12-25-2015, 01:45 AM
You don't need to look specifically for "lines". Just any kind of contrast will do, even if they are tiny specs of pollen on a for the rest featureless background.
It can happen that one misfousses when there is a repeating pattern, where PD AF may think one contrast edge is another, and then says things are in focus. But that happens very seldom, and mostly only with line type AF points.
It can also happen that there are edges of contrast, but only in the vertical orientation where the AF point used has a vertical line sensor point. Then it can happen the AF does not see any edge. This too is very seldom.
PD AF misfocus happens in two circumstances: Apparently, it is said that Nikon cameras with Sigma lenses (probably also with other lenses) can somehow misfocus in artificial light, because the light "bends" differently and the AF conclusions are not valid as they are with white light and the AF with the lens does not get corrected for that. Not sure if that is a correct explanation, or if something else occurs, but I have read a number of times of someone complaining that with artificial light OOF results are produced.
And then there is the problem when the lens is not calibrated correctly "for the camera body", and steps and especially final steps in the AF process are different than what one expects, resulting in an OOF result. Often the results then are different when starting the Af from infinity than when starting from MFD.
Also, aberrations can influence what the AF sensor sees, being different from what the image sensor gets.
With CD AF, you can get OOF results from lenses/cameras overstepping exact focus and not realizing that, since they can't actually see when something is in focus, and only have a difference in contrast to go on. And at darker circumstances the live view image stream can get too noisy to do contrast comparisons from frame to frame.
It can happen that one misfousses when there is a repeating pattern, where PD AF may think one contrast edge is another, and then says things are in focus. But that happens very seldom, and mostly only with line type AF points.
It can also happen that there are edges of contrast, but only in the vertical orientation where the AF point used has a vertical line sensor point. Then it can happen the AF does not see any edge. This too is very seldom.
PD AF misfocus happens in two circumstances: Apparently, it is said that Nikon cameras with Sigma lenses (probably also with other lenses) can somehow misfocus in artificial light, because the light "bends" differently and the AF conclusions are not valid as they are with white light and the AF with the lens does not get corrected for that. Not sure if that is a correct explanation, or if something else occurs, but I have read a number of times of someone complaining that with artificial light OOF results are produced.
And then there is the problem when the lens is not calibrated correctly "for the camera body", and steps and especially final steps in the AF process are different than what one expects, resulting in an OOF result. Often the results then are different when starting the Af from infinity than when starting from MFD.
Also, aberrations can influence what the AF sensor sees, being different from what the image sensor gets.
With CD AF, you can get OOF results from lenses/cameras overstepping exact focus and not realizing that, since they can't actually see when something is in focus, and only have a difference in contrast to go on. And at darker circumstances the live view image stream can get too noisy to do contrast comparisons from frame to frame.