Performance

Operational Speed

The K-3 is a fast and responsive camera in-use. From hitting the power switch to being able to take the first shot takes a fraction under a second - a respectable, if not quite immediate, result. In single shot mode there's around a half second pause between being able to take another shot, which again is a the sort of performance you should expect at this level. The Info menu appears almost instantaneously, as does the main menu, so there's very little that makes you feel you're waiting for the camera.

AF System & Performance

Continuous autofocus is one of the main areas of improvement that the K-3 promises over existing Pentax models. Our experiences of it were pretty mixed. There are so many variables inherent in continuous autofocus testing that it's essentially impossible to characterise performance in an entirely repeatable (let alone 'definitive') manner. However, in discussion with Ricoh, we did our best to get a sense for what the K-3 was and wasn't capable of.

We tested continuous focus with a variety of Pentax lenses, having discussed relevant settings with Ricoh. We used some lenses focused using the camera body's built-in motor (70mm Limited and HD 50-300mm), some with ring-type SDM motors (55mm and 50-135mm F2.8), and others with DC micromotors (18-135mm and 20-40mm Limited). Of these, the DC motor-based 18-135mm lens was the fastest to refocus and performed best in our tests.

The K-3 offers seven AF point selection modes when set to continuous autofocus (C-AF). We found its SEL (27 points) mode did a pretty good job of staying on the specified subject, even when it wandered off the originally specified focus point. Adjusting Custom Setting 18 (Hold AF Status) defines how sensitive the camera is to the subject falling off the selected focus point, allowing you to tune the behavior to the type of motion you're shooting. Between these two settings, the camera was able to retain focus lock on a subject moving in the X and Y axes of the image.

What the K-3 struggled with rather more, every time we tested it, was subjects coming towards it (movement in the Z, distance axis), and we weren't ever able to get the camera to pretty successfully track a subject in both the Z and Y axes, i.e. moving around the frame and towards the camera. However, it appears that a side-effect of the 'Hold AF Status' setting not only reduces the system getting distracted by things passing in front of the subject or the subject temporarily moving off the current AF point, but also reduces its willingness to refocus at all, which isn't quite what you want.

Much of the problem with Z-axis tracking appeared to relate to the focus speed of the lenses, with even the fastest lens we used struggling to work at much above walking pace. Subjects moving laterally (and therefore requiring little refocusing) weren't a problem, but even the fastest-focusing of the lenses we used would struggle to get a good number of shots in perfect focus with an approaching subject.

Frame 1
Frame 2
Frame 3
Frame 4
Frame 5
Frame 6
Frame 7
Frame 8
Frame 9
Frame 10
This sequence is an example of something a phase-detection system should be able to do well - a subject travelling towards the camera in a fairly predictable fashion. I'd estimate the subject's speed as under 10mph, so above walking pace but by no means fast.

These images are part of a longer sequence, where the rider has already been under the central focus point for at least five frames preceding the ones shown, so there's little ambiguity about the subject. AF Hold set to Off, Focus Priority for shots in a continuous sequence.

Ultimately, the camera's autofocus performance appears to be limited by the lens, since it appears to be doing a good job of understanding what the correct subject is, but struggling to refocus fast enough to cope with approaching subjects. This means it's quite possible the K-3 performs more effectively with the system's fastest-focusing lenses (the 60-250mm F4 being an oft-cited example). However, having tried to use a cross-section of appropriate lenses, we found that the camera wasn't able to work to its full capability - a limitation we think many users will hit.

Frame 1
Frame 2
Frame 3
Frame 4
Frame 5
Frame 6
Frame 7
Frame 8
Frame 9
Frame 10
This sequence is an example of something a phase-detection system should be able to do well - a subject travelling towards the camera in a fairly predictable fashion. I'd estimate the subject's speed as under 12mph, so above walking pace but by no means fast. We had considerably less success if the subject deviated from the central focus point.

Here AF Hold set to High, which has resulted in the camera not attempting to refocus at all during shots 1-5. So, while shots 1-4 (not actually the beginning of the sequence), are acceptably in focus, this is only partly down to the camera choosing the right answer, and partly down to user settings preventing it trying to choose a different result.

Having shot it side-by-side with the Fujifilm X-T1, the K-3 ultimately fell short in comparison. This suggests that, while the K-3 is a potentially very capable camera and a big step forward for Pentax, the availability of lenses means it will rarely match the best of its peers across a broad range of action photography.

Continuous Shooting

There are three different continuous shooting speeds offered on the K-3: low, medium, and high. Pentax claims speeds of 3.0, 4.5, and 8.3 fps, respectively. To put the K-3 to the test, we used one of the fastest cards on the market: the SanDisk Extreme Pro UHS-3 SDXC card. The DNG Raw format was used for these tests.

High speed

Timing
Large/Fine JPEG
Raw
Raw+Fine JPEG
Frame rate 8.8 fps 8.9 fps 8.7 fps
Number of frames 68 shots 23 shots 24 shots
Buffer full rate 2.1 fps 2.2 fps 2.3 fps
Write complete * 9 secs 12 secs 19 secs
* Point at which photos can be taken at full speed

In what will be a common thread throughout this section, the K-3 exceeds its advertised numbers at high speed. The buffer is clearly quite large, given the number of photos that can be taken, but it takes quite a while for it to clear to the point where you can enter playback mode or start another full-speed burst.

In live view mode, the images displayed on the LCD are not in real-time, which is the case for all three continuous shooting speeds on the K-3 - intead the previously shot image is shown, which needs to be considered if you're hoping to pan or follow a subject.

Medium speed

Timing
Large/Fine JPEG
Raw
Raw+Fine JPEG
Frame rate 4.7 fps 4.6 fps 4.6 fps
Number of frames To card capacity 33 shots 34 shots
Buffer full rate N/A 2.0 fps 1.2 fps
Write complete * 5 secs 12 secs 23 secs
* Point at which photos can be taken at full speed

Low speed

Timing
Large/Fine JPEG
Raw
Raw+Fine JPEG
Frame rate 3.1 fps 3.1 fps 3.1 fps
Number of frames To card capacity 53 shots 39 shots
Buffer full rate N/A 1.3 fps 0.9 fps
Write complete * 3 secs 16 secs 22 secs
* Point at which photos can be taken at full speed

Battery Life

The K-3 uses the D-LI90 lithium-ion battery, which is used in several other Pentax-branded ILCs. This battery packs 14Wh of energy, which allows you to take 720 shots per charge (measured with the CIPA standard).

If that's not enough, one can purchase the D-BG5 battery grip, which can hold an additional D-LI90 or six AA batteries. Ricoh includes an external battery charger with the Pentax K-3, though it's very slow, requiring a whopping 6.5 hours to 'fill up' the D-LI90. Faster chargers are available from third parties.