Pentax K-3 Field Test Part III
Pentax K-3 Field Test Part III
Off to the races!
And finally, we come to the last section of my Field Test -- and my decision whether to buy the Pentax K-3 for myself. This section has been a while coming, for which I apologize -- every time I've gone to return to shooting and writing, something else has come up. Just one example: The subject I'd chosen for my planned autofocus and burst performance testing didn't cooperate. I'd been intending to shoot with the K-3 and K-5 side by side at a local drag racing strip to get a good sense for how the two cameras' performance compared. Come the day, heavy rains had flooded the pit lane and the track day was canceled.
Off to the races. I went in search of alternatives, and finally settled on a (somewhat) nearby go-kart racing track: Xtreme Racing Center in tourist-packed Pigeon Forge, Tennessee. Their karts are advertised as capable of 40 miles per hour, and while I think that's a bit optimistic -- I'd estimate closer to 25 mph -- they're still faster than the karts at other local tracks. I set myself up at the end of the third-longest straight on the track, with the carts coming towards me and the afternoon sun behind my shoulder.
I had two lenses with me for the test, both consumer-grade. (The only higher-end glass I had access to during my review were the Limited primes, none of which seemed a good choice for this shoot.) I opted for my 18-135mm F3.5-5.6 ED AL [IF] DC WR as an example of a lens with a built-in motor, and my 50-200mm F4-5.6 ED WR as one relying on screw drive. Since go-karts are fairly small subjects, I shot both lenses at their telephoto position almost the whole time, trying to draw the action a bit closer. With the karts passing just 20 feet from my position, it made for a reasonably challenging test of continuous autofocus. Both the K-3 and K-5 were configured with the default Focus Priority, rather than release priority, ensuring the shutter wouldn't be tripped unless the camera believed itself to be in focus.
The lab vs. the real world. Going into the shoot, I wasn't really sure what to expect. Our lab testing had suggested that the K-3's phase-detection autofocus system was just a little slower to lock focus than its older sibling, likely because of the significant increase in the number of focus points. I hadn't really noticed any difference shooting side by side with my K-5, though. (Live view autofocus is noticeably swifter on the K-3, but subjectively the PDAF performance felt about the same, to me.) The difference in burst performance between the two cameras was subtle but noticeable with focus locked, but would there be perhaps be a bigger difference when the cameras had to focus (and confirm a lock) between frames?
A surprising result. As it turned out, the answer was a little surprising: Both cameras turned in about the same performance, at least in terms of capture rate. Regardless of whether I was shooting the K-3 or K-5, if I held the shutter button down and tracked a kart coming towards me down the length of the straight, either camera hovered at right around five frames per second -- occasionally a little more or less, but only by a frame or so.
Of course, the K-3 has a lot more data to handle with each shot, so simply maintaining the same performance is in itself an upgrade, but it wasn't the added performance I was hoping for. Part of the reason for this is likely that, thanks to the bright sunlight, both cameras had to stop the lens down for each exposure, then open it back up to focus again before the next shot, limiting continuous burst speed. The fact that I was in focus priority rather than release priority -- ensuring the camera believed it had a lock, but also increasing the delay while a lock was achieved -- will have also slowed things down somewhat.
Once I got the photos on my laptop, though, it was clear that there was more to it than pure performance, anyway. I sat down and figured out how many of my photos -- I'd shot a few hundred, between both cameras -- were in focus, nearly in focus, or out of focus. Here the K-3 had a clear advantage, especially when shooting with the 18-135mm lens. With my K-5 body, perhaps one in four burst shots with that lens was significantly out of focus, and around a third of the remainder were close, but didn't quite nail the focus. With the Pentax K-3, though, about 90% of my shots with the 18-135mm lens were in focus or very close to it, and for the great majority of those, focus was spot-on.
Not just that, but the K-3 was also better able to handle focusing on the karts as they got closer to my shooting position, where greater focus adjustments were needed between shots. Pretty consistently, I got an extra couple of in-focus frames with the K-3 at the end of each burst, where the K-5 was returning significantly defocused frames. So -- a clear win for the Pentax K-3 here -- even if it wasn't shooting any faster than the K-5, it was giving me more useful results.
How about screw drive? But that was all with the faster-focusing DC motor. What about my 50-200mm lens? I switched and had another go. This was an even greater challenge, given both the stronger telephoto and the reliance on screw drive AF. Subjectively, I didn't notice any difference in screw-drive performance between the two, but once I got the photos on my laptop it was clear that the K-3 still had an advantage, although it too was struggling now. Both cameras had a hit rate of about 40%, with perhaps another 20% of shots nearly in focus, but the K-3 again managed to keep up with focusing after the K-5 had surrendered.
So no real advantage in accuracy for the K-3 with my screw-drive zoom lens, but nevertheless, it got more keeper shots than did my K-5. And while both cameras turned in about the same burst performance after accounting for focus adjustment, the Pentax K-3 also did a much better job following moving subjects with my 18-135mm DC lens. That's enough for me to add this to the list of things swaying me towards a K-3 purchase.
