Pentax 645Z Review

Detail and Moire

Back in 2011, when we compared the Pentax 645D (40MP medium format) to the Nikon D3x (24MP FF) and Pentax K-7 (14MP APS-C), we were blown away by the amount of detail that the 645D was capable of rendering.  Three years down the road, advancements in sensor technology have not only led to increased resolution, but also considerably less noise from all 3 sensor sizes.

On this page, we will compare the resolution of the Pentax 645Z (51MP medium format) to the Nikon D810 (36MP FF), Pentax K-3 (24MP APS-C), and Pentax K-50 (16MP APS-C) in a controlled studio setting, using a detailed model locomotive as our subject.

We'll start with a little teaser:

Pentax 645Z 100% Crop

As you can see, if it's detail that you're interested in, the 645Z doesn't disappoint.  But how much better is that 51-megapixel resolution alongside top-of-the-line sensors in smaller formats, especially as we increase the ISO?  To find out, we shot the scene below with each camera at ISO 100, 1600, and 6400. 

Test Scene (test 1 in yellow, test 2 in orange)

The yellow box (test 1) represents a 100% crop that's 448 pixels wide in the original 645Z image.  We upsampled the crops from the other cameras (using bicubic interpolation) to match this resolution, allowing you to easily compare the amount of detail present.

The orange box (test 2) encompasses a much larger area for which we provide unedited crops.

Test Conditions

All cameras were mounted on a tripod and fitted with prime lenses having roughly-equivalent horizontal angles of view.  Slight distance adjustments were made as needed.  The self-timer release mode was used to negate the effects of camera shake/mirror slap, and aperture adjustments were made to maintain comparable depth of field.  The shooting distance was approximately 1 meter.

The images on this page were developed from RAW (DNG or NEF from the Nikon) with no noise reduction.

Test 1

Take a look at the results in the table below, then scroll down for our commentary.

ISO 100
645Z
D810
K-3
K-50
ISO 1600
645Z
D810
K-3
K-50
ISO 6400
645Z
D810
K-3
K-50

The difference between each notch in sensor resolution is evident.  We asked an unbiased third-party individual to attempt to read the "Siemens" logo at ISO 100, starting with the K-50 image, and attempts were unsuccessful until we presented the 645Z image.  From the D810 image, only "S--M-NS" was recognizable.  At ISO 1600, the text is still clear from the 645Z whereas only the "NS" remains legible in the D810 image.

Even at ISO 6400 the 645Z holds its ground very well when it comes to detail, while the D810 begins to struggle and the K-3 more or less gives up.  It's also interesting to note that at ISO 6400, the K-50 image is superior to that of the K-3, which was overwhelmed by noise.

It should therefore be clear that the 645Z shines for landscape and product photography, where the rendering of fine details and the ability to crop are important.  Whether or not you need its high resolution over that of the K-3 or D810 will of course depend on how big you want to make your prints, and what ISOs you plan to use.

Test 2

Seeing a larger crop of the image that hasn't been scaled allows us to better evaluate the final product from each camera.  The 645Z's detail rendering is even more impressive here, as not only do we see dust, but also scratches and imperfections in the materials.

Click on any link in the table to open a full-size crop in a new window.

645Z ISO 100, 100% crop (click to enlarge)

ISO "Equivalence"

Finally, we asked ourselves the question of how high one needed to crank up the ISO to get roughly the same usable resolution from each camera.  Based on the results below, it appears that at ISO 12800, the 645Z offers similar performance to the D810 at ISO 6400, which in turn offers similar performance to the K-3 at ISO 1600.

In other words, at high ISOs the 645Z's resolution can give it a 1-stop advantage over the D810 and a 3-stop advantage over the K-3.

645Z ISO 12800
D810 ISO 6400
K-3 ISO 1600

Moire

The 645Z (and D810 and K-3) employs a sensor design without a low-pass filter for maximum sharpness, just like the 645D.  It is therefore slightly more susceptible to false color/moire artifacts than a traditional sensor would be.

In practice, however, we found moire to be a non-issue with the 645Z.  It did manage to make its way into some of our landscape photos (namely on cacti), and there's an extremely faint trace of it in the full locomotive cab photo (test 2) above.  As a general rule, if there is moire, it will only be visible in extremely fine, repeating details.  The larger the sensor resolution, the finer the details need to be before moire becomes visible.

The 645Z's playback mode offers an effective in-camera moire reduction tool.  We've shown examples of this tool in action in our Pentax K-3 review.

Moire atop cacti - click for full size JPEG

Print Sizes

The table below shows maximum print sizes you can expect from a single 645Z image at low ISOs.  We recommend a resolution of 300dpi (118 px/cm), which will produce very high quality prints and is the default embedded in the 645Z's JPEGs.  The maximum usable print size will start deteriorating quickly at ISO 6400 and above, as evidenced by our forthcoming studio tests.

Verdict

The 645Z renders never-before-seen amounts of detail at its price point.  More importantly, however, it offers better detail at higher ISOs compared to the D810, which wasn't quite the case when we compared the 645D to the D800E in the studio.  It's therefore a considerable improvement over the 645D not only in terms of resolution, but also noise.

Sensor technology has certainly been evolving quickly in recent years, and there's no doubt that the other cameras we tested on this page deliver great results given their pricing and target audience.  However, high-resolution still imagery from a tripod is the 645Z's domain, so it's hardly surprising that it dominated this series of tests.