autofocus points picture

Autofocus points/sensors are distributed over the picture frame. The fewer the number of autofocus points the more they will be placed clustered around the centre of the picture frame. This makes focusing on subjects that are off-centre difficult and the photographer may need to centre the subject first for focusing, before re-framing for the shot. The more autofocus points there are, the more the area of the frame is covered, thus the more reliable and accurate the autofocus is able to be.

Autofocus sensors work by measuring the difference in contrast across the sensor. A sharp point of focus will have a strong edge contrast, with the edge weakening as it becomes more blurry. This is the principle on which the autofocus processor determines that maximum sharpness has been reached.

The Mechanics Of Autofocus

Autofocus essentially works by a system of trial and measure. First the lens turns a little and a measure is made to determine whether sharpness has improved. Based on this, the sensor will determine by how much and in what direction the lens should turn to improve focus. Further adjustments would continue until maximum sharpness is reached. Usually this all happens in less than a second, but in conditions where edge sharpness is difficult to find, like in evenly lit, evenly textured scenes, it can take longer. While the process remains the same for all cameras, the speed varies. This is influenced by the accuracy of the autofocus processor but largely determined by the speed of the motor moving the lens. For this reason a motor that uses ultrasonic waves to drive the lens is generally recommended for faster focusing.

How Your Camera Selects Which Point To Use To Focus With?

Your camera doesn’t know what the subject of your photograph is and unless it is equipped with face detection, it can’t even recognize the presence of humans. The decision on which point in the frame to use is based on the brightness, the level of contrast and the amount of movement detected across the sensor. Cross-type sensors, that have both a lateral and vertical line of detection are more accurate than the single vertical line sensors.