Ken Nielsen
25th of June 2009 (Thu), 17:31
I hear cameras that employ the Foveon method of image gathering, like the Sigma SD14, produce sharper images than the Bayer method used by most digital cameras on the market today?
Is this true? For bird photography should we be exploring cameras that employ this technology?
A typical CCD or CMOS sensor is monochrome, and the appearance of color is provided by the bayer matrix. You only know the luminosity of one color for each pixel, and the camera interpolates the other colors from neighboring pixels. If you ever look at individual channels of a digital camera image, you'll find the green channel is the most detailed, because half the data gathered is in the green wavelengths, and only a quarter in red and blue.
Kodak came up with a twist on bayer filters where a certain number of pixels have no filter at all, and with some fancy interpolation can result in shorter exposures under the same conditions.
Yet another scheme is rather than RGB, they use a bayer filter with cyan, magenta, and yellow filters, and subtract neighboring pixels from each other to achieve a typical RGB image.
Some special purpose cameras are monochrome. Security cameras often are because they can see better in the dark, blocking less light. Also, astronomy cameras often lack a filter -- the astronomy geek often creates four separate images -- one luminance and 3 color images. You can get more detail with less exposure time that way, which is important when single exposures measured in minutes or hours and your subject doesn't change. It also allows one to do neat things like use narrow-band filters. Many nebulae emit mostly in the hydrogen-alpha band, so using a HA filter knocks out most of the light pollution humans create.
A Foveon sensor gathers red, green, and blue data in every single pixel, kind of stacking 3 sensors in each pixel instead of 1. So pictures with the same number of megapixels are sharper from a foveon sensor, with about double the resolving power. However, they've still got issues with noise, especially with longer exposures.
Thanks,
Ken
Is this true? For bird photography should we be exploring cameras that employ this technology?
A typical CCD or CMOS sensor is monochrome, and the appearance of color is provided by the bayer matrix. You only know the luminosity of one color for each pixel, and the camera interpolates the other colors from neighboring pixels. If you ever look at individual channels of a digital camera image, you'll find the green channel is the most detailed, because half the data gathered is in the green wavelengths, and only a quarter in red and blue.
Kodak came up with a twist on bayer filters where a certain number of pixels have no filter at all, and with some fancy interpolation can result in shorter exposures under the same conditions.
Yet another scheme is rather than RGB, they use a bayer filter with cyan, magenta, and yellow filters, and subtract neighboring pixels from each other to achieve a typical RGB image.
Some special purpose cameras are monochrome. Security cameras often are because they can see better in the dark, blocking less light. Also, astronomy cameras often lack a filter -- the astronomy geek often creates four separate images -- one luminance and 3 color images. You can get more detail with less exposure time that way, which is important when single exposures measured in minutes or hours and your subject doesn't change. It also allows one to do neat things like use narrow-band filters. Many nebulae emit mostly in the hydrogen-alpha band, so using a HA filter knocks out most of the light pollution humans create.
A Foveon sensor gathers red, green, and blue data in every single pixel, kind of stacking 3 sensors in each pixel instead of 1. So pictures with the same number of megapixels are sharper from a foveon sensor, with about double the resolving power. However, they've still got issues with noise, especially with longer exposures.
Thanks,
Ken