Image quality

Out-of-camera JPEG.
Olympus 12-100mm F4 Pro | ISO 200 | 1/40 sec | F8
Photo by Dan Bracaglia

With what Olympus is describing as the same sensor and processor as the Olympus E-M1 Mark II (and therefore related to the E-M1X), it's no surprise that image quality is very similar among these three cameras.

  • Good Raw image quality, on par with the best Micro Four Thirds has to offer
  • Pleasing color rendering
  • Good JPEG noise reduction, but overaggressive sharpening
  • High-res mode comes with a noise benefit, as well as bumping resolution
  • An electronic front-curtain shutter is well-implemented and enabled by default, contributing to sharper results at all shutter speeds
  • There is some occasional possibility of PDAF artifacts in backlit images or images with flare

Studio scene

Our test scene is designed to simulate a variety of textures, colors and detail types you'll encounter in the real world. It also has two illumination modes to see the effect of different lighting conditions.

Raw

Right off the bat, we can see that detail capture from the E-M5 III is basically identical to the E-M1X, which means it's quite good. Panasonic's G95 does a better job with text legibility on the smaller lines, and while the Fujifilm X-T30's X-Trans tech tamps down on false color, and shows a minor detail benefit from its larger sensor. It's the same story across other areas of the scene as well, with the Olympuses performing well, the Panasonic showing fewer artifacts (see how the word 'RED' appears almost broken up on the Olympus cameras), and the X-T30 looking good but not amazing in comparison. The Sony a6600 starts to show a bit more detail than the Four Thirds cameras, but introduces some false color as a tradeoff.

As the ISO value climbs, we start to see a clearer advantage to the Fujifilm's larger sensor (though Fujifilm looks to be employing some noise reduction on its Raw files), while all the Micro Four Thirds cameras look extremely similar. Once you reach ISO 12800, we think you're solidly in 'emergencies only' territory for the Four Thirds cameras.

JPEG

Switching over to JPEG, we can see the pleasing colors we've come to enjoy and expect from Olympus cameras in their default 'Natural' color profile. Reds and yellows are nicely saturated and greens are a tad warm (which we like).

JPEG sharpening leaves something to be desired - the slight 'haloing' around the gray patches is indicative of large-radius sharpening on the Olympus, which is less pronounced on the Panasonic. While large-radius can give the appearance of better results depending on how you're viewing your images, it's not the best method of sharpening for extracting maximum fine detail, which the Panasonic does a better job at. On the other hand, the Panasonic suffers from some 'stair-stepping' artifacts on curves or diagonal lines that you don't see as prominently on the Olympus cameras. That said, there are still some text artifacts on the Olympus cameras.

At higher ISO values, the Olympuses and Fujifilm all perform pretty similarly in terms of detail retention, with the Panasonic coming out ahead, but at the cost of a lot of artifacts. However, Sony's context-sensitive noise reduction, which does an impressive job balancing noise reduction and detail, pulls then slightly ahead of the Panasonic. That said, we find the luminance noise that the Olympus cameras leave behind to be more natural and less 'splotchy' than the Panasonic and Sony options here (we like the grain that Fujifilm leaves behind as well).

High-res mode

The E-M5 III offers a high-resolution mode that produces 50MP JPEGs or 80MP Raw files; Olympus says this is done by shifting its sensor 8 times in increments of 0.5 pixels. As you might expect, you get a massive resolution benefit, so long as your subject is non-moving, and downscaling to lower resolutions shows impressive results.

We should also note that the E-M5 III's 80MP high-res Raw files will respond well to quite a bit more sharpening than our standard studio test protocol, so we'd recommend downloading the files yourself to check out how they perform for your workflow.

Lastly, as you'd expect from sampling the same scene 8 times, using high-res mode will give you a significant noise benefit, especially if you downsize to a lower output resolution. Just keep in mind that the high-res mode maxes out at ISO 1600.

Dynamic range

The E-M5 Mark III performs identically to the existing E-M1X in terms of dynamic range, so it's in line with the best that Micro Four Thirds has to offer right now. In our ISO invariance test, we choose an exposure appropriate for ISO 1600 on m43, and then shoot ISO values up to 1600 using that exposure, brightening lower ISO Raw files as needed in post. Any differences in noise performance across ISOs must come from the camera (which is then being overcome by the additional amplification being applied at the higher ISO settings).

The sensor in the E-M5 III is a very good performer, showing essentially no noticeable noise difference from shooting at ISO 200 and brightening in post, as opposed to exposing at ISO 1600 at the time of image capture. So if you're trying to shoot a low-light scene that also has bright highlight detail you'd like to capture, you can keep your ISO value low and brighten the image later on: There will be a negligible noise cost and you'll retain multiple stops of extra highlight detail by not adding amplification.

You can safely shoot the E-M5 III at ISO 200 and brighten to an equivalent of ISO 1600 with little noise cost, but save multiple stops of highlights

The other way of exploiting a large amount of dynamic range is, in bright light, to reduce exposure to capture additional highlights, then brighten the shadow region. Reducing the exposure like this inherently increases the noisiness of your images, since you end up capturing less light and hence are more likely to see the randomness of the light (photon shot noise). However, some cameras perform better than others due to lower read noise. Comparing the E-M5 III's images shot in this way in our exposure latitude test again shows identical results to the E-M1X. This also means that sometimes, there's no substitute for a larger sensor like that in the a6600 which will respond better to these sorts of post-processing manipulations.

Shutter modes

Olympus has updated the shutter modes in the E-M5 Mark III relative to its predecessor. Without having to enable an 'Anti-shock' mode, the E-M5 III uses an electronic front-curtain shutter (EFCS) below 1/320 sec in both Single Drive and Sequential Low mechanical shutter modes, and switches to a full mechanical shutter above 1/320 sec. However, if you engage Sequential High, the camera will only use a full mechanical shutter at all shutter speeds.

The E-M5 III automatically uses EFCS below 1/320 sec, above which it switches to a full mechanical shutter

Olympus tells us the 'Anti-shock' mode is still present to allow users to customize the post-shutter release delay time, when working on a tripod, for example. In any case, we're pleased to see this solution. Automatically switching between EFCS and the full mechanical shutter means there is less for the user to worry about, and ensures the sharpest possible images at all shutter speeds.

If you want to shoot silently (or at the highest burst speeds the E-M5 III is capable of), note that the electronic shutter scans pretty quickly. Under certain lighting conditions or very fast pans, you may notice banding or rolling shutter artifacts, but we're confident it won't be a significant issue for most people unless you're shooting high shutter speeds under artificial light. And though it's a nice touch that the mechanical shutter goes to 1/8000 sec, the electronic shutter lets you shoot at up to 1/32,000 sec if necessary.

PDAF artifacts

Most mirrorless cameras these days come with on-sensor phase detection autofocus systems, which is mostly a good thing - you get reliable focus accuracy at the imaging plane, and lens calibrations are rarely necessary. Sometimes, though, we do see the impact of these autofocus systems in images, in the form of stripes or bands of pixels.

The dots within this purple flare are likely caused from the E-M5 Mark III's on-sensor PDAF system, but this occurs very rarely. Click through for the full image.
Photo by Richard Butler

The E-M5 Mark III is generally resistant to these sorts of effects, which most often show up in images with lots of backlighting or flare. Over the course of capturing thousands of images, we only found one instance in which any artifacts were visible, shown here. In the end, because of just how rarely we encountered this effect, we don't consider it a major shortcoming of the camera.