
It has been almost ten years since Pentax launched their first DSLR, the *ist D, almost ten years ago. I did buy my *ist D very shortly after it hit the streets and it may be interesting to read back my old textural review of it which was first published back to fall 2003.
Okay, nine years something has elapsed and now the new Pentax cameras have climbed up to 16 megapixels in pixel count and resolution. Yet, as we all know, they are all equipped with an APS-C sized cropped (still!) sensor in 16M. The noise level of the new CMOS sensor is undoubtedly lower and they are video capable. But how about the image quality besides resolution? What have we gained, or, actually lost?
So, I decided to do a shootout test myself, side-by-side and head-to-head for my two old and newer Pentax DSLRs, namely, the *ist DL2, which is the last *ist body actually, and the K-5, which is the first Pentax DSLR using the same never-upgraded but ever recycled Sony 16M CMOS sensor or its variant in all the newer Pentax DSLRs up till now! :-o
Test Conditions:
Both cameras set in RAW mode, manual white balance in daylight, fixed at ISO 200, the same lens of DA21/3.2 was used in Av mode, the same scene was then shot consecutively in pairs;
Custom image mode was set to "Natural" with all centre/zero/default settings except sharpness was set at +1. Pictures were then all converted from RAW to 6M Jpegs without any adjustment (using camera settings) with original Pentax Digital Camera Utility (Laboratory) Version 4.3, for both the *ist DL2 and K-5 pictures. The results are posted in the following album of mine:
http://smg.photobucket.com/user/RiceHigh/library/K-5/Colour_Test_Vs_DL2?sort=9&page=1
Full size original images are available for downloading per popped-up picture page in the above.
What I can tell is that the *ist DL2 images are far more natural and faithful to the actual scene in real world than all those made with the K-5! The K-5 images are yet whiter than the DL2's, though, despite that both cameras were actually set to daylight preset manual white balance already.
If you look at the leaves of the trees, the green colours of the K-5 pics are actually looking rather fake and odd. Ditto for the brown colours of the trunks of the trees. In fact, the flower I shot is in orange red, but not that red colour that K-5 has produced and rendered!
Besides, to my surprise, the dynamic range of the *ist DL2 pics is shown to be yet better than that of the K-5 despite that the DR mark measured by DxO is far higher for the K-5 and its 16M CMOS sensor! Just observe the shadow details.. :-o
In fact, the favourable colour rendition was top one big reason why I still stayed with Pentax system regardless of the low performance of the Pentax bodies plus all the various technical and quality issues that were associated with the Pentax system and Pentax themselves. Now, it seems that this main point and reason has also gone with the later and latest Pentax bodies! >:-[
At the end of the day, I am afraid that now putting those excellent Pentax glass on a Canon body do produce much more favourable results even in term of colour rendition, see my this set of recent examples here.
After all, you may have a different opinion than me, but I think my pictures have clearly spoken for themselves! And, this test should be scientific enough as I have already tried to keep all the possible variables and affecting factors to minimal, FWIW.
Simon · 611 weeks ago
Tried RAW and DPP (in my case it is a 5DII), but I couldn't reach the rendition of Nikon(D600)/Sony(a350)/Pentax(K200D).
I don't know exactly, but the Canon colors have a yellowish tint, pictures from Nikon D600 (out of the cam) look way better.
I think Sony sensors are superior to the Canon ones in terms of pleasant color rendition and even sharpness, but I don't know the 5DIII,6D,1Dx.
Kindly Regards.
Anonimity is Great · 611 weeks ago
Z.T. · 611 weeks ago
I have compared images coming out of Q to similar images taken with my old Minolta A1 digital camera pictures (A1 and A2 used a CCD). Well, even 7 years after, and with all the great tech advancement, the old CCD makes much more pleasing looking colours at usable ISOs and by default. And Yes, the feeling of the brightness of photographs goes towards natural warm whites, not bleached out, artificial whites (which give an illusion of a wider gamut).
We have gained something through the process of switching everything to CMOS tech, but we have lost something too. I think too much has been sacrificed to gain purely on noise performance, because that's what sensor tech was all about in the last 5 years — a mad race towards less noise, less noise. But we've lost the magic of colour and tone — the very essence of photography.
zosX · 611 weeks ago
Z.T · 611 weeks ago
foton · 611 weeks ago
Yoku · 611 weeks ago
Peter · 611 weeks ago
Victor · 611 weeks ago
davescliches 52p · 611 weeks ago
Lurker · 611 weeks ago
Kind of like a modern optical mouse and the old ones mechanical (balls) ones. They still send signals to the computer using wires (I know, now there are wireless mice), but how they collect input is different.
But still, you can't deny that the differences between CMOS and CCD images are perceptible. Could it be the change in image processing that has changed?
Dominique · 611 weeks ago
Pancakemaster 9000 · 569 weeks ago
Mind you, I HAVE only used one CCD sensor camera (and three CMOS sensor cameras), namely the M8, so I can't really say for sure that that's the case for all CCD sensor cameras... but I have read up on this issue a fair bit, and other peoples' anecdotal evidence does seem to support my statement at least somewhat.
The cameras I had before were the 400D (fairly good colours, though the blues were difficult to work with), the 5D (overall kind of excellent to work with in terms of colour, but nowhere near the M8) and the Fuji X100 (the colours, other than skin tones, which were brilliant, were near impossible to get to a place where they are even somewhat pleasant to me).
Also, the colour differences may actually have more to do with the AD converters and how they work within the system than the actual colour filters and processing algorithms (would make the observation that CCD files respond better to processing, especially colour processing make more sense), which is to say, that overall the tonal response would be different. Apparently that's one of the main practical(which is to say that there are many, many other differences, but in terms of tonal output most of the variables along the pipelines either have no practical effect because software could easily compensate, but this one is pretty fundamental to how the image is recorded, and that's currently the thing being discussed) between CCD and CMOS sensors.
robirobi 1p · 566 weeks ago
ccd yan · 566 weeks ago
paco · 516 weeks ago