10-06-2010, 11:43 AM
[quote name='Lomskij' timestamp='1286353937' post='3522']
Vieux loup:
You're soo politically incorrect - instead of "pixel peepers" you should have used something less offensive, "diminutive detail examiners" for instance :-)))
Brightcolours:
I was browsing the maxmax website and stumbled upon this:
http://www.maxmax.com/b&w_conversion.htm
It's not to prove my point or anything, but I found this rather interesting.
[/quote]
Yes, I agree that BW sensors are interesting. They do have a big disadvantage though.. with a colour sensor you can determine what the BW will look like, you can choose how the red, the blue, the green.. or even the yellow, the magenta and the cyan convert to BW. You lose that advantage with a BW sensor.
To come back to the by you noted blocky detail in colour areas (particularly with the red hens in the car dials) the D200 examples which I blamed on the not so great RAW converter used.
I showed you a link earlier to a RAW file from a Canon EOS 5D which had its AA-filter removed. It is not the most ideal image to pixel peep and dissect (taken with the old 24mm f1.4 at f5.6 (3.5 in APS-C terms, no not great DOF/sharpness in the scene), a lens that shows clear CA (green/magenta in this lens' case), but for now we will have to do with it, for lack of other RAW files from AA-filter-less cameras.
I will provide some crops, with comments on what the crops show. I converted it with a good raw converter, RPP 4.1.7.
Note that you do get to see a dpreview processed version of it, you can not link to "originals". But it is in original resolution. Crops at 400% so we can clearly see individual pixels, with a 100% crop inset.
![[Image: 8707693878B34C5DB00E9864CEC1C9BB.jpg]](http://g2.img-dpreview.com/8707693878B34C5DB00E9864CEC1C9BB.jpg)
Crop 1, 2, 3 and 4 show relatively clearly that the colour detail is NOT blocky as we saw with the (according to me badly converted) D200 crops. Sadly, due to the not very sharp image (due to the chosen aperture and scene) we do not have the sharpest possible edges. But still, I think it illustrates well the detail per pixel (not like the "grouped" D200 sample pixels).
Crop 5 shows a luminance moire pattern (with CA from the lens) which would also show up with a Sigma/Foveon sensor, due to the lack of AA-filter.
Crop 6, and much more strongly (more in focus too) crop 7, show moire colour artifacts, which you would NOT see with a Sigma/Foveon sensor. Crop 7 also shows luminance moire, which would also show up with a Sigma/Foveon sensor, due to lack of AA-filter.
Vieux loup:
You're soo politically incorrect - instead of "pixel peepers" you should have used something less offensive, "diminutive detail examiners" for instance :-)))
Brightcolours:
I was browsing the maxmax website and stumbled upon this:
http://www.maxmax.com/b&w_conversion.htm
It's not to prove my point or anything, but I found this rather interesting.
[/quote]
Yes, I agree that BW sensors are interesting. They do have a big disadvantage though.. with a colour sensor you can determine what the BW will look like, you can choose how the red, the blue, the green.. or even the yellow, the magenta and the cyan convert to BW. You lose that advantage with a BW sensor.
To come back to the by you noted blocky detail in colour areas (particularly with the red hens in the car dials) the D200 examples which I blamed on the not so great RAW converter used.
I showed you a link earlier to a RAW file from a Canon EOS 5D which had its AA-filter removed. It is not the most ideal image to pixel peep and dissect (taken with the old 24mm f1.4 at f5.6 (3.5 in APS-C terms, no not great DOF/sharpness in the scene), a lens that shows clear CA (green/magenta in this lens' case), but for now we will have to do with it, for lack of other RAW files from AA-filter-less cameras.
I will provide some crops, with comments on what the crops show. I converted it with a good raw converter, RPP 4.1.7.
Note that you do get to see a dpreview processed version of it, you can not link to "originals". But it is in original resolution. Crops at 400% so we can clearly see individual pixels, with a 100% crop inset.
Crop 1, 2, 3 and 4 show relatively clearly that the colour detail is NOT blocky as we saw with the (according to me badly converted) D200 crops. Sadly, due to the not very sharp image (due to the chosen aperture and scene) we do not have the sharpest possible edges. But still, I think it illustrates well the detail per pixel (not like the "grouped" D200 sample pixels).
Crop 5 shows a luminance moire pattern (with CA from the lens) which would also show up with a Sigma/Foveon sensor, due to the lack of AA-filter.
Crop 6, and much more strongly (more in focus too) crop 7, show moire colour artifacts, which you would NOT see with a Sigma/Foveon sensor. Crop 7 also shows luminance moire, which would also show up with a Sigma/Foveon sensor, due to lack of AA-filter.