In the current issue of the most popular HK Digital Camera magazine (printed), the DiGi Bi-weekly, there is a full shootout on the above DSLRs:-
The same lens model was used to do all the tests, it is the Tamron A18 (18-250mm/3.5-6.3 Di II).
For specs comparison, the K-7 wins in size and weight (well, we all knew), the 50D wins in pixel count and ISO range, the D300 wins in AF system and the viewfinder. The A700 wins nothing on paper.
Btw, here is a concise summary on specs differences of the four APS-C DSLRs:-
|Body||Launch Date||Pixel Count||ISO Speed Range (Normal / Expansion)||Weight (with Battery)||Size||Viewfinder||AF System (Total Number of Points / Number of Crossed Points)||X-Sync|
|K-7||2009-5||14.6M||ISO 100-3200 / 6400||750g||131 x 97 x 73mm||100%, 0.92X||11 / 9||1/180s|
|50D||2008-8||15.1M|| ISO 100-3200 / 6400-12800||822g||146 x 108 x 74mm||95%, 0.95X||9 / 9||1/250s|
|D300||2007-8||12.3M||ISO 100-3200 / 6400||903g||147 x 114 x 74mm||100%, 0.94X||51 / 15||1/250s|
|A700||2007-9||12.3M||ISO 100-3200 / 6400||768g||142 x 105 x 80mm||95%, 0.9X||11 / 1||1/250s|
Well, the lab tests and measurbations begin..
1. Dynamic Range Test
Test description and methodology: All cameras set at ISO 200, finest JPEG, on tripod, metering by external incident light meter, Av=f/8, shot target was Kodak Gray Scale, final images analysed by Zoner Photo Studio 11 (clipped highlights and shadows will be shown). Exposures from -2 EV to +2 EV in 1 EV step were made with with highlight and/or shadow optimisation turned off and on.
An extracted summary on the results in the Numbers of Steps Clipped for the Highlights / Shadows are as follows:-
|Camera||-2 EV||0 EV||+2EV|
|50D (Default)||1 / 7||1 / 6||3 / 0|
|50D (Highlight Optimisation On)||1 / 6||1 / 6||1 / 0|
|50D (Highlight Optimisation On with Auto Leveling)|
Note: Very low contrast images were produced.
|0 / 0||0 / 0||0 / 0|
|D300 (Default)||0 / 0||0 / 0||0 / 0|
|Active D-Lighting On: Highest||1 / 6||1 / 6||0 / 0|
|K-7 (Default)||1 / 0||1 / 0||1 / 0|
|K-7 (Highlight Optimisation On)||1 / 7||1 / 0||1 / 0|
|K-7 (Highlight Optimisation On + Shadow Compensation)|
Note: All images had become rather Yellowish.
|0 / 0||0 / 0||0 / 0|
|A700 (Default)||1 / 0||1 / 0||1 / 0|
|A700 (D.R.O Auto On)||1 / 0||1 / 0||1 / 0|
|A700 (D.R.O Advance/Manual +5)|
Note: There is no significant effect regardless of the settings!
|1 / 0||1 / 0||1 / 0|
The editor further particularly notes that the K-7 images had large amount of noise appears when the compensation functions were turned on, but the DR results could be even worse!
2. Noise Test
Test description and methodology: The same set of test objects were shot under studio environment with different ISO speeds available for the cameras. Same aperture and exposure values are used; same WB, highest resolution and finest JPEG; All built-in compensation functions for distortion and vignetting etc. were disabled; other settings remained at factory defaults.
Cropped portions of the images obtained at the same viewing size are printed.
Results: For what I can see, the K-7's ISO 800 is a bit worse than the ISO 1600 of the 50D and the D300 (and that the 50D images seem slightly better). And, the ISO 800 images of the K-7 are still a bit worse than those of the ISO 800 of the A700. Of course, with the NR turned on, noise are suppressed at the cost of less details.
3. Auto White Balance Test
Test description and methodology: The GretagMacbeth ColorChecker was shot in Auto WB mode and ISO 200 under daylight, fluorescent and tungsten as well as an accurate 4000K light sources. The images were then imported into the Imatest and the colour accuracies were checked.
