Robert writes: > Excuse my ignorance but what is the logic doing > it this way instead of resample it directly to > the resolution you want?
It seems to give a better final result, as opposed to one single large downsampling step, although I have not been able to rigorously verify this. If you downsample from 1000 pixels to 10, for example, you get a blur, even after sharpening. If you downsample in multiple steps of no more than 1/2 at a time, the result at the end seems a lot more recognizable. I think this is because steps larger than 1/2 tend to lose information from intermediate pixels. If you downsample in steps and unsharp mask each time, details tend to leave traces in adjacent pixels that survive the next downsampling step. The result is a final image that contains more pixels that resemble important details of the original. It's actually probably less accurate than a single-step downsampling, but to the eye, it looks more like the original, because key details are more likely to survive (in exaggerated form, but that's what you need to make them obvious). ---------------------------------------------------------------------------------------- Unsubscribe by mail to [EMAIL PROTECTED], with 'unsubscribe filmscanners' or 'unsubscribe filmscanners_digest' (as appropriate) in the message title or body