if one is too stubborn to buy a sharpening tool, one can duplicate a layer then sharpen that layer and, using the luminosity blend mode, non-destructively sharpen one's image! or am i missing something?

Check out the free sharpening tools here:

http://www.thelightsrightstudio.com/

You can also look at the scripts themselves to see what they do, modify them to your own requirements etc. You'll learn a lot more than by just using the canned solutions of commercial products.

wow, that's easy. thanks for the tip.

Sharpen how much and for what device? That's the key! That takes a LOT of testing to a lot of output devices. The mechanics after that isn't rocket science.

You should probably read this too:

http://www.creativepro.com/story/feature/20357.html

I've read it, but...

Even within the bounds of contone printers, sharpening can appear differently. For example, if I send a file to both a ZBE Chromira and a Durst Lambda the apparent sharpness is different ( Lambda sharper) due to the way the file is printed, i.e laser v LED. Does any software allow for such differences? I'm not sure, but knowing how the monitor displays sharpness - varies depending on monitor- and how the specific printer reproduces it surely has as much to do with the final result as software. IOW, there are a lot of variables involved....

Best,

Mark

If I was going to sharpen without 3rd party plug ins, I'd use the PS workflow described here:

http://www.shutterfreaks.com/Tips/joukosharpening.html

I has one refinement that I'd never seen described in a sharpening tutorial before, using the "blend if" split sliders in the blending mode dialog box.

That's an excellent technique, I think I'll use that from now on, at least with pictures I can't sharpen easily. For some reason the action gave a horrible result, but doing it manually (I used smart sharpen instead of usm) created excellent results.

Thanks for the link!

Something Fraser came up with a long time ago....

thanks, all great stuff. that technique just mentioned with the 'blend-if' method is great! i'll be studying that method with the scripts and actions.

regarding one of the scripts mentioned earlier in the post, i used the TLR output sharpening script and it seemed to not sharpen as much as i thought. settings were standard, for use with 240ppi output, inkjet.

should i be using the fact that it didn't seem to be a huge difference to my eyes (100% view) as an alert that i've been over-sharpening? up until now, i've been doing two forms of USM: 1st - 100-200% radius 1 (sometimes .6) threshold 0 and 2nd - 10-20% radius 30-70 threshold 0. this on raw 10D photos.


thanks

steve

Maybe! At 240ppi you don't want to be too aggressive. I do everything at 360dpi and it definitely looks sharp on the screen. I find that if the image has been over-sharpened at the capture stage, the output will be too sharp. I use Smart Sharpen (a bit of a dog) for this as I need to see what I'm going to get. Sharpening is a subjective thing which is why I don't go for canned solutions other than for output. The chances of me remembering six months later what I used as output parameters is zero!

BTW, you can improve on the scripts just by isolating the parameters you need and rewriting it as an action. For example, you really only have to do the high-pass once then duplicate it. Left as an exercise for the reader ...

playing around a bit more tonight, i might be over thinking sharpening in terms of trying to nail down a certain process. i think i need to back off a bit you brought up a good point, trying to remember what was done. though there's ways around that too. sharpen to taste w/o getting artifacts or accentuating noise is probably the best way to think of it. i'm glad for all the samples/examples posted. i suspect i'll settle into a nice little store of techniques to draw on.

Actually, no. . .given a level of input in terms of camera, file size, ISO and edge width and a known output resolution, there are indeed knowable parameters for sharpening based upon the resolving capability of the human eye-which is known...

Creative Sharpening, or the local sharpening or smoothing may be an aesthetic decision rather than based upon image parameters. That must be done carefully and in a manner that will not break the image down the road.

While The Lights Right Studio actions and scripts can be useful, they do NOT offer any definitive sharpening parameters. That is their weakness.

The other problem is that using a low resolution device-a computer display-simply is not useful for determining either how or how much to sharpen an image. You can not accurately judge image sharpening on computer displays unless the final output IS the display. The -ONLY- way to determine the exact sharpening required for an image that will be printed is to make prints and adjust the parameters until optimal. Trial and error...

If that don't float your boat, and I'll be the first to admit that is a lot of work to do accurately (that's why free actions and scripts are free), I would suggest that professional level sharpening tools that have been created by experts and backed by extensive testing will save you a lot of time-time that you can spend doing other things...

If it's all science, why capture sharpen on a layer and allow you to adjust the opacity/layer mask to taste?

One reason would be to compensate for the variation of AA filter strength between different cameras of the same resolution.

I just took another look at PK, comparing the output sharpening of Inkjet 360 Matte with that of the corresponding settings in TLR ... by eye on screen and with Calculations and there's not a lot in it.

These are the things that I'd like to see in the next version of PK ... and may help me part with my dollars:

1. Previews, rather than what is clearly a bunch of actions dressed up as a filter.

2. Some refinement to the input types. For example "4x5 Positive Film". What does this mean? From my Flextight or a cheapo flatbed? Fujichrome or more grainy Ektachrome? Scanned at 1000ppi, 2000ppi, 3000ppi? You get the idea.

3. Internal use of Smart Sharpen. Adobe have put a lot of work into this and I'd like to see it used.

I hope you'll take this as constructive input.

So far I've been pretty happy with Photokit, and that's after trying several other things like Nik Pro (allthou I do like Nik's RAW Presharpener), various Fred Miranda things, and PS itself.

I was initially very interested in Nik's 'distance based' output sharpening, but to my taste it oversharpens the prints.

Well. . .I guess you didn't get the part about viewing an image on screen -NOT- being a good indicator of how and how much to sharpen? While we do have previews now in PhotoKit Color 2, the next version of PhotoKit Sharpener still