Some of the more important results on the Accuracy of Reproduced Skin Tones / Greys are summarised as follows:-
|50D||Good / Good||Good / Good||Serious Orange Cast / Serious Orange Cast||Better|
|D300||Good / Good||Orange Cast / Yellow Cast||Serious Orange Cast / Serious Orange Cast||Good|
|K-7||Magenta Cast / Blue Cast||Good / Good||Serious Purple Cast / Serious Orange Cast||Good|
|A700||Slight Magenta Cast / Slight Blue Cast||Slight Orange Cast / Good||Serious Orange Cast / Serious Yellow Cast||Fair|
4. Resolution Test
Test description and methodology: The ISO 12233 Resolution Test Chart as well as a real scene were shot; All cameras set at f/8; ISO 100; Highest Resolution; Finest JPEG; Cameras mounted on Tripod; SFR (Spatial Frequency Response) at MTF50 by Imatest were measured.
Crops of the resolution chart test shots and the real scene images are also printed (in the same viewing size).
As for the Imatest SFR figures, here they are:-
|Camera||Resolution (Image Centre)||Resolution (Image Corner)|
|50D||2,324 LW/PH (Line Widths Per Picture Height)||2,168 LW/PH|
|D300||1,921 LW/PH||1,935 LW/PH|
|K-7||1,893 LW/PH||1,856 LW/PH|
|A700||1,958 LW/PH||1,904 LW/PH|
The K-7, despite of having a pixel count close to that of the 50D, it loses "hands up" for both the resolutions at image centre and image corner. And, the real scene sample photo crops are obviously "soft", and so do the test chart crops.
Well, the above is *exactly* what I found previously for my own hands-on test of the K-7 (against 5D), the K-7 is just producing very soft and low resolution JPEGs even at ISO 100 (and noise control is not good as well). Too bad.. :-(
My bet is that soon when the Dpreview (DPR) resolution test and review is out, once again the K-7 will score very low in the IQ/PQ department and it will be commented as a "Soft JPEG" machine again, as usual, for what Pentax DSLRs have long been renowned for!
In fact, for the low PQ/IQ of the K-7, it is really really sad for all Pentaxians (and it actually becomes hopeless), especially for all those who are waiting for a *real* upgrade and a better body from Pentax, maybe for many years already, including me of course. With the low Image Quality of the K-7, no matter on resolution and noise level (and the AWB is not really quite as good as others as well, according to the above AWB test in the lab environment), the K-7 is really a big let-down.
At the end of the days, I just wish to ask: Without good IQ and PQ, what else is/are more important for a DSLR?? I guess one of the main culprits for the low IQ/PQ performance of the K-7 should be the new version of the Samsung sensor, which has the speed increased but not so for the image quality and noise control, but the reverse! (So, for those who are really having high hopes on the upcoming NX, my advice is don't let yourself have *high hopes*!)
Furthermore, in view of the machine/system performance, no Pentax could actually match and catch up with C and N machine guns and AF monsters anyway, as proven and verified again in this test.
I originally planned to buy the K-7 the first time I heard about it and learnt something more about its specs. But the reality and many reviews just show that it just cannot live up to the hypes Pentax/Hoya (more so by the Internet community actually) have created and its actual performance and (the low) IQ are just falling short of in many ways.
Objectively, it is really rather unwise for me to invest in this body as I am afraid that its PQ (both resolution and noise) are just even worse than the old K20D which is now still available brand new at half of the current price of the K-7 brand new (although the K20D has so many known issues and problems, too). In the meantime, I think the only model of DSLR body in the Pentax land which I could recommend is the K-m (Yes it is an entry level model, but it is the ONLY *Sensible* Choice right now, IMVHO), for which I have acted recently - I've acquired a special edition, anyway.
Besides, snowing is now all over the Pentax land for all Pentaxians worldwide in view of the ridiculous high prices of all those Pentax lenses now on sales as well as the crazy increased prices of those Pentax accessories as well. I've checked today the price of a 360FGZ flashgun has now been increased for +75% (which is just about 20% more expensive than a Canon 430EX II flashgun!) and that of the 540FGZ is now more expensive than the Canon top-of-the-line Pro usage weather-sealed 580 EXII too. I think Pentax the brand will really soon go out of business in one to two years to come (if they still do not make better products and re-establish a more sensible pricing strategy and revise all the nonsense (increased) new prices of their lenses and accessories).
Last but not least, if you are in Hong Kong, I highly recommend that you should buy the magazine (if you're interested) to see all the test pictures and results as well as more comments by the testers/editors, you can buy the magazine before August 7 (next Friday).