Saturday, November 15, 2008

Does the Pentax K20D Perform Better with non-DA Lenses?

by Miserere

When the K20D was first announced, a collective groan went up when Pentax unveiled the pixel count of its sensor. 14.6MP sounded like too many to provide low-noise images with a wide dynamic range. Pentax were quick to calm us down by pointing out that the individual pixels in the new CMOS sensor were actually of a similar size to those of the K10D; what they had done was shrink the inter-pixel space where all the circuitry is located so they could now fit more pixels onto the chip without making them significantly smaller.


They also improved the micro lenses that sit on top of each pixel. By bringing them closer to the surface of the pixel they improved the light-gathering capacity of each pixel because light reaching the micro lenses from different angles will now be better directed towards the pixels instead of being lost in the spaces between them.

While DA lenses have been designed so the light leaving the lens from the rear-most element is as perpendicular to the sensor plane as possible, older lenses were not designed with this requirement because film is sensitive to light coming in from almost any direction. This is one of the reasons why some older lenses don't perform as well on digital SLRS as they did on film SLRs.

All these improvements should help the K20D provide better IQ with non-DA lenses than previous DSLRs from Pentax. If this is true, it would be a very important reason to upgrade for those of us who own a lot of old lenses.

As you might imagine, I don't have a K20D, so I cannot test this myself. It would be interesting if blog readers owning a K20D and one of the other Pentax DSLRs could perform some experiments to show whether Pentax's claims are true or not. Yvon, I know you have both a K10D and a K20D; this sounds like a nice weekend project for you :-) Of course, you'd post the results here for all to see.

I wouldn't be surprised if this were true, and it's another big plus for the K20D that Pentax isn't advertising emphatically enough.

Run some tests, and post your results in the comments section.




Editor: Yvon Bourque said...


Good article. That's an idea. I can compare the 10MP CCD equipped Pentax DSLR with the 14.6MP CMOS sensor with older lenses. However, I don't have a setup such as Phil Askey for instance. What do you suggest for test? Bright colors with fine details? Same photo at same location, unretouched at all in RAW then converted directly to JPG or just JPG? Will the naked eye be able to see the difference?
I will be glad to conduct these non-technical tests along withthe other DA/DA* lenses test for use with full frame.Of course Pentax is not getting on the bandwagon for full frame, but one can revert back to your Love letter to Pentax. They may still do it.

robin said...

When the K20D came out there were a lot of posts about how suddenly existing lenses looked better... in ways that the pixel count alone could not account for. Part of this was focus accuracy and perhaps more was due to the rendering the new pixel arrangement made possible. Anecdotal evidence, sure, but if it was a random subjective effect I'd expect some posters to remark on the opposite. No-one did.

Sushil Kumar said...

Thanks for the information. I don't understand why Pentax doesn't advertise such improvements in the technology while other brands do?

Miserere said...

Yvon, I think the test should be pretty straight-forward. Just plonk the cameras on a tripod and choose a scene with plenty of detail in the corners of the frame. A shot in a park with lots of trees is probably the best.

Use a pre-DA lens, preferably a wide angle such as a 24mm. I'm sure wide angles is where most of the difference will be due to their wide image cone.

If you didn't live on the other side of the country I would let you borrow my FA 20mm :-)