"But at low ISO nothing can beat this camera." - CEO SIGMA

Started 8 months ago | Discussions thread
Ray Ritchie
Veteran MemberPosts: 3,779
Like?
Re: At low or high ISO, there are quite a few.
In reply to D Cox, 8 months ago

I think you may also be missing a key point - or at least choosing to leave it out of the discussion. The Bayer sensor in your example does, in fact, have more resolution in the luminance channel, because it has more sensel locations, or physical 2-dimensional sampling points. It is true that each of these is filtered detect primarily one color, but the different colors can all contribute  a similar resolution in luminance. The Bayer sensor has more locations devoted to green than blue or red because that's the way human vision works - it's roughly twice as sensitive to green.

The 16 MP Foveon sensor in your example has fewer sensel locations, but more information about color at each location - potentially allowing it to have more accurate color. But more accurate color at fewer physical locations does not equate directly to higher resolution in the X-Y plane. The Sigma marketing mantra that the sensor is "48 megapixels" is misleading, since although there are three "pixels" at each physical sampling location, but these "pixels" are also not of equal quality, because they are located at different depths in the chip and have different levels of noise and color shift.

What it amounts to is that a 24 "MP" Bayer sensor has higher luminance resolution than a 16 "MP" Foveon, and lower color resolution. I think of the Bayer sensor design as being somewhat like conventional color television design, where a higher resolution black and white image is essentially "colorized" by a much lower resolution wash of color.

I don't think you can exactly compare the resolutions of a 24 MP Bayer and and a 16 MP Foveon by any known analytical technique, because not only are their "pixels" are not the same, but neither samples all three color channels with equal resolution at each physical location. You end up making comparisons of "perceived resolution," or something of the sort, and it's made more "apples and oranges" by the fact that most comparisons are of images processed by totally different software packages, so that it becomes difficult to separate the inherent capabilities of the sensor from the ingenuity of the algorithms used to render raw sensor data into final images. And of course, it's pretty rare to find a Foveon image compared to the exact same image shot on a Bayer camera, as well.

It would be interesting to see some 16 MP Foveon images in black and white compared to the same images shot with a 16 MP Bayer sensor with no AA filter. I suspect the two would look much more similar than most of the color comparisons we see in these debates.

-- hide signature --
Reply   Reply with quote   Complain
Post (hide subjects)Posted by
110New
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark post MMy threads
Color scheme? Blue / Yellow