Quote:
Originally Posted by
Hanson Yoo 
So the basic issue is that dpi in relation to printing has zero to do with an LCD screen or its resolution. All of those comparisons Jobs made between printed matter and pixel density were complete rubbish.
We've got a profound communication problem, then. I wish we could talk live. This would be much easier to explain with a whiteboard and color markers :)
At the heart, it's all the same thing. DPI, in the simplest view, is the number of black dots per inch (think monochrome laser printer). It is the number of discrete image units per inch. This is a direct analogue to a (monochrome) display pixel.
The link you sent me refers to a specific type of printing, where dithering is used to fake color by printing multiple colors in close proximity. But this is not the only method of printing. It can also be done by directly combining multiple colors to create the actual desired color on a given spot. And in this case, we can still speak of minimal image element per square inch, which again is an analogue to active display mediums.
But still, that "dithering" view of dpi is a useful illustration, as it's representative of how display technologies work: While the iPhone 4 display has a pixel density of 300 ppi, it really has a higher "dpi", in the printing sense, because each pixel is composed of three sub-pixels (RGB) that cannot be resolved by the eye and appear as a single color unit. Likewise your old CRT: each pixel is really from a triad of RGB phosphors being turned on to create a full-color pixel. But we don't speak of sub-pixel resolution; we cheat and talk about the equivalent pixel resolution. And so I do when talking in general about printing results.
Quote:
My thought experiment is this -- once you cannot see pixels, no matter how many you add, you cannot resolve any more than the point where you cannot see them. This is common sense. At 12", I cannot see pixels (and I have 20/20 vision thanks to the miracles of laser eye surgery). I have shown the screen around -- no one can see them.
I basically agree. And the well established science is that 1 arc-minute is the approximate resolution limit for humans. In many cases, depending on contrast and the particular colors adjacent to each other, and the rate of image motion, people will resolve only larger features. But you want to design to the limit, to the "best viewing" case. And for a handheld device about 8" away from a person with properly corrected vision, that's about 300 pixels-per-inch.
To deny that is to deny human physiology and basic physics :)
Actually, w/ imaging sensors, in most cases, the quoted pixel count is not the true res (like we're equating for ppi in this thread), but is typically just the (Bayer) interpolated res where the original sensor elements are actually equivalent to the dots in "dpi", insteaad of the pixels in "ppi", so to speak...
