Pentax K-5 Hands-on preview

Pentax K-5 with vertical grip D-BG4 andlens DA* 55/1.4

Photokina2010, the world's largest and most influencialphotographic fair, has just closed its doors last Sunday.Monday, the day before Photokina officilally openedits doors in Cologne, members of the international presshad a chance for a sneak preview and Pentax managed toshow prototypes with firmware 0.30 of their new PentaxK-5 camera in the afternoon. I later consolidated my impressionsat a further visit. However, any sample shots shown hereare from Monday where all over Photokina, journalists werehappily forcing their SD cards into whatever camera whichappeared to be new. Pentax then better controlled it duringthe official show but, e.g., Nikon simply did not haveenough staff to prevent people from putting their SD cardinto the new D7000. Me being no exception :)

So, I put the following disclaimer upfront: Any sample shotfrom the K-5 found on the internet today is not endorsedby Pentax and no conclusions must be drawn from it. It isfrom a prototype camera with preliminary firmware flewn intoCologne by Hoya the day of the announcement (2010 September,20).

These are my preliminary impressions from my short testdrive:

1. The announcement

The Pentax announcement was a bit unlucky.Corporate Pentax was late for the event and announcementshave beenfull of little mistakes, contradictions and appeared to berushed out. This did not exactly help in creating any enthusiasmfor the K-5 among the press. Pentax missed the opportunityto hold a press conference on Monday or Tuesday too. Maybewith an invited speaker from Tokyo and showing some pridein the product. What a pitty. So, the general reception ofthe K-5 is that ofa K-7Superora K-7mkII.

So, now it is important to have a closer lookat the new Pentax flagship model which is going to hitthe marketsoon. Maybe, the camera is able to create the excitementall by itself which is what happened with the popular K-xcamera.

2. Market positioning

The Pentax K-5 is a semi-professional or enthusiast-levelSLR camera and competes with its peers from Nikon and Canon,like the Nikon D7000, Nikon D300s and Canon 7D. Esp. theD7000 looks like an identical copy of the K-5, and if theK-5 is just a refreshed K-7, then it would be clear who copiedwho ;)

It is exciting that the K-5 is defending Pentax' marketposition in the market spot pretty much created by the K-7(robust and compact and suitable for professional use). Aspot which obviously attracted Canon and Nikon and now appearsto be the most interesting market spot above $1000. Accordingto a recent DPReview poll, D7000, Fuji X100 and K-5 are themost exciting Photokina news and then K-5 may be just arefreshed K-7.

3. Availability and price

The MSRP of the K-5 body is $1,399.- USD. Pentax is confusedenough to not acknowledge this fact with their internationalcorporationand to publish list prices as high as $1,599.- USD. I understandthat Pentax was caught with their pants down when Nikon launchedtheir D7000 body at $1,199.- USD just days before Pentaxwas going to announce their K-5 at a price, I am pretty sureabout, closer to that of the D300s or 7D. But hey, if PentaxGermany can respond in a timely fashion, why not the corporation?

The German MSRP as announced at Photokina is 1,449.- € incl.19% VAT and e.g., French online store digit-photo.fr alreadylists the K-5 body at 1,249.- € TTC and the D7000 body at1,169.- € TTC, only 80 € less.

Both, K-5 and D7000 are announced to ship in October 2010and you read it here first: when they hit the market, theywill sell at equal price. And both cameras should sellvery well, according from the first feed back I see.

4. Specs and first touch

That's easy: it looks and feels like a K-7. For many, thismeans they fall in love with the camera immediately. Forsome, it means the body is a bit small and the grip thenmay helpor not. I tried the D7000. I am partial, I admit (for a subjectivequestion like this). But the D7000 could not create the samekind of emotion whenI triedit. At high ISO, one sees the lower noise floor in zoomedview, with the 18-55 kt lens, one feels the more aggressiveAF, but other than that, it feels "K-7". With one notableexception: contrast AF. It almost feels like Pentax leapfroggedCanon and Nikon.

I hate redundancy on the web. So, rather than repeatingwhat is easily found elsewhere, I refer to the sources forthe specs and walk-thru:

• Specifications and original press release from Pentax Japan (English version)
• User manual (when it comes out, unavailble ATM)
• DPReview walk thru menues and preview

Therefore, here I will focus on exclusive content maybenot found elsewhere. If you aren't familiar yet with theK-7 or Pentax cameras, you may want to start readingwith any of the links above.