Multi Auto White Balance. This is a new addition to Pentax's DSLR line, inherited from new parent Ricoh's compact cameras. In response to a reader request, I gave it a quick test and found the effects, while probably still worthwhile with difficult subjects, to be quite subtle. Essentially, the function aims to balance complex lighting from multiple sources.
To test it out, I set up a couple of scenes (one closeup, one rather wider), each with uneven illumination from a cheap fluorescent bulb on the left, and an incandescent on the right. Sure enough, the color is balanced a little better across the exposure when using Multi Auto White Balance mode, rather than standard Auto white balance.
If you often use fill flash to balance out sunlight, it could be quite handy, and it's a nice touch that Pentax lets you specify that flash exposures use Multi Auto White Balance through the Custom menu. I don't shoot a lot of portraits, though, so seldom wanted to use the function myself.
Auto White Balance
Multi Auto White Balance
The images above were lit with a fluorescent bulb on the left, and an incandescent bulb on the right. Although the difference is subtle, Ricoh's Multi Auto White Balance (right) does handle the differing light sources better than does the standard Auto White Balance function (left).
Flash. Speaking of flash, although it's not significantly changed from that in my K-5, I did give flash -- both internal and external using my AF540FGZ strobe -- a try on the Pentax K-3. Exposures were reasonably accurate, and range of the internal strobe was plenty for typical subjects.
And it's nice that Pentax's wireless flash system works with the internal strobe as a master, although I find it can sometimes make my subjects blink. (It communicates via extra bursts of preflash rather than radio, and with some subjects that extra flashing before the exposure is a problem.)
Video capture. One last feature I wanted to test before rendering my verdict on the Pentax K-3 is its video feature set, which has had a pretty radical overhaul. I was pretty pleased to see that Pentax has added a headphone terminal to the K-3, but didn't find it terribly useful for two reasons.
Audio issues. For one thing, my headphones of choice -- Sennheiser PXC 450 cans -- didn't work with the K-3 at all, even though they're well below the impedance required by the K-3. I initially wondered if perhaps the camera body didn't work with TRS (tip, ring, sleeve) connectors, but when I tried some Sony earbuds with a TRS connector, they worked fine. So, too, did some HTC earbuds with a TRRS (tip, ring, ring, sleeve) connector.
More of a problem, though, was that there's no volume control for levels monitoring, just for the audio you're recording. Even with the recording levels set to their maximum, sound on the headphone jack in record mode was extremely faint -- far too quiet to have any idea if the audio was clipping. If I immediately played back the video I'd recorded and dialed up the playback volume control, I could easily hear the audio, so the camera body is quite capable of providing sufficient output -- it just doesn't do so in record mode. That meant the headphone jack was effectively useless to me except for reviewing videos I'd already recorded.
The Pentax K-3 no longer uses mechanical image stabilization during video capture. Instead, it uses digital stabilization, and while it does do a fairly impressive job of keeping the scene steady, it introduces some unattractive distortions and artifacts. Above, you can compare an unstabilized video (top) with a stabilized video (bottom).
Digital stabilization. In common with some rivals using sensor-shift stabilization -- and unlike its predecessors -- the Pentax K-3 disables its mechanical vibration reduction in movie mode, and relies solely on software stabilization. I gave it a try, but found it a rather poor substitute. Sure, the mechanical system made noises that could easily be heard in your movie's audio, but that could be resolved with an external microphone or audio recorder. Now, though, you're faced both with a focal length crop when using Movie SR, and with some artifacts of the purely software-based technique. These manifest themselves as distortions and a slight flickering that's very unsightly. I'd very much like to see the mechanical stabilization return as an option, if not the default.
Much better compression. That's the bad news. The good news is that the Pentax K-3 turfs its predecessors' inefficient Motion JPEG storage in favor of MPEG-4 AVC/H.264. In one fell swoop, file sizes have been slashed to just one-quarter of their previous levels, meaning I don't feel the need to reencode every single clip when I transfer it to my PC, to cut down on storage consumption.
Finally, autofocus! And that's not all. With the K-3, Pentax finally delivers what it first showed us five years ago with the original K-7: autofocus during video capture. Five years ago, that would've been big news. Now, it's a commonplace feature in the competition, yet Pentax shooters have had to do without.
Five years after we first saw a prototype Pentax DSLR with the ability to autofocus during capture, the company has finally made the functionality available at retail. It's not the fastest we've seen, somewhat prone to hunting, and only allows single AF operations, but it's still a whole lot better than nothing for consumer videographers.
Sadly, there's still no full-time autofocus in movie mode, just single-servo AF. It's also a bit slow, and the focus operation will sometimes rack all the way through the focus range, even though only a small change is needed. But for consumers, it's still better than nothing. (And if you don't like it, you don't have to use it, obviously.)
But on the plus side, with my 18-135mm DC lens there's relatively little noise recorded by the K-3's internal microphone. I'd imagine the same will be true with SDM lenses. You'll want to avoid screw-drive lenses, though, as the noise from these will very clearly be recorded.
