My K-7 Hands On Test and Findings

A full production unit was tested and my observations and findings are as follows:-

Pros:-

1. Autofocus is fast (not with SDM, which is slow by itself): With the 18-55 WR kit lens, it has been able to track moving vehicles running at speeds at about 30 to 40 km/hr at middle to close distances. Used in conjunction with the Continuous Hi frame advance mode, the shooting rate can be kept at about 3 to 4 fps and the AF hit rate was high (only the Central Single AF point was selected for the test);

2. Shutter is very quiet (but not so smooth). Mirror action is very gentle, too;

3. Shutter lag is shortest for a Pentax DSLR ever (similar to Canon 5D but obviously still not as short as the D300 or 50D etc., nothing comparable to the EOS 1 series DSLRs);

4. Rigid, lightweight and compact body, well balanced in hands, very comfortable to hold for the grip - very good ergonomics;

5. Flash raised at a high position and pop-up action is smooth and gentle. The closed RTF is rigid and there is no play;

6. Auto WB accuracy and reliability seems to have been improved - the best amongst any Pentax DSLRs ever.

Cons:-

1. It does not feel metalic for the metal alloy body - very different from the MZ-S. Actually, Pentax can make it plastic as the metal alloy shell actually doesn't feel cold nor it has any metal feel and texture. The K-m might even feel better, IMHO;

2. Shutter/mirror action is not silky smooth as it can feel and heard for the "rough" moving sound;

3. Viewfinder is near 100%, but it is not that bright (even a F2.8 prime was mounted) and the magnification is not that high - Still some kinds of tunnel viewing effect is there, nothing comparably to Full Frame DSLR viewfinders;

4. The 77-segment metering is too conservative and too sensitive: It counts too much for the bright area(s) anywhere in the frame. Whenever there are highlights, even those areas are small and they are off-centre, the picture will be underexposed! (The K-m 16 segment metering works better!);

5. AF seems not to be as accurate as it should be, especially for an APS-C DSLR with that pixel count, which might contribute to the soft images obtained as resolution of even sharp optics are decreased (but not the main reason, I suspect, read more below);

6. The expensive WR kit lens is of low optical performance. Blurs are too obvious at corners near wide opened (not extreme corners). The image centre is not tack sharp neither, even at closed down;

7. Images are blurry, in-camera Jpegs are very soft throughout the frame;

8. Images are noisy, at least two stops disadvantages than the 5D and surely will be even more if it is compared to the current mainstream Full Frame DSLRs like the D700 and 5D MkII;

9. LCD display is always turned on "to ease operations" by default which is considered unnecessary but just a waste of battery power, especially considered this is an advanced classed camera, not an entry level one without top monochrome status LCD;

10. Mode dial unlock button hinder operations and is considered unnecessary. The button is small and turning mode dial whilst pressing it is somehow difficult;

11. Colours are a bit too saturated and somehow look unnatural than real even for the Natural image mode used, might be a result of underexposure which produces more saturated images. But it also seems that the Samsung CMOS sensor produces images that look less favourable than those Sony or Canon sensors (that can be somehow subjective anyway).

12. Video function is buggy, there is sudden flash (drastic change in exposure) for a short clip I recorded without any good reason. CDAF under LiveView / Video recording is damn slooow. It hunted a lot very slowly and sometimes even missed the focus and needed to search again.

13. The video quality of the movie is just so-so. I shot in 720p mode. What I have to say it is not better than the 848x480 H.264 Movie from my Casio Exlim S10 and I would say the playback of the S10 movies could just be smoother.

I suspect the reason for the Soft JPEGs is yet once again not because of inadequate processing power of the camera, but inherently a noisy sensor (the new Samsung one) of which the images it produces are needed for in-camera noise reduction and as a result the images are made soft. This is the same as what Photography BLOG's test shows: You could get Either Soft JPEGs Or Noisy RAWs (with details but together with noises), even at ISO 200, with the K-7. Trade-off and compromise have to be taken.

Okay, last but not least, here comes some sample images (100% Crops from in-camera Jpegs, 5D's ones further upscaled by 115% (in area) in IrfanViewer for almost the same viewing sizes, saved as lossless 24-bit BMPs):-

Camera / ISO	Scene 1 Image Centre	Scene 1 Image Corner
K-7 / ISO 100
5D / ISO 100
K-7 / ISO 200
5D / ISO 200
K-7 / ISO 400
5D / ISO 400
K-7 / ISO 800
5D / ISO 800
K-7 / ISO 1600
5D / ISO 1600

Camera / ISO	Scene 2 Image Centre	Scene 2 Image Corner
K-7 / ISO 100
5D / ISO 100
K-7 / ISO 200
5D / ISO 200
K-7 / ISO 400
5D / ISO 400
K-7 / ISO 800
5D / ISO 800
K-7 / ISO 1600
5D / ISO 1600

Remarks on the test conditions:-

K-7: Factory Default Settings, "Natural" Custom Image, Matrix Metering (No Exposure Compensation), ISO NR starts at ISO 800, Medium setting, AF-S with Central Single AF point selected, no frame re-composition after AF achieved, FA 28 (Photodo MTF Score 3.8) @ f/8;

5D: Factory Default Settings, "Faithful" Picture Style, Matrix Metering (No Exposure Compensation), no ISO NR, MF with Central AF point selected, no frame re-composition after focusing point measured, FA 43 (Photodo MTF Score 3.6) @ f/8.

Field/Angle of View of both Cameras are the Same.

I have repeated similar tests with my new K-m and the results against my 5D is not of such a huge difference! It verifies that the K-m produces sharper images with more details and less noise even for the same viewing/zoom-in sizes, with more favourable colour response, too. My Olive DAL kit lens is also way sharper than that WR version which triple the price of an ordinary DAL 18-55, too.

Btw, I shall do a more systematic shootout test and post the results later on - K-m Vs 5D, when I have the time. Update: Here is the test - K-m Vs 5D Pixel Peeping!