5. Image quality

Currently, everything I've seen let me hope that theimage quality turns out to be impressive when studiedin more detail. As far as I know, Pentax uses the SonyExmor HD 16.2 MP APSC CMOS sensor with integrated column-parallelA/D-converters. Which may be the same as in the NikonD7000. Pentax confirms using a sensor with embedded A/D-converterand that's all it needs ;)

5.1 High ISO noise

Pentax K-5 firmware 0.30 ISO 6400 rawimage (Lightroom 2 default development)

[Download link to the -> 30MBDNG raw file]

Because the interest is huge and because high ISO JPEGsamples are already floating the web, I decided to publishone ISO 6400 RAW sample. I took it Monday when all journalistsplayed their card party but I respect Pentax interestto keep image files under closure. The RAW sample isless affected by changes to the firmware which is whyit may be in the interest of Pentax to have at leastone RAW if JPEGs are already in the wild. The high ISOJPGsamplesbenchmarkthe in-camera v0.30 JPG engine first of all and shouldnot be compared to D7000 JPGs.My RAW may provide an alternative.

The RAW published here has been already developed usingvarious raw converters and NR settings, esp. in a veryvivid thread at pentaxforums.com and by GordonGoodman at dpreview.com. The general conclusion Idraw from the discussion is:

1. K-5 allows it to use AUTO-ISO ranges of up to ISO 6400 without hesitation (if it used to be 1600 like on my K-7). Some noise is visible, but is well treated by NR.
2. At ISO 6400, the advantage wrt K-7 may be about 2 stops (almost). As I refuse to run quantitative lab tests yet, this is impossible to say exactly. This has to await until Pentax thinks it makes sense to run a lab test.
3. Looking at Pentax K-x and Nikon D7000 and D5000, the performance looks similiar. The K-x has less pixels to be taken into account too. For me, it is currently difficult and too early to rank the three cameras.
4. At lower ISO settings like ISO 1600, the advantage wrt K-7 is less. As to be expected if the progress comes from lower read-out noise rather than a higher quantum efficiency. ISO 400 shooters may be able to increase to ISO 800 only. Physics can't be beaten.

So, a first assessment is possible: The high ISO noiseof the K-5 is within the best of its class.

Gordon B. Good went a step further and analyzedthe raw histogram in the DNG (cf. link above). Hisconclusion (which can only be preliminary) is thatpixel read-out noise (dark noise) is even less thanwith the Pentax K-x which excels in this regard.He found too that ISO 6400 is created from an ISO1600 capture then boosting EV by +2 in camera.

5.2 Dynamic range, colors, pattern noise, AA filter,resolution

To the right is an unprocessed image out of thePentax K-5 camera at ISO 100. So yes, the cameraworks at ISO 100 :)

If Gordon's preliminary analysis holds true (helooked at an ISO 1600 PEF too), then a pixel's fullwell capacity is on par with a K-x and the minimumrecordable luminosity of a K-5 pixelshould beat the value of 0.021% forthe K-x (dynamic range as defined by an SNR of 0dB).E.g., a value of 0.017%(corresponding to Gordon's 2/3 black noise remark)at K-5's resolution would correspondto a dynamic rangeof 13.0 EV in DxO's terms. That would bea stunning value for an APS-C camera. 0.017% would use12.5 bits (out of 14) too which makes the 14 bitformatnota complete loss,compared to 12 bits which is an industry's first.If this analysis can beverified to hold true after production cameras ship.

Update: more recent remarks (made after analyzingthe ISO 80 RAW black noise) claim dynamic range tobe even better than 13EV. I'll soon edit this sentenceto reflect a common preliminary assessment by Gordonand myself.

I did not evaluate color rendition, pattern noise,AA filter, and resolution.

5.3 Shutter blur

I'll take a detailed look at that as soon as I can.So, while I refuse to comment on the issue for thetime being, I can still quote others who tried tofind out,can't I? Ok, several other tried to see the shutterblur on the K-5 (in magnified playback) and told meto have a good feelingbecause they couldn't find it. Would they have seenit with a K-7? I don't know. But it is a positivesign andtherefore, let's take it as such.

6. Autofocus performance

6.1 Phase AF

The K-5 uses a similiar AF module (SAFOX IX+) to theK-7 (SAFOX VIII+), but with larger AF lenses and/or moresensitive AF sensors. It has 11 AF sensors, 9 which ofare cross type. All sensors are f/5.6, so it has no f/2.8dual type center sensor (as far as Pentax Germany knows).

