Olympus E-M5 and Panasonic GX1 Sensor Dynamic Range Comparison

Discussion in 'This or That?' started by Amin Sabet, Aug 20, 2012.

  1. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    Sony is known for making the sensors with the highest dynamic range (DR) in the industry today. In practical terms, this means that the shadow regions of files from Sony sensors are clean and can be "pushed" to reveal lots of usable detail. I compared the Sony sensor from the Pentax K-5 to the Panasonic sensor from the Panasonic GH2 in this DR comparison.

    We now know that the Olympus OM-D E-M5 uses a Sony sensor, but there haven't been many rigorous demonstrations of E-M5 DR compared to the newer 16MP generation of Panasonic sensors. Our member Jman presented a great comparison of processed E-M5 and GH2 files here, and DHart recently presented a very nice comparison of processed RAW files from his new G5 and OM-D.

    Having obtained a Panasonic GX1 to go with my E-M5, I spent a few minutes today on a dynamic range comparison. The GX1 isn't Panasonic's newest sensor - that would be the G5. Also, according to DxOmark, the GX1 has almost a stop less dynamic range than the GH2 at their respective base ISOs, so this shouldn't be taken as a comparison of Panasonic's best vs Olympus' best. Nevertheless, these are both high-end Micro 4/3 cameras that I love a lot, and they are the ones I had on hand for testing.

    Here is the scene I chose to compare, viewed from within Lightroom, Panasonic GX1 file on the left, Olympus E-M5 on the right:


    Below are 50% crop comparisons of the highlight areas after a 4 stop pull (-4EV) in Lightroom 4.1. Pay no attention to the TV, which can vary in brightness over the course of the exposure. The door and curtains in the back are lit by 12 ceiling lights and are a good subject for judging recovery from highlight clipping. As you can see, the GX1 highlights are more clipped:


    Below are the shadow areas from the same two RAW files after a 4 stop push (+4EV). Much cleaner, more detailed shadows from the Sony sensor in the E-M5:


    Test details: Tripod, self timer, Panasonic 14mm f/2.5 lens used at f/5.6 on both cameras, IS off (E-M5). Exposure time for the GX1 was 4s (base ISO 160) vs 3.2s for the E-M5 (base ISO 200). Lightroom sharpening/NR settings were left at default values.

    Here are the RAW files for your evaluation:

    What does it all mean?

    If you push the heck out of your shadows, the E-M5 files hold up better.

    How often do I do this?

    Almost never. So infrequently that to me these sensors are basically equivalent.

    For both low and high ISO in both low and adequate light, both sensors deliver great image quality. On any given day, my choice of cameras comes down to operational issues (eg, GX1 is smaller, OMD has body IS) rather than any differences between the sensor. Of course if you tend to push the heck out of your shadows, then the differences shown here are likely to matter much more to you than they do to me!
    Last edited: Aug 21, 2016
    • Like Like x 22
  2. Bhupinder2002

    Bhupinder2002 Mu-43 Hall of Famer

    Thanks Amin
    I also picked up GX1 and love it so far. The only thing which I dont like is the video on GX1. My Oly 720p is much cleaner and I have full control over video , apart from that its a very fast camera . I will pick up OMD in due course and till then will enjoy GX1 .
    Thanks for this comprasion again and I can see Harrison and Physiology books .
  3. CarlB

    CarlB Mu-43 Veteran

    For how I shoot in daylight, the Oly will have a *slight* edge. I underexpose some so that the sky and clouds don't go cyan or completely washed out. Then I push the landscape in post while masking-out push from the sky.

    But, having that great sensor and ease of use in something almost as small as the E-PL1 ... Dang!
  4. Steven

    Steven Mu-43 All-Pro

    May 25, 2012
    That's quite a difference in favor of the Olympus. Very nice to have for night photos.
  5. madmaxmedia

    madmaxmedia Mu-43 Veteran

    Feb 20, 2010
    This is the kind of thing that doesn't make a big deal in most usage, but it's nice to know that you can push a photo later pretty hard and still good good IQ.

