This came up in a current thread on an unrelated issue, in which Matus Kalinsky commented:
I haven't come across this idea before. Could someone elaborate a little on how this is done? Are there downsides/things to look out for?I have recently (finally) adopted the strategy where you scale down the original image (as a copy - you keep the original master file of course), do all your adjustment (levels, curves, adjustment levels, masks, etc ...) and once you are happy with that you scale it up again and paste the original image in as a new layer - so only the very last steps (final cropping, sharpening, etc) will take long. It really helps when your original image is 2000 dpi scan from 4x5 (cca 70 Mpix image - 400MB in 48bit color mode)
Thanks.
Bookmarks