The AF performance in good office space lighting withthe 18-55 mm screw drive kit lens is outstanding: Itmakes anoise which is so short that it sounds a bit like theAF hit the limiter. The algorithm hasthe SAFOX gene: if the focus has to travel (isn't closealready), then the lens may be refocussed a second (orin rare cases) even a third time. But it is so fast oneis not necessarily aware of it. I think it is a goodthing that the Pentax AF still verifies focus "onarrival"and doesn't trust it blindly which may not be as accurate.But it can increase shutter delay. I didn't check ifthe new shutter priority modes in the K-5 AF settingsbring a change in this respect.

I heard reports (caution: hearsay!) that the new DA35/2.4 lens has a very remarkable high focus speed aswell.

Wieland (aka blende8 from pentaxforums.com) and myselfbrought two DA* 16-50/2.8 SDM lenses and compared headto head,K-5 vs. K-7. It was difficult to tell ifthe K-5 was faster. The SDM speed is the same as forthe K-7, limited by the lens motor. Wieland(who hasit on video) and myself verified this. Lens travel wasthe dominating part in focus time. But for short lenstravel (prefocus) Igot the impression that the remaining focus time forthe K-5 is significantly smaller. This must be benchmarkedreally. Moreover, Pentax Japan specifies autofocus operationto be possible at -1 EV. That may be false information,requiring the AF assist beam or be an impressive figure.

I didn't test continous autofocus. Accordingto pentaxeros.com,the throughput in usable photos in AF.C may havedoubled, relative to a K-7. Overall, my preliminaryimpression is that the progress made in autofocusperformance from the K20D to the K-7,and from the K-7 to the K-5, may be comparable. Thismay or may not suffice to compete with a D7000 or 7Din this department.

I made a short test if the K-5 tries to track AF points,i.e., if the K-5 tries to predict the hop of a subjectfrom its active AF point to the next. My preliminaryimpression is thattheK-5 doesn'tdo that and it would be difficult with 11 AF points anyway.

I tried the AF in the new Nikon D7000 and the new SonySLT55 as well. I wasn't overwhelmed by the D7000 (thelens didn't have a fast focus though) and the SLT55was cool. Sony had a half pipe with skate boarders andbikersand plenty of action. The SLT55 AF.C performed considerablyfaster than my K-7 (I loved to have a K-5 there butPentax staff would have tortured me afterwards ;) ).But the K-7didn't fail entirely and the SLT55 didn't perform flawlessly.In the end, the number of usable photos per half pipeturn may have been twice in Sony's advantage. Twice"only", maybe not out of reach for a K-5.

6.2 Contrast AF

Asmall video of the Pentax K-5 contrastAF at work
© 2010 by YouTube user SmalltimR

The contrast AF seems to be a complete rewrite. Whenyou half press the shutter (or whatever activates AF)the live view image zooms into the appropriate focusrectangle and confirms with green border and beep, thenzooms out. The entire process takes about 1 s. Comparethat to phase detect AF which typically took about 0.4s and then contrast AF is only slower by a factor 2-3.

It's even possible to switch contrast AF into AF.C.But the performance is poor: it only refocusses every4 secondsor so when the change in contrast is obvious. But thisis for a prototype firmware. Pentax may still work onthis.

Another positive surprise is that contrast AF doesn'tpump a lot (anymore).

Overall, my preliminary impression is that the K-5 (andK-r too) are the first Pentax SLR cameras where contrastAF is really usable. At least for subjects which don'tmove fast.

The new contrast AF should now be fast enough to allowAF during movie recording. But firmware 0.3 shows nosigns of such a feature: pressing AF in video has noeffect and there is no menu option for it. Pentax Germanysays that they received no information regarding AF duringvideo and it's not on the Japanese spec sheet either.

7. Live view

Except AF operation, LV looks the same as with the K-7.In manual focus mode, there still are additional 8xand 10x magnification steps. However, unlike with theK-7,the image quality doesn't seem to increase in the additionalmagnification steps: diagonal edges had a very visiblestaircase effect which was symmetric. The K-7 reducedthe effect in one direction with the additionalmagnification steps at the expense of a slower framerate. Overall, it's probably about the same to manualfocus in live view. The resolution still seems to belimited to that of HD video. There is no auto-zoom inmanualfocus, e.g. by half pressing theshutter.

