GX7 and GM1 and noise

Discussion in 'Panasonic Cameras' started by MacBook, May 23, 2014.

  1. MacBook

    MacBook Mu-43 Regular

    Jul 24, 2010
    South Carolina
    The June Popular Photography issue reviews the GM1, which it says has the same sensor as the GX7. Their evaluation of noise from the GM7 seems remarkably different than the review they did in the November 2013 issue of the GX7.

    For the GM1, noise remains "moderate" 2.4 through ISO 12,800, and only becomes "unacceptable" at ISO 25,600.
    For the GX7, noise remains "moderate" 2.8 through ISO 800, and becomes "unacceptable" 4.0 at ISO 1,600.

    The article seems to attribute improved performance to the Silkypix software by writing:
    "Usually we end up talking about how the Silkypix RAW conversion software that Panasonic includes with its cameras doesn't vary the amount of noise reduction as the ISO changes. But that's no longer the case. The new version of this software applies a different amount of noise reduction for each ISO and yielded very favorable results."

    Do these rather stunning changes in performance seem attributable solely to the software change? Does that mean their initial test would look quite different with Lightroom?

    The results for the GM1 are beyond the Olympus EM1 performance as well, which leaves "moderate" 2.6 at ISO 3200 and becomes "unacceptable" 3.3 at 6400 (and "unacceptable" 4.4 by 12,800).
  2. Darren Bonner

    Darren Bonner Mu-43 All-Pro

    May 1, 2013
    Poole UK
    If this is true, I better load that version of Silkypix to my computer.
  3. pellicle

    pellicle Mu-43 All-Pro

    Feb 10, 2010
    Southport, OzTrailEYa
    Real Name:
    That has been my observation since about the first generation of micro43 Panasonic.


    Probably. My view is shoot raw and avoid the camera upgrade (essentially a subscription) cycle and instead focus on knowing your camera better and taking images with it :)
    • Like Like x 2
  4. T N Args

    T N Args Agent Photocateur

    Dec 3, 2013
    Adelaide, Australia
    Real Name:
    call me Arg
    To publish in Nov 2013, the test was probably done in Sept -- when there was no Lightroom support for GX7 (I think). So there could be no 'quite different' test result with Lightroom.

    If you are asking about the current version of Lightroom vs the current version of Silkypix -- who knows? Both have probably improved their converters.

    But I found the recent Silkypix GM1 result fascinating -- thanks. Maybe we should ask someone for side-by-side high-ISO Silkypix output from GM1 and GX7?
    • Like Like x 1
  5. poopstick

    poopstick Mu-43 Regular

    Aug 9, 2013
    Burlington Ontario
    I too would be interested in seeing the results of some testing on this.
  6. MacBook

    MacBook Mu-43 Regular

    Jul 24, 2010
    South Carolina
    That would be very interesting and potentially useful.
  7. T N Args

    T N Args Agent Photocateur

    Dec 3, 2013
    Adelaide, Australia
    Real Name:
    call me Arg
    Has anyone bought the GM1 or a recent Panasonic? What version of Silkypix SE came in the package?
  8. m4/3boy

    m4/3boy Mu-43 Veteran

    Jul 21, 2013
  9. Wasabi Bob

    Wasabi Bob Mu-43 Top Veteran

    In reading such reviews you have to remember that the final results are often the opinion of the person doing the review, and often based on a visual assessment. When processing RAW files, it’s up to the person processing the file to decide how much of an adjustment is enough or too much, often deviating from the software’s default settings. To confirm this, read several reviews and you’ll find they will differ.

    If these reviews were done, using DxO, you’d see the noise comments greatly decrease. I just started using it with GH4 and the results you can achieve, even at extreme ISO’s , are incredible.

    The important point to consider is that any review would be so much more helpful and factual if they used industry recognized software like Imatest. This software is what most manufacturers use the spec their cameras and lens. Specific parameters such as signal-to-noise rations and lines of resolution and color accuracy are measured, resulting in quantifiable numbers that remove the editorial comments from the process.
    • Like Like x 1