Panasonic GH1 ISO 4000+ Equivalent Sample Images

Discussion in 'Open Gear Talk' started by Amin Sabet, Jul 18, 2010.

  1. Amin Sabet

    Amin Sabet Administrator

    Jul 3, 2010
    While the new Panasonic G2 and G10 have been well received, some current Micro 4/3 owners have been left wondering when we'll see a sensor upgrade in these cameras. It has been my impression that the performance of the Panasonic GH1 sensor is a small step up from that of the Panasonic sensor found in the other current Micro 4/3 cameras. This impression is consistent with the results of DxOmark testing. According to Panasonic, the reason for not using this sensor in other Micro 4/3 cameras is that it costs more (source).

    The one drawback I and others have found with the GH1 image quality is a susceptibility to banding, a problem which can be addressed using the debanding feature of Nik Dfine. Some have found that later production GH1 sensors are less susceptible.

    Most likely, we'll see a GH1 replacement at Photokina (September 21-26, 2010) or thereabouts, and I'm hopeful that we will see a new sensor from Panasonic at that time. In the meanwhile, I wanted to share a couple high ISO GH1 images taken yesterday. Each of these images was taken at the ISO 1600 setting, which according to DxOmark is actually ISO 2154. Both images were pushed 1 stop (Exposure set to +1.00) during RAW processing in Lightroom 3, bringing them to the equivalent of ISO 4308.

    Many review sites test high ISO performance in good light with fast shutter speeds, and that will give a rosy impression of sensor characteristics. These were each 1/25s (handheld, 45mm f/2.8) exposures in a dark room lit by a single halogen bulb. Images were processed from RAW in LR 3 (default settings for NR and sharpening) and then debanded with additional chroma NR only in Nik Dfine.


    Full-res JPEG:
    RAW file for download:


    Full-res JPEG:
    RAW file for download:

    For my purposes, this is more than acceptable image quality for ISO 4000+ equivalent images. If the GH2 can improve on this performance, it may be worth the wait.
    • Like Like x 5
  2. bilzmale

    bilzmale Super Moderator Emeritus

    Jul 17, 2010
    Perth, Western Australia
    Real Name:
    Bill Shinnick
    The GH2 just might replace my E-P2/G1 pair. Then again it will be hard to let them go.
  3. Herman

    Herman The Image Stimulator

    Jul 11, 2010
    The Netherlands
    Real Name:
    Fantastic shots with great atmosphere, well captured !
    I assume that Pan will bring nice stuff to table soon.
    • Like Like x 1
  4. soundimageplus

    soundimageplus SC Top Veteran

    Jul 6, 2010
    Good stuff, and I think I'll give Nik Dfine a look. Having a look at the DxO ISO results, and the differences between the claimed and measured ISO I was struck by the differences between the GH1 and GF1 and the E-P2. The Panasonics were consistently recording higher ISO's than they claimed, while the Olympus was recording lower ISO's than claimed.

    This relates to a couple of samples I shot to show the differences in low light performance between a GH1 and a Leica X1. I was scratching my head at the results, which didn't seem to make sense. The GH1 seeming to be able to have almost a 1 stop advantage in light gathering. The DxO results now explain why that was so.

    They haven't tested the X1 yet, but have tested an M9, whose claimed ISO was pretty close to the measured one. Assuming that Leica are pretty fastidious in their approach to this I would assume that the X1 is probably also pretty close.

    Having shot a lot a film, I still have a lot of shutter speed/aperture/iso combinations in my head and I've often raised my eyebrows at some of the readings my digital cameras have given me. These tests make it easier to understand why that is.
    • Like Like x 2
  5. Amin Sabet

    Amin Sabet Administrator

    Jul 3, 2010
    I read somewhere that the Panasonic G2 doesn't underrate ISO like the G1/GF1/GH1. When I do shootouts/comparisons, I always match shutter speed/aperture and try to normalize the brightness/apparent exposure at the end in order to get around these nominal vs actual issues, but there's no perfect way to do such tests. We don't even know for sure whether the shutter speeds and apertures are accurate.