(Japanese, Google English Translation Here)
As usual, we have full ISO samples, with different NR settings. And, there are also images taken for the same scene with different colour modes (Custom Image in Pentax' term). There are also some other real-life samples. The most important and meaningful thing is that they are all original in full size, directly out of the camera!
Btw, I don't see the super 14EV high range DR in action, at least the highlight is clipped as usual, as for most digital SLRs:-
http://dc.watch.impress.co.jp/img/dcw/docs/405/733/html/122.jpg.html
In contrast, I recall that my 5D still preserves better highlights than the above K-5 sample, see:-
http://ricehigh.blogspot.com/2009/03/some-new-macro-photos-to-share-dfa-100.html
My above bee on white flower photos under the bright sunlight are very similar to the situation of that K-5 sample, as it can be easily seen, the highlights and details of mine are just preserved better.
Besides, I also see the new King of Purple Fringing, just see this sample:-
http://dc.watch.impress.co.jp/img/dcw/docs/405/733/094.jpg
(Warning: Large File in 12MB!)
For your easy viewing, I provide a full-sized crop of the above sample:-

Finally, there are also two original video clip samples that can be downloaded. I do not feel good for the video quality nor the video efficiency. E.g., this first train bridge video is about 300MB for only 30 seconds. As for most of the time there is nothing in action within the frame and even though there are some, the moving portion is yet small, I am quite sure if there is some kind of video motion compression, the file could be drastically decreased.
Read Also:-
K-7 Wins Over K-5 in IQ (Specifically)!
Ten Reasons for Thinking Twice before Decided on a K-5..
Michael · 751 weeks ago
RiceHigh 110p · 751 weeks ago
http://ricehigh.blogspot.com/2010/03/standard-zoo...
jaad75 · 751 weeks ago
ArcticExposure 45p · 751 weeks ago
RiceHigh 110p · 751 weeks ago
Anu · 751 weeks ago
I do suspect that Pentax autoexposure programs don't take advantage of the high DR to preserve the highlights, as that would fool several pixelpeeping review sites into thinking that the sensor is noisier than what it is (ie, it exposes roughly the same the competition does, instead of exposing less, thus preserving the highlights, and having a curve 'pushing' the midtones and shadows to compensate for the lesser exposure - this would take advantage of the lower read noise of the sensor and still offer as good midtones&shadows the competition does, instead of offering better ones, like now, when exposed identicallly).
Anyhow, I am sure the camera has an option to expand the DR of the jpgs (as my old K20D has it), and raw-shooters ought to understand to expose as required instead of blaming the camera for underperforming in unproficient hands.
RiceHigh 110p · 751 weeks ago
jaad75 · 751 weeks ago
K-5 headroom is about 4EV above middle grey. That means you can use spot metering on hightlights and use +4EV compensation, to preserve the most information in the shadows. It's just your stupid idea that camera should meter exactly where middle grey is. Lightmeter is just a tool and you decide how to use it.
And a proof? Ok. here you have +4EV from a screen background (BTW, the scene contrast is 11.3EV and Blacks are at 10 if you know what that means): http://i140.photobucket.com/albums/r25/jaad75/K-5...
satriak · 751 weeks ago
The high DR of K-5 sensor shall be obtained from RAW instead.
bandymelis · 751 weeks ago
Another thing - chromatic abberations. These are do to the 18-55 lens used for testing. Not a good lens to decide about CA's, as it produces a lot of those. Take a standard 50 mm lens and then check for CA's.
Otherwise, the results seem fine.
Wacko · 751 weeks ago
Dick · 751 weeks ago
USUK SendMoney · 751 weeks ago