Here is my take on the discussion of the merits of using medium format at 2500 dpi vs. using 35 mm film at 4000 dpi. Figuring that a picture is worth 1000 words, I have created some images to illustrate the issue. This table illustrates what the image simulations were meant to represent.
10526 dpi | 14 mm negative | |
6421 dpi | 22 mm negative | |
4000 dpi | 36 mm negative | |
2421 dpi | 58 mm negative | |
1473 dpi | 94 mm negative | |
947 dpi | 152 mm negative | |
632 dpi | 247 mm negative |
In the above table, the top image represents a scan of an image where the grain size of 10526 grain fields per milimeter matches the scan resolution. The bottom image is scanned at a much lower resolution, so grain is less evident
To create these images, I started with an image 1600 by 1200 in size and painted it with random noise to simulate an extreme case of film grain. The top image, being a scan using a resolution matching the grain size, is very noisy. Each subsequent image was created by resizing the 1600x1200 image by the next integer multiple of the golden ratio. The negative size indicated is calculated from extrapolating from 36mm in either direction using this multiplication factor (golden ratio).
Thus, each image represents a picture taken on the same type of film but on different film formats. All images are identical in size because they represent the same image.