Now this is cool: The Pentax K-3 can create 4K time-lapse videos in-camera. If you have a 4K display, you can see the full video (top) above. If you want to get an idea for the detail available in a 4K video but only have a Full HD display, watch a crop from the center of the video (bottom), saved at high quality using Adobe Premiere Pro CS5.5.
4K video (with a twist.) The Pentax K-3 also adds a new feature: the ability to record 4K video. There's a catch, though: It works only in interval mode, creating 4K time-lapse movies. And likely due to lack of processing power, they're saved using inefficient MotionJPEG compression, rather than H.264. That means truly epic file sizes -- around 50MB/second, or three gigabytes per minute of captured video. You can set an interval time of anywhere from two seconds to to one hour, and the maximum duration varies depending on the interval.
Although I don't have access to a 4K display on which to view the results, I did find the function rather fun. (And you can record at Full HD or HD resolution, if like me you lack a 4K display.) I used it to capture the somewhat cliched clouds-rolling-across-the-sky that most every timelapse video contains, as well as to shoot a video while driving around town at night. I did find myself wishing for a shorter interval time, though -- something on the order of 1/2 to one second would've been ideal.
Although I couldn't watch the full 4K video without access to a suitable display, I was able to crop just the center 1,920 x 1,280 pixels from a 4K clip, and use that to judge image quality. The verdict: 4K interval movies shot with the K-3 have very good resolution, which isn't surprising considering that they are essentially a sequence of downsampled still images. I'm not entirely convinced that there's a use-case for 4K video at home, with the display sizes and viewing distances in the typical living room, but if you're 4K-ready you'll likely enjoy fiddling with interval videos of your own.
You can combine the 4K interval video capture with manual exposure, if you wish. Here, I set the shutter speed for a reasonably sharp exposure from a moving car at night, then cranked the aperture wide open and the sensitivity to the maximum available, ISO 3,200.
My Flucard has the flu. One last thing I'd hoped to test with the Pentax K-3 was the Wi-Fi equipped Flucard accessory that enables remote live view and control of the camera. (And indeed, I did test it as far as was possible, using both Windows and Android devices.) Sadly, I had several problems that prevented a fuller review of the card's capabilities, and haven't been able to arrange a replacement of my review unit.
The remote live view feed has relatively little lag, and the Wi-Fi range is reasonably good compared to what I've seen from in-camera Wi-Fi on other models, as well as when using an Eye-Fi card. I was able to manage a distance of perhaps 50-60 feet with a wooden floor in between before the signal degraded to the point of unusability.
I was also able to adjust focus by clicking on a specific subject on the live view feed, and to trip the shutter remotely, taking a photo. And once photos were captured, I was able to view them at full resolution.
What I couldn't do, though, is adjust any other settings. The required controls couldn't be accessed on my laptop screen, and my ability to access the Flucard from any Android device was curtailed almost immediately that I first used the card.
Specifically, what happened is that I connected from a Google Nexus 10 tablet using Google's Chrome browser, then attempted to change the password as recommended on first connection. The password change didn't "stick", and on connecting to the card for a second time, I had to log back in with the default password. Shortly afterwards, I received a "400 Bad Request: Unsupported method" error, and from then on, the card simply didn't work with any Android device after the password was entered and accepted. My HTC One X+ smartphone, too, received the "400 Bad Request" error when logging in.
Subsequently, I was able to change the password from my PC, and this did "stick" -- the PC and Android devices had to use this new password to gain access -- but that didn't resolve the Android issue once signed in. Nor did the Reset Transfer Settings option in my K-3's menu system.
So... the long and the short of it is that the Flucard, as was, didn't work very well for me. While it showed promise, I also had stability issues. I'm hoping to revisit this accessory at a future date, asthe Pentax 645Z -- which I'll be reviewing -- uses the same card. Once I get the chance to test it properly, I'll either update this review or simply point it to the relevant section in our 645Z review. Watch this space!
My verdict. And now, the moment of decision that you've all been waiting for: Will I be buying the Pentax K-3? I skipped the previous geneation because it was too modest an update, but this time around, it's a clear step forward. Although the Pentax K-3 isn't quite everything I'd hoped -- most notably, it trails its burst-shooting performance claims -- it is certainly the most complete enthusiast DSLR I've ever used.
The already great ergonomics are now better than ever (save for the AE-Lock button location, anyway), and there's a quite significant step forwards in resolution, while retaining the previous camera's image quality in other areas. The new autofocus system stands head and shoulders above that in my K-5, and while I never had much trouble with that camera's metering system, the K-3 seems to perform a little better in this regard too. I love the fact that I can now shoot HDR without having to forgo shooting in raw mode. The addition of USB 3.0 connectivity and a second card slot are the icing on the cake.
I've started saving already, and I'll be buying a Pentax K-3 of my very own, just as soon as I can. Now, does anybody want to buy a second-hand Pentax K-5? ;-)