Sharpen _before_ resize?

[efunc]

I've always been of the school that adjusts exposure, levels, contrast, etc, then resize for web (800x533) and finally sharpen (USM 65%, 0.8, 0.3).

However I'm noticing that at whatever settings, and however subtle I set USM in CS4 I get hasher results than I would like on my my small 800 pixel versions. My original 4272x2848 images when sharpened however look beautiful. I've gone ahead and resized one of these sharpened images down to 800x533 and seem to have acheived a subjectively more natural image than the post-sharpened version. This is in no way conclusive and I haven't tested this extensively, however I wonder what best practice out there is?

post-sharpening seems most logical, however there may also be some sense to sharpening the high res version instead since there is much more detail and more pixels to interpolate and acheive smoother edges, before resampling down. any thoughts?

 

[blue max]

Sharpening is very subjective. And different images require differing treatments. Foliage can be awful if over-sharpened, whereas man-made objects see to lap it up.

You may find that a different setting on your reduced pictures do the trick.

If it's only to be viewed on screen, it's whichever way gets the results that you are happy with that wins the day.

Graham

 

[STARRIDER]

Sharpening should be the very last thing you do

 

[admirable]

I did a search on here and found a couple of sharpening tutorials,

High Pass Sharpening


Here's another I found

 

[efunc]

Sharpening should be the very last thing you do

Thanks. that's been my usual practice, but i was just curious if it was always correct because I've been so disappointed with the results lately.

 

[efunc]

I did a search on here and found a couple of sharpening tutorials,

High Pass Sharpening


Here's another I found

I've only been using unsharp mask, so these will be useful. lots to learn. thanks.

 

[Beagletorque]

I always process for the biggest print size, sharpening in the midtone layer last of all.
I then use this pic as a starting point for everything else. When I resize (for web or whatever) I sharpen again, last off all, at an appropriate amount for the pic at hand.

 

[Chaz Photos]

I've always been of the school that adjusts exposure, levels, contrast, etc, then resize for web (800x533) and finally sharpen (USM 65%, 0.8, 0.3).

However I'm noticing that at whatever settings, and however subtle I set USM in CS4 I get hasher results than I would like on my my small 800 pixel versions. My original 4272x2848 images when sharpened however look beautiful. I've gone ahead and resized one of these sharpened images down to 800x533 and seem to have acheived a subjectively more natural image than the post-sharpened version. This is in no way conclusive and I haven't tested this extensively, however I wonder what best practice out there is?

post-sharpening seems most logical, however there may also be some sense to sharpening the high res version instead since there is much more detail and more pixels to interpolate and acheive smoother edges, before resampling down. any thoughts?

why not post two copy's one of each method to show what you are getting.
For me sharping is the LAST thing done before saving with new name