8. Video

One frame from a K-5 720p HDvideo (click for 100% size)

The K-5 sports 1080p HD video at 25fps (up to 25min or 4GB). And smaller sizes at 25fps and 30fps.So, thesensor's read-out rate or processor performance arenot sufficient for 1080p at 30fps and only Pentaxknows whythen they didn't incorporate a 24fps mode.

Video qualitylooks good but that's only a preliminary statement.I can see traces of pixel artefacts in the frameas are typical for all current video SLR camerasbecause the sensors must be subsampled. Note thatthe quality in production cameras can even be substantiallybetter.

There are no manual controls (for shutter and ISO, thatis) in movie recording, just manual aperture, exposurecompensation and lock. The audio recording in the K-7was fixed gain (manual only so to speak) and it may bethe same for the K-5. I did not check it.

Pentax Germany had requested manual video controls inJapan. Obviously, it did not help much.

The K-5 has a programmable RAW button (Fx). The pressmaterial mentions that the video mode can be activatedthis way. Which is a great thing because turning themode dial every time is awkward. Moreover, one couldtheoreticallyswitch the mode dial to "M"and activate video,hoping that ISO and shutter controls are enabled (hopedieslast). But I couldn't test because the video functionfor the Fx button was still missing in the firmware.I hope the press material is right about the possibilityto map Fx to video. Because turning the locked modedial every time is awkward.

9. HDR

In-camera HDR (high dynamic range) photography waspioneered by Pentax with the K-7. But they ommittedthe feature to align images in-camera too which madethe feature obsolete without a tripod. I demonstratedina LumoLabs article right here that the K-7 hardwareshould be capable enough to do the alignment in-bodyaswell. Pentax ignored me but they did not ignore Sonywho then implemented the full feature in their NEXmirrorless cameras. So, the K-5 does now supportin-camera image alignment too and additionally providesadditional rendering parameters. Below is my free-handHDR example using the various settings:

(click on the thumbnails for full size samples,appropriate to inspect alignment quality -- hoverover a thumbnail to see the HDR setting)

Most of the time, the alignment works well and theHDR feature is a powerful extension of a photographersarsenal. The processing time is acceptable. However,the alignment isn't flawless (e.g., look at the "strong2" sample in full size). Maybe, Pentax wants to licensethe LumoLabs algorithm? ;) Another disappointmentis that the processor still doesn't write a linearDNG raw file without any tone mapping applied. Andeventually, the alignment could have been availablein multiexposure mode too but isn't. Sony calls thelatter feature "steady night shot".

Nevertheless, Pentax and Sony are pioneering newcreative possiblities with their cameras and thisgives them an innovative edge wrt Nikon and Canon.

10. Many more things ...

The sections below are answers to questions I receivedvia comments to my blog or forum posts.

10.1 RAW file bit depth and buffer size

The K-5 has an increased bit depth (from its A/D converterto RAW storage/processing) and increased number ofpixels. An uncompressed RAW frame is 28.3 MBor 1.31x as much dataas it was on a K-7. Which would reduce the buffer capacityto 10 raw frames (down from 14-15 on the K-7) somewhatin line with the Pentax specification of 8 raw frames.At7fps, this means that thebuffer fillsin 1.1 s (down from 2.7 s on the K-7). I confirmedby trying it out that the hi speed burst lasts for 8raw frames. This means that the overall buffer size andwrite speed to SD card have remained the same.

The fps or bit depth cannot be configured and Pentaxhas no plans to release a mid speed burst mode wherethebufferfillsmore slowly. Pentax thinks that action shotbursts are done in JPG and RAW bursts are mainly forbracketing and similiar applications.

It wasn't specified but the K-7 had an easter egg: Witha high speed SD card, low quality JPG setting and allsettings manual, the K-7 could burst at 5.2 fps untilcard full! The K-5 cannot do it anymore. Even at lowestquality settings and with a Sandisk Extreme III card,the burst breaks slow after about 29 JPGs or 4 seconds.

10.2 Prime II processor

The press material creates the impression that the PrimeII image processor is changed. This isn't true. The DRAMmemory (2GB AFAIK), the clock speed and processor itself(Prime II) are the same as with the K-7. Maybe, someof the algorithms are different (more pixels, 14 ratherthan 12 Bits) but that does not qualify the image processorto be considered "upgraded".

I compared the processing speed for lens corrections(CA+distortion). The image on the display of the K-7appears just a second or so before that on the K-5. Whichis exactly the expected difference with an extra of 11%more pixels to process on the K-5. For an image processorof equal speed that is.