    That's one thing I loved about the NEX-5N. Would be really curious to see how the E-M5 stacked up against the 5N (although I'm sure both are excellent.)
  6. danska

    danska Mu-43 Top Veteran

    May 21, 2012
    Portland, OR
    Real Name:
    Gradation set to auto in the EM-5? Interesting results, I would like take the time to use your files and see what the difference at -2ev or so would be.
  7. goldenlight

    goldenlight Mu-43 All-Pro

    Jan 30, 2010
    Real Name:
    Particularly impressive from the Olympus but it's worth noting that we won't be able to push the shadows like that at high ISO, where DR will be less in any case. Also, if we don't sometimes allow shadows to be shadows, fading into obscure murkiness, we become even further removed from the film "look" that many of us covet. Particularly impressive from the E-M5, yes, but here we have two great cameras with two great sensors and more important issues to influence our choice between them.

    Thanks Amin, this was really informative. :smile:
  8. Ray Sachs

    Ray Sachs Super Moderator

    Apr 17, 2010
    Near Philadephila
    The gradation setting shouldn't matter with raw files.

  9. With_Eyes_Unclouded

    With_Eyes_Unclouded Mu-43 All-Pro

    Apr 17, 2012
    Real Name:
    Thanks Amin, this was a great presentation!


    The thing with the OM-D is that you can expose to the right more freely. It doesn't burn the highlighs. In fact, that is one more point for the "film" behaviour.

    ETTR is my prefered method, I then have huge latitude to bring the image to "correct" exposure with minimal noise (which is, IMO, "pleasant" grain noise when it appears, anyway).

    Yes, but NR actually does. I don't know how the GX1 sensor handles NR, even at base ISO.
  10. DHart

    DHart Mu-43 Hall of Famer

    Jan 7, 2010
    Scottsdale, Arizona
    Real Name:
    Great comparison, Amin... Thanks.

    I'm eager to see what improvements are brought in the next wave of pens, GH3, and GX2. If the E-PL5 has the same sensor from the E-M5, that would be sweet news indeed and I'll readily swap my E-PL3 over.

    And I'm very eager to have Adobe LR4 adoption of RAWs from the G5, so I can redo my brutal G5/E-M5 DR test at the kitchen window... proocessing apples to apples, this time, no Sillypix allowed! :smile:
  11. mfj197

    mfj197 Mu-43 Regular

    Aug 20, 2012
    Guildford, UK
    Real Name:
    Thanks Amin, good post.
    Certainly true. However it's also worth noting that the E-M5 is very good at retaining high dynamic range as the ISO increases. Indeed, according to the review on Techradar the E-M5 still has a DR of 12.5 EV at ISO 1600, by which time the GX1 has dropped to below 9. Indeed, according to the same review the E-M5 has the same dynamic range at ISO 12,800 as the GX1 at base ISO (160)! I don't quite know how Sony do it ...

    Whether this is of any relevance to the camera user completely depends on the type of shooting they do. I have bumped up against the limitations of DR on my G3 a few times, which seems to be manifest in green and magenta blotchiness in areas of extreme shadow lift. However this is in very high contrast scenarios, and in most of the shots I take it's not an issue. Just a shame that I really like shooting the high-contrast stuff!

    • Like Like x 1
  12. Kiwi Paul

    Kiwi Paul Mu-43 Top Veteran

    Aug 15, 2011
    Aberdeen Scotland
    And once you are finished you can eat the apples too...or make apple pie!!! lol :redface:

  13. jnewell

    jnewell Mu-43 All-Pro

    Jun 23, 2011
    Boston, MA
    Useful posts here at mu43 are surprisingly frequent, compared to other forums. This post, however, crosses over into the category of "wise." Thanks! :2thumbs:

  14. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    Thanks, all. Glad this was helpful. I wish the big review sites would show the data instead of just curves and numbers for dynamic range. Worse still is when they give DR results for JPEGs only, but that's just my bias as a RAW shooter.

