Monday, August 22, 2005, 12:51:26 AM, Graywolf wrote: G> That one is easy, Shel, each bit doubles (or halves if you are G> going that way) the dynamic range of the image. Powers of two, you G> will recall.
Hi Graywold, I think there is abit of a mistake. AFAIK dynamic range, if understood as ability of the film/digital to capture range of light, is not basically dependant on bits per pixel. DR is dependant mainly on the type of sensor, and also the electronics behind it (AD converters,etc.). Some sensors can capture a range from pure black to say 5 stops of light, some only 4, some 6, et cetera, before the pixels "fills" with electrons and doesn't register any more light falling on it. More bits per pixel only allow it to resolve the levels more finely, thus a 14-bit sensor can resolve more details in darkest stops, but it doesn't increase the range from pure black to supernova bright. Noise gets into this to complicate matters a "bit", though - all the dynamic range in the world doesn't help you if all you get from the darker stops is noise. And sure, sensors with more dynamic range need more bits per pixel to actually resolve the dynamic range in fine enough to be practical steps. Thus to get better dynamic range better sensors are needed - like these in medium format backs or the Fuji hybrid sensor, but also more bits are needed with these sensors to resolve the additional levels of light. The bits are just like markings on a ruler, while the length of the ruler is the dynamic range. Of course a 10m ruler with 2m markings won't be fine enough, whil 5m ruler with 10cm markings won't help you with the highlights that are "brighter" that fall from 6m to 10m... At least that's my understanding of the issue. Frantisek