10.3 Shake reduction

I could find no differences in the shake reduction feature.The specification didn't change either. However, thereis a small change I noticed nevertheless: When you halfpress the shutter, the SR needs about a second to getready.Aboutthe same for K-7 and K-5. But if you unpress the shutter,the SR stays ready until the VF INFO row is switchedoff a couple of seconds later. With the K-7, it didn'tand another half press always had the delay for SR toget ready again. This may be a huge difference in certainsituations. If it wasn't simply a display bug, then ornow.

10.4 Top LCD light

The top LCD illumination is either on or off (dependingon a menu setting) which is like the K-7 and unlike K20D.Most people like the K20D method better.

10.5 Bulb mode and DFS

DFS (Dark frame subtraction aka Long exposure noisereduction) can now be switched off even for bulb mode(exposures longer than 30s). I made a test with a 66sexposure atISO 1600 and the image displayed immediately and thecamera was ready again -- without a delay (firmware 0.30).ISO 1600 is the highest ISO setting for bulb mode. Thisis no limitation for raw shooting though because higherISO are made in camera from ISO 1600 by exposure boostpost processing; something easily done in the raw converteras well.

10.6 Tethering

The K-5 ships without a true tethering solution. AnEyiFi card, remote trigger and HDMI connection are theonly means available now. Pentax Japan confirmed to PentaxGermanythat theirrequest for a tethering solution has been acknowledgedwithout saying if Pentax is going to release a tetheringsolution or not.

10.7 Flash sync time

The K-5 has a flash sync time of 1/180s. The rumor abouta sync time of 1/250s, maybe in conjunction with fasterPentax flashes to be released soon, has been disclaimedby Pentax Germany.

10.8 SXHC support

The K-5 will receive SXHC support via a firmware update.This has been made public by the Pentax Imaging USA website.

10.9 Intervall shooting

The K-5 allows to select the number of intervall shotsto be between 1 and 999 (up from 99 on the K-7). Theselection is by digit.

10.10 User modes

The K-5 now has 5 user modes. However, there isonly one user mode position on the mode dial andyou'll have to select which one of five modes touse.

10.11 Level meter

The K-5 has a level meter for two axes (roll and pitch).The K-7 has roll only. However, the K-5 only shows theroll level in the viewfinder and the top LCD. The rearLCD shows both.

10.12 Exposure bracketing with MLU

The K-5's Fx button can be configured to enable exposurebracketing to later produce HDR images -- with a singlebutton press. However, on the K-5 and K-7, mirror lock-up(MLU) and exposure bracketing are mutually exclusivedrive modes. Theoretically, the Fx button solution couldsolve this. However, an Fx button press selects the exposurebracketing drive mode and thereby deselects the MLU drivemode if it was selected before. No luck.

Nevertheless, exposure bracketing with mirror lockupis available on both the K-5 and K-7: use exposure bracketingfrom within live view (LV). Because of the improved LVfocus performance this is a serious option with the K-5actually. The Nikon D7000 calls it "quiet mode" and hasit as a new feature.

10.13 Shutter noise

I compared a K-5 and K-7 side by side. They haveidentical shutter noise (at least without MLU andwith no burst -- as I didn't check for those). Whichmeans that K-5 continues to be state-of-the-art inthis field.

Conclusion

Pentax K-5, the engineering masterpiece (photo © 2010 Falk Lumo)
Photokina 2010 exhibit, went back to Japan well beforethe show ended.

This concludes my hands-on preview report of the PentaxK-5 semi-professional APS-C SLR camera. It combines whatmay currently be the best APS-C sensor with a camerawhich has almost professional grade attributes but ina package which pleases the outdoor and street as wellas the enthusiast photographer. It shares the genes ofthe applauded K-7 and improves on the threepoints which could have been considered its only weakness:ISO performance, autofocus speed, uncommon sources forblur (maybe).

The K-5 has only two rivals which however, can't beatit (except maybe for very fast action and long lenses):the very similiar Nikon D7000 and the bulkier and moreexpensiveCanon7D. If you now consider that only Pentax managed to packa good sensor-shift stabilizer into the body which isthemost elegant, sturdy, lightest and smallest of the threeand is done by the smallest company of the three, thentheK-5 is an engineering gem and a photographer's pleasure.No wonder it is the only SLR to receive this year's PhotokinaStar 2010.