>> I think push processing is a very good analogy. You can't strictly >> speaking change the ISO on a digital camera - the sensor has a fixed >> sensitivity. ... >> > > ... at a specified voltage input/output level. If you change the > voltage levels, you are changing the response curve and thus the ISO. > Changing ISO is not the same as multiplying the pixel values from a > sensor with a fixed voltage level, which is basically what the > exposure compensation slider in a RAW converter does unless the > programmer who designed it includes additional processing of > differential values. > But doesn't the ISO setting in the camera amplify the voltages after the data is read from the sensor? I really can't see how this is different from multiplying the pixel values. Or it's different, as in one case you multiply the values before they are converted to digital, in the other after - but in principle there's not much difference, and the final image quality ought to be the same. Unless the A/D that chops off real data in the low end when the gain is 1 (i.e. when the max value of the A/D matches saturation of the sensor), but I gather that it doesn't. The theoretical extra bits that get thrown away, are assumed to contain only random variation caused by noise.
Also, ISO for a digital sensor is apparently defined in terms of signal-to-noise ratio, and that obviously doesn't change when you multiply the signal. - Toralf -- PDML Pentax-Discuss Mail List [email protected] http://pdml.net/mailman/listinfo/pdml_pdml.net
