And now for something completely different… or is it?

I am sure that you all have heard of Ken Rockwell's blog. Ken has a very popular blog where he tells us his opinion on everything photographic. His opinion changes (lately from being apologetic for RealRaw -whatever that means- and then for the Leica M9), but that makes him more entertaining.

His post from Friday, October 23rd 2009 took my attention. You'll find the post on this page:
http://www.kenrockwell.com/tech/2009-10-new.htm

In this post, Ken posts a capture from the M9 which looks stunning. Here is a crop for your reference:

(this picture and all the next ones in this post are cropped at the pixel level, just click them to display the original file)

There is no denying that this is a pretty sharp picture. How did he do it?

There is no denying that the Leica 21mm f/2.8 ASPH is a fine lens (but why don't you go back a month in this blog to see how sharp the venerable Minolta 20mm f/2.8 is, mmh?), but the fine print says something interesting about the M9's DNG being sharper than the M9's jpegs.

Wait a minute, I have read something similar somewhere. Here it is:
http://artaphot.ch/dslrs/163-a900-detail-resolution-raw-vs-jpeg

(artaphot is another fine blog, but mainly about Sony/Minolta. Maybe a bit less entertaining, but Stephan Kölliker does not change his mind as often than Ken Rockwell does)

This is the difference that Stephan finds between jpeg and raw capture on the A900 (small sample from the article, please check the link for the whole story).

Could it be that indeed the raw capture from the A900 is as good as the raw capture from the M9? Let's match the image from Ken and the image from Stephan in contrast and saturation:

Quite close isn't it? And keep in mind that I did not work on the original raw for the A900, but used Stephan's jpeg instead.


So what does this tell us? Is the M9 that much better than the competition? Apparently not as much as Ken appears to think. Is the problem in jpeg, then? Probably not, all what you are seeing from this site is saved in jpeg at some point. So what is it?

The problem is the jped engine in the cameras. Both the M9 and the A900 lose a great deal of details when converting their pictures to jpeg internally.

So, are we discussing something completely different or are we still discussing the comparison between a full frame A900 and a Canon Ixus P&S? On the previous post, they look quite similar, remember? But everything was taken from jpegs, whereas all the other lens tests on this blog come from raw captures. What happens if we compare the Ixus (necessarily jpeg) to the A900 raw?

This is the A900 full frame capture, converted internally to a 12 Mpix jpeg:

This is the Canon Ixus capture (jpeg, obviously):

This is one of the pictures from the "35mm" test page, resized to 12 Mpix in photoshop:

It looks as if I should redo the comparison between full frame, aps-C, four thirds and small sensor P&S from raw files, doesn't it?