4 ms·
Thanks for the info, and to the contributors below for their comments too. I also wonder -- in what vertical / horizontal configuration did they do the test?
by supernova87a 6y ago
Thanks for the info, and to the contributors below for their comments too.
I also wonder -- in what vertical / horizontal configuration did they do the test? Lens pointing up, compressing the body? Or horizontal as if taking a normal photo? From their photo in part 1 of the story, it shows lens aiming at the ceiling.
I wonder if the weight/flexure of the body causes an appreciable change in especially plastic camera components, and therefore the lens-plane distance? Anyway, regardless I suppose the variation in camera bodies is the point, and holding in a different orientation would nonetheless produce this finding. (unless flexure only happens in a certain orientation)
- formerly_proven 6y agoThe Denz tester is the lens: https://denz-deniz.com/wp-content/uploads/2016/09/Backfocus_Off_V4.png https://denz-deniz.com/wp-content/uploads/2016/09/Backfocus_... I wondered how they measure the distance so accurately, since normal optical measurements would be quite hard in this setting (optical sensor filter + sensor cover glass + microlenses + sensor surface), so instead the measurement is done using the sensor itself, so the effective optical flange distance is measured. > Evaluating the flange focal depth is based on projecting a red bar and a pair of green bars from different optical axis onto the chip. In the right distance from the projective system the red bar comes to lie right between the green bars forming a symmetrical image that can be easily recognized looking at a monitor that is connected to the video output of the camera. > In case the flange focal depth of the camera suspect to measurement is perfect (namely 52.00mm in case of a PL mount camera) the FDC multi’s scale ring attached to the hand wheel will show that “0” correction is necessary and the combined image actually is symmetrical. > In case that the flange focal depth is not right, the hand wheel of the FDC multi has to be turned, changing the position of the projective system along the optical axis until the combined image is symmetrical again. Now one can read the amount of correction, that has to be applied to the cameras mount, from the scale ring. The adjustment of the flange focal depth itself involves putting or removing a combination of shims under the camera mount. Thanks to the exact instruction painful trial and error procedures are history.