You do have to sharpen differently for different cameras, different lenses, different subject, different papers, and different printers - really!
Here are some ballpark guidelines that work well for me with a 5D II and which can produce very sharp prints at large sizes.
I do my sharpening using "smart filters" in Photoshop. (In virtually all cases - there are some situations in which some sharpening in ACR makes sense, but I'll leave that out of this description.) I add two smart layers for two different but complementary types of filtering.
1. Apply a SmartSharpen filter (in CS4) with the display at 100% magnification. Settings can very depending on "stuff," but a decent starting point is to set the top value ("amount") to 150 and the lower value ("radius") to 1.0. I would rarely increase either of these (with one exception mentioned below) but I might well lower them a bit. In particular the radius value may be lower if the photo was made with a particularly sharp lens - say a prime. It is not unusual for me to set radius as low as .8 or .7.
Exception: In an image with a lot of very fine detail I may start with SmartSharpen values that go back to an old Canon recommendation - amount: 300 and radius: .3.
2. Reduce magnification (25% or less) and apply a USM (unsharp mask) filter, also as a smart layer. Typically the three settings here are 12, 50, 1. The amount might be a bit less in some cases.
The first step increases the resolution of fine details and the second increases local contrast - you'll see the effects of the two methods when you try them.
These two processes are pretty much used by default on all image. I do not necessarily wait until all other layers, etc have been applied - keep in mind that I can return to Smartlayers later and change values if necessary.
Then I do a third "output sharpening" pass when I make a final conversion or do a print. Since your question relates to prints, I'll leave out the output sharpening I do for jpg screen images.
The idea of the final print output sharpening is to slightly
over-sharpen the image - someone described the effect as "crunchy." Save your working file before you do this,a and then duplicate the file. Working on the copy, first flatten the image. Work at 100% and go for a bit of over-sharpening, noticeably beyond what you think you'll actually want. This step is intended to compensate for the tendency of the ink to spread when it hits that paper. The plan is that by over-sharpening on the screen you'll get "just right" sharpening after the ink hits the paper. (And, yes, you ultimately do have to try different settings with different types of paper...)
Typical starting points for me, again using a Smartsharpen layer might be in the amount: 200 and radius: .3 range. Amount can go as high as about 250 and as low as about 100. Radius is most often .3 but it can occasionally need to be as high as .4 or rarely as low as .2.
One more thing. I don't do any interpolation to upsize the image. I simple set the image size where I want it and let the resolution fall where it may. As long as it is at least 180 this works quite well, indeed. (I didn't believe it when I first heard of this technique either - but I tried it and my skepticism was replaced by relief...) It makes sense that you would want higher resolution for smaller images that are likely to be viewed very closely and that lower resolution would be fine for very large prints.
And, as the previous posted pointed out, sharpness is not all just about, uh, sharpness. For example, work with curves (often masked to constrain them to only portions of the image) can increase mid-tone contrast and produce the impression of a sharper image.
Good luck.
Dan
--
---
G Dan Mitchell - SF Bay Area, California, USA
Blog & Gallery:
http://www.gdanmitchell.com/
Facebook:
http://www.facebook.com/gdanmitchell
Flickr:
http://www.flickr.com/photos/gdanmitchell/
Twitter:
http://twitter.com/gdanmitchell
IM: gdanmitchell
Gear List: Cup, spoon, chewing gum, old shoe laces, spare change, eyeballs, bag of nuts.
