Photophys.com: The Science of Photography

At page 64 of the book you wrote: "Each pixel typically contains one sensor element so nothing smaller than one pixel can be resolved." In my opinion this quite inaccurate and here is what I consider as a proof of that: when one takes a photograph of a dark power line against a bright sky from a distance large enough to result in an image of the power line much thiner than a pixel, none of the pixels will be as dark as the power line is but the power line will be very visible. A harsh application of your statment would imply it should not be visible at all... This example is the first one I actually encountered in real life (knowing the height of a distant building I photographed, an approximation of the diameter of a power line wich was next to it and the number of pixels in the heigth of the sensor, I was very disturbed to see the power lines in the resulting image!) An other common example is that the stars don't disapear brutally when their apparent diameter get lower than a single pixel. This depends of the contrast but on thin textures like leaves on a distant tree my guess is that this play a significant but certainly less obvious role. Besides that, because I make many stitched panoramas and because common "half-true" rules about perspective, DoF, diffraction, etc, often don't apply at all in this case, I find your book extremely useful. Example: http://www.autopano.net/forum/t6722-une-plante-grasse-de-ma-terrasse-a-fleuri http://www.autopano.net/forum/t6731-want-larger-dof-use-f22-or-f32-aperture-shoot-more-images

Posted by Georges Lagarde on August 09, 2009 at 11:54 AM EDT #