The metadata for this image claims it was shot at f/16. That's wrong; it was more like f/2.5 or so. This was an artifact of the too-clever-by-half way Leica M cameras estimate the f stop. There's no mechanical link between the aperture ring and the camera body, so instead they estimate the f-stop with a separate light sensor that's compared with the brightness of the recorded image. This works reasonably well, except when you use an ND filter (as here), which confuses it to no end.
mattblaze
drahardja