Archived posting to the Leica Users Group, 2014/03/22
[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]Further to my earlier post, I took another photo with the NEX-7, but instead of setting the FL of the kit 18-55 zoom to 24mm so as to get the same FOV as the D800 with its 35mm prime, I set it at an actual 35mm. This produced the same image scale on the sensor as the 35mm prime does on the D800 sensor. What this does is to show the image resolution that would result from a 35mm lens working with a 54MP FF sensor. (The FF sensor has 1.5 times the linear dimensions of the NEX?s, so that a FF sensor with the same pixel size would have 1.5 x 1.5 times as many pixels.) The result is quite interesting. The uprights on a deck railing about 300 meters away are substantially clearer and better resolved in the NEX image than in the D800 image, and much better than in the NEX image taken with the zoom set to give the 35mm-equivalent FOV. This shows not only that the increase from 24MP to 36MP yields more information, but that the inherent resolution of even Sony's inexpensive lens may not be exhausted by going to 54MP. I think this is illuminating, and it certainly supports my earlier impression that the pixel count that will be needed to exploit the inherent resolving power of the best modern glass will be in the range of 50-100MP. For comparison, the best general-use 35mm film seems to have a resolution of about 150 lp/mm, or 300 pixels/mm, for FF content of somewhere around 80MP. Of course, resolution is not the only important attribute of a lens or an image, but for critical work involving large formats, high crop factors, or close inspection, it is AN important attribute. It will be interesting to see where camera manufacturers set the ?sweet spot? for pixel size/MP count as the evolution of sensor technology makes FF chips in the 50-100MP range practical. I can foresee a new variable at the disposal of the digital photographer of the future, the ability to set the ?binning factor? of the camera. Because tiny photosites on a high-MP chip generate more noise than larger ones, pooling the signals from 2x2 or 3x3 arrays of photosites results in a better S/N ratio and is often done in deep-sky astroimaging (which represents the ultimate in available-light photography!). A user of the Leica M typ 960 could use the native resolution for APO-Summicron scenics that will be printed large, but bin 3x3 for noise-free 11MP nocturnal street-photography images at ISO 6400! I know that different resolutions can be selected with many cameras now, but I don?t know whether this file-size reduction is accomplished by binning or just by down-sampling, or whether it does anything to decrease noise. I?ll experiment with the different resolutions available for the JPEG files on the typ 240. ?howard