Jman13
Online
• • • • • •
Upload & Sell: On
|
The issue is that some users and uses do require an image to be interpolated up in size (not just down!) to such an extent that anything that will show up at 100% view will be easily seen in the print. The better the per pixel, 100% view, the more I know I can push the print up in size when needed.
But it doesn't matter. We're comparing two different cameras here. Obviously, if a camera is better per pixel than another of the same resolution, it will be better in a larger print, but if you have a 12MP camera on one hand, and a 16MP camera on the other, but the 16MP camera is only 5% noisier at the pixel level, it will still produce a cleaner print at ANY size.
Let me see if I can make an analogy that will make this clear. My definition for noise in this analogy is simplistic, since noise is random, but it is valid for how noise will appear in print.
- Camera A produces files that are 6000x4000, or 24MP on the nose. We'll describe the noise by saying that it affects things about 5 pixels wide. (assume a 'grain' of noise 5 pixels wide).
- Camera B produces files that are 4500x3000, or 13.5 MP. Let's say this camera produces less noise than Camera A, but not orders of magnitude better. Say this noise is 4 pixels wide.
Now, pixel level noise on camera B is 20% better than camera A. That's a pretty big difference.
Let's print these images. Let's say you want to print a very large print, say, 60 inches wide.
Camera A will put 100 linear dots in each inch of print (6000 pixels/60 inches), and the width of each 'noise grain' in this example would then be 0.05 inches wide on the print
Camera B will put only 75 linear dots in each inch of print (4500 pixels/60 inches), so the width of each 'noise grain' for this print would be 4/75, or 0.053 inches wide on the print.
Now, Camera B, despite having 20% better noise control at the pixel level, will actually have slightly MORE visible noise on the print (albeit in this example, an imperceptible amount). These ratios will hold for any print size, large or small.
I think the easiest way to think about this is that when you are viewing images from two different cameras with different resolutions on screen at 100%, you are viewing them at different effective sizes, since the output of your monitor is a fixed DPI. The average monitor is between 72-96 dpi. When viewed at 100% on screen, if you are viewing a 12MP image and your monitor is 96 dpi, you are effectively viewing that image at a size of 44" wide. (4250 pixels wide at 96 dpi).
When you view, say, a 16MP image at 100%, you are effectively viewing that image at a size of 51" inches wide. If the noise appears identical at 100%, you can make the print 7" wider with the same visible noise with the 16MP image. It stands to reason that when printed at the same size, that noise gets magnified in the smaller image, or shrunk in the higher res image.
|
