Unsharp Mask how much?

[iphoto]

In my series of probably dumb question here is 2nd one. more to follow.

I understand that for presenting image on web, unsharp mask should be used...I wonder what are average or common settings or a range that is treated as best...it may change from type of photo etc but some common values which might act as baseline

your comments please

 

[ksmattfish]

I'll tell you how I do it (I got my advice from Doxx), but I'm sure there are other ways and opinions. I started out scanning from 8"x10" prints, and this method seemed to do the best job recapturing sharpness lost in the scan. Now that I have a neg scanner I use the same procedure when scanning from negs too. I'm usually looking to make a web friendly image that is 500 pixels across the long dimension.

I start with a file (scanned from print or neg) that is approximately 2000 pixels across the long dimension.

USM amount: 80%
radius: 0.8 pixels

Then I cut the dimensions in half, to 1000 pixels across.

USM amount: 50%
radius: 0.7 pixels

Then cut it in half again, so I have my final image at 500 pixels across.

USM amount: 25%
radius: 0.5 pixels

I don't mess with "threshold" at all.

 

[Digital Matt]

There's so many different variations on how to use USM, and each picture will need different amounts.

Here's two links for you.
http://luminous-landscape.com/tutorials/understanding-series/understanding-usm.shtml
http://www.luminous-landscape.com/tutorials/contrast-enhancement.shtml

 

[JohnMD]

Hi there
I find its best to only perform one sharpening operation and make sure that it is right at the end of the line. In other words, perform all other image editing (levels, brightness, contrast etc) at image mag 100% then resample image in one step to your final web size image (photoshop bicubic resampling does a perfectly adequate job here) and then sharpen. You are correct in assuming that actual sharpening values are likely to vary but as a starting point try;
Amount 75%
Radius 0.7
Threshold 0 (raising the threshold is useful in over noisy images)
and have a play around from there!