO on my E-M1 is set at 3200 compared to 2000 on my E-M5.

    Going even further, I actually prefer the output of the E-M1 to the E-M5 at any ISO.
     
  20. Dave in Wales

    Dave in Wales Mu-43 All-Pro

    Nov 5, 2011
    West Wales
    This strikes me as extreme 'pixel-peeping' with a vengence.

    Go and take some pictures and see if it makes a blind bit of difference.
     
    Last edited: May 14, 2015