    Only if you push the shadows. If you leave them dark (ie, how they look when you import them into LR), then there is very little difference.
    • Like Like x 1
  15. danska

    danska Mu-43 Top Veteran

    May 21, 2012
    Portland, OR
    Real Name:
    I've noticed this in real world settings already. I still have a lot of workable shadow detail at much higher ISO settings than I do with my GX1. If you like to shoot dimly lit bars and night scenes it can be particularly helpful. I'm still relatively new to this but nailing exposure in these settings can be extremely challenging, so having some wiggle room is an added benefit.
  16. Luckypenguin

    Luckypenguin .

    Oct 9, 2010
    Brisbane, Australia
    Real Name:
    The usefulness of a higher dynamic range sensor is dependant on how you like to process and display your images. Going back over a year when I preferred a higher contrast look and liked playing around with various film filters a higher DR sensor wasn't such a big priority. These days however I often try to display a much broader DR when processing an image, even to the point where I will go back and re-process old images where I feel that I have previously destroyed way too much detail. The E-M5 has been a real eye-opener in how much DR you can squeeze out of a sensor and it has prompted me to re-evaluate how I shoot my GH1 as well to maximise it's DR. These two cameras best suit the look I am trying to achieve at the moment, even though there is almost three years between them.
    • Like Like x 1
  17. With_Eyes_Unclouded

    With_Eyes_Unclouded Mu-43 All-Pro

    Apr 17, 2012
    Real Name:
    Thank you for saying this so clearly Nic. I suddently feel less alone in a way! :tongue:

    DR latitude is and will always be the Holy Grail of digital camera design overall. And, if you think about it, it's the single most important feature that expands the artistic vision of the photographer. Specs and tech aside, having a wide DR simply frees you mind and let you concentrate on the important stuff, like making the picture YOU want to make.

    Apropos to this, I used to sometimens shoot bracketed HDR with Canon gear, in order to expand DR (not because I specifically wanted the "HDR look"). This is seldom the case with the E-M5.
  18. arentol

    arentol Mu-43 Veteran

    Jun 29, 2012
    Actually wide dynamic range is not really what matters. What really matters is the cameras USEABLE dynamic range. You can have two cameras, 12stop DR camera and a 10 stop DR camera. If the 10 stop one can recover 4 stops of highlights and 4 stops of shadow before it starts looking bad enough to be unuseable while the 12 stop one can only recover 3 stops of highlights and 3 stops of shadows before it starts to look bad, then the 10 stop actually had better dynamic range performance, even though it doesn't have as wide of a dynamic range.

    Basically current dynamic range testing techniques are no longer applicable. Current tests are a test of the outer limits where detail can be distinguished. This is fine, and we need to know this, but what really matters is the range in which the image is functionally useable. I really don't care if a camera has a 13 stop DR if I can only actually use +3 to -4. As far as I am concerned that is a 7 stop dynamic range camera, and that is the information I want from tests. The old testing method was more relevant back when it equated fairly well to actual useable range across all sensor from all manufacturers. Now Sony sensors are destroying that rule, and the testing methods have not come close to catching up with this change.
    • Like Like x 1
  19. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    I think the eyeball test I used here is pretty good for showing the usable dynamic range and also for giving some insight into tonal range. The Sony colors also hold up nicely to all the pushing.
  20. Amin Sabet

    Amin Sabet Administrator

    Apr 10, 2009
    Boston, MA (USA)
    That's really more of an indictment of Techradar's testing methods than it is a testimony to Sony's excellence in sensor design. Here's a comparison of the GX1 at base ISO (ISO 160) with the E-M5 at ISO 12,800.

    I shot the GX1 at nominal ISO 160, 14mm, f/5.6, 4s, which ended up with a touch more highlight clipping than the E-M5 at nominal ISO 12,800, 14mm, f/5.6, 1/20s (4-stop pull):


    Here are the shadows after a 4-stop push:


    Without a doubt, I could have exposed the GX1 file shorter to preserve the highlights (like the E-M5) and still ended up with more usable shadows than that purple mess from the E-M5 :biggrin:.

    Here's the whole frame without any pushing/pulling:


    RAW files from this comparison: