E-M1 vs. E-M5 Noise/Hot Pixels at long exposure

Discussion in 'Olympus Cameras' started by ginobonetti, Oct 5, 2013.

  1. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    Hi all,

    I got my E-M1 yesterday, had some fun hours with it so far. However, I tested it against my E-M5 at long exposures and noticed something weird..

    Here's the test scenario:

    10, 30 and 60 seconds exposures
    m.zuiko 45mm
    f5.6
    ISO 200
    Noise Reduction and Noise Filter both turned off
    IBIS deactivated
    lens cap on, self-timer on and cameras put in a dark box

    The result:
    While the E-M5's 60 sec raw file is almost clean at 100% in LR5, the E-M1's raw is littered with noise/hot pixels.

    Here's two screenshots of LR5:

    E-M1_noise.png

    E-M5_noise.png

    Am I missing something? - Does anyone have an explanation for this? Shouldn't the E-M1 perform at least as good as the E-M5 under the same conditions?

    Can anyone of you E-M1 owners confirm this behavior? - Maybe I got a faulty unit..:frown:

    - Gino
     
  2. Serhan

    Serhan Mu-43 Veteran

    Joined:
    May 7, 2011
    Messages:
    439
    Location:
    NYC
    I don't have e-m1 but the sensors should be similar. From nex-5n to nex-6, the sensor got a small hit per dxo tests due to addition of af points like they did w/ e-m1. But in real world shots, it is hard to see that much difference.

    Other things causing additional noise/hot pixels might be the sensor heat, eg if you were using e-m1 before the test. You need to cool down both cameras before you can test it. Also you can remap the sensor pixels and see if that helps.
     
  3. bcaslis

    bcaslis Mu-43 Veteran

    Joined:
    Jul 3, 2011
    Messages:
    302
    Location:
    Wilsonville, OR, USA
    Real Name:
    Brian Caslis
    Lightroom 5.2 is seriously weak for the E-M1. I found you can't create your own profiles, it looks like the current profile is baked in and the red channel is seriously over done. I'd wait until the next release to judge the RAWs at this point.
     
  4. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Joined:
    Aug 6, 2010
    Messages:
    4,177
    Location:
    Davis, CA
    Real Name:
    Dara
    Any chance that you ran pixel mapping on the E-M5 at some point, but not on the E-M1?

    With long exposure NR off, I would definitely expect to see a fair amount of noise from both.
     
  5. christofp

    christofp Mu-43 Regular

    Joined:
    Jul 21, 2012
    Messages:
    95
    When I look to your test picture, I support "dhazeghi"s guess: pixel mapping.

    I had the same effect for my E-M5 when I got it back from LCD repair. I was shocked about the hot pixels but after thinking about it the reason was simple:

    Olympus had set the cam to factory defaults and the pixel mapping was resetted too.

    I did the pixel mapping in the utility menu once again and my camera is back to normal behavior since then.
     
  6. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    I'm not so sure on that.. I thought pixel mapping is for canceling stuck or dead pixels, not so much for hot pixel from long exposures.

    Olympus FAQ confirms this: http://www.olympusamerica.com/crm/oneoffpages/ask_oly/crm_e_ask_oly_03_09.asp

    Nevertheless, ich ran pixel mapping in the E-M1, but that did not work out..

    I ran another test, 60 sec. again.. same result. Looking awful..

    em_1_60.jpeg

    em_5_60.jpeg

    If I activate noise reduction in the E-M1 and the dark frame is taken, the result matches the E-M5 without noise reduction.

    Something's really wrong with this thing...
     
  7. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    I get the same results in OLYMPUS Viewer 3. And the differences are visible in the JPEGs as well..
     
  8. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    So, this is basically the rundown of the tests I've done so far:

    long_exposure_e-m1_e-m5.jpg
     
  9. LovinTheEP2

    LovinTheEP2 Mu-43 Top Veteran

    Joined:
    Feb 15, 2011
    Messages:
    610
    Location:
    Toronto
    Wow. Oi, is the em1 heading to a halt and delay like the pro lens? That's a pretty serious poorly functioning true pic 7 processor.
     
