In article ,
(Fjx1) wrote:
I’d like to print an 8×10. The pic was taken with enough resolution to print at 300 dpi. When does one click resample, when not? I see different things in different books. Is constrain always checked?
The main reason for upsampling is to avoid the appearance of artifacts, _viz_ ‘pixelation’, due to inadequate resolution. This can’t create detail that wasn’t there to start with, of course, but it can make the ‘jaggies’ disappear, producing a reasonably smooth, albeit often somewhat blotchy, result. When upsampling by a large amount it’s generally best to do it in stages; my rule of thumb is to enlarge to no more than double the file size — about a 140% increase in the linear dimensions — in one go.
The three main reasons for downsampling are to reduce the file’s size, to prevent the rendering of details smaller than can be reproduced, and to make the appearance on screen a better predictor of the sharpness of the output. (I’m thinking of halftone screening for prepress as the intended destination: some different considerations apply to other types of imaging, like dithering as done by an inkjet printer or like the ‘stochastic’ screening sometimes used in high-end offset printing.)
Whenever in doubt, resize without resampling; in this case all you’re changing in effect is the rulers, the image itself remaining untouched — and it happens more or less instantly. Once you’re at the final print size you can evaluate the resolution, sharpness, &c., and decide whether resampling will be required.
Constraining proportions is the default, most likely to avoid the accidental distortion of an image by disproportionate scaling, which should be considered a ‘special effect’ rather than a standard behaviour.
—
Odysseus