Sony FE 12-24mm f/2.8 GM Review
inksandpaper wrote:
That's strange. Shimming tilt at the mount does not cause FC to increase significantly. I wonder what's going on. What do you mean by the focus distance reported was less accurate? How much less accurate was it? Is it commensurate with the focus shift you would expect on-axis for a 10-micron correction induced at the screw, which is ~5 microns only on-axis?
Yes, I was susprised to see the induced FC given how thin the shim was. Past about 1m, the in-camera distance readout was increased significantly. E.g. objects at ~2m were sharpest for a readout of about 4m, whereas unshimmed it was actually surprisingly accurate. I rely on such distance readouts to do my focus stacking, so this strange readout remapping might make that quite challenging. I'm not sure how to calculate how much focus shift I should expect.
But the whole affair has remined me why I was so glad to see more high IQ native lenses released for the Sony E system after messing around with shimmed adapters in the early days.
** edit *** Maybe I'm underestimating the shim thickness due to the lens mount screws exerting less compressive force than I think. But I don't think it could be more than 20um. Whatever the effective thickness was, it was about the right amount to balance corner sharpness, so not a viable solution for me due to the FC and distance readout problems.
That's strange. Shimming tilt at the mount does not cause FC to increase significantly. I wonder what's going on. What do you mean by the focus distance reported was less accurate? How much less accurate was it? Is it commensurate with the focus shift you would expect on-axis for a 10-micron correction induced at the screw, which is ~5 microns only on-axis?
Yes, I was susprised to see the induced FC given how thin the shim was. Past about 1m, the in-camera distance readout was increased significantly. E.g. objects at ~2m were sharpest for a readout of about 4m, whereas unshimmed it was actually surprisingly accurate. I rely on such distance readouts to do my focus stacking, so this strange readout remapping might make that quite challenging. I'm not sure how to calculate how much focus shift I should expect.
But the whole affair has remined me why I was so glad to see more high IQ native lenses released for the Sony E system after messing around with shimmed adapters in the early days.
** edit *** Maybe I'm underestimating the shim thickness due to the lens mount screws exerting less compressive force than I think. But I don't think it could be more than 20um. Whatever the effective thickness was, it was about the right amount to balance corner sharpness, so not a viable solution for me due to the FC and distance readout problems.