  10. bcaslis

    bcaslis Mu-43 Veteran

    Joined:
    Jul 3, 2011
    Messages:
    302
    Location:
    Wilsonville, OR, USA
    Real Name:
    Brian Caslis
    I wouldn't expect NR off on a 60 second exposure to look clean on any camera. Are you sure the E-M5 has NR off? Maybe the camera applies NR automatically even when off, there are cameras that have been like this before.
     
  11. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    Me neither, but for a ISO 200, 1 minute shot, the result is pretty bad.. Even my GX1 does better without NR..

    Yep, everything's switched off! If NR was turned on, it would result in 2x exposure time because of the dark frame and that would not go unnoticed..:big grin:

    Ideally another E-M1 owner could perform the same test, so I could at least be sure there is nothing wrong with this particular unit.
     
  12. dhazeghi

    dhazeghi Mu-43 Hall of Famer

    Joined:
    Aug 6, 2010
    Messages:
    4,177
    Location:
    Davis, CA
    Real Name:
    Dara
    Dark frame subtraction means the exposure takes twice as long, so it should be fairly clear if it's there or no. Standard noise-reduction settings (noise filter in Olympus parlance) only apply to JPEGs/RAW processed in Olympus Viewer
     
  13. michaeloneseven

    michaeloneseven New to Mu-43

    Joined:
    Oct 7, 2013
    Messages:
    2
    Sorry if it's a bad suggestion, but perhaps you can post over on the Dpreview M4/3rds forum? It seems quite a few people have received their E-M1 there.
     
  14. Ulfric M Douglas

    Ulfric M Douglas Mu-43 Hall of Famer

    Joined:
    Mar 6, 2010
    Messages:
    2,734
    Location:
    Northumberland
    Are they cooking the RAWs in the e-M5?

    Do the PD-AF pixels on the E-M1 cause the noise here?

    Just my two possibilities.
     
  15. gotak

    gotak Mu-43 Regular

    Joined:
    Nov 28, 2012
    Messages:
    185
    Location:
    Toronto
    I think the problem is the e-m1 doesn't have dark frame on by default. This really isn't an issue.
     
  16. LovinTheEP2

    LovinTheEP2 Mu-43 Top Veteran

    Joined:
    Feb 15, 2011
    Messages:
    610
    Location:
    Toronto
    You don find it an issue that the EM5 greatly out performs the state of the art latest and great PRO EM1.. Really HUH????

    As for dark frame on .. isn't that the feature that causes you a full extra shot to be taken and then zeros out the noise.

    The EM-5 didn't have it on so why should it be turned on for the EM1

    Since both cameras were setup the same.. why should the EM1 be handicapped needed the dark frame subtraction..

    OP.. can you redo the test at another ISO like 800 and see if you get similar profiles for the EM1 and EM5.
     
  17. MikeWhitten

    MikeWhitten Mu-43 Regular

    Joined:
    Dec 29, 2012
    Messages:
    56
  18. gotak

    gotak Mu-43 Regular

    Joined:
    Nov 28, 2012
    Messages:
    185
    Location:
    Toronto
    You seem to misunderstood. We thing the e-m5 simply have dark frame on by default.

    You take any camera out there and you don't turn on dark frame subtraction you are going to get some noisy images with long exposure. It's just the nature of the beast with imaging sensors today. The longer you run them the hotter they get and the more noise you get.
     
  19. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    Hmm.. could be possible! No clues from Olympus support this morning (I called them on that whole matter but they told me they did not have any experience with the E-M1 so far).
     
  20. ginobonetti

    ginobonetti Mu-43 Rookie

    Joined:
    Jul 30, 2012
    Messages:
    15
    Did that.. results look even worse.. Few more hot pixel in the E-M5 raw file, while E-M1 is off the charts..

    60 seconds with lens cap on, ISO 800, f9.0, IBIS off, Noise Reduction and Noise Filter off, screenshots at 100% magnification (this time from Olympus Viewer not LR, although LR looks almost identical):

    E-M5:
    e-m5.png

    E-M1:
    e-m1.png
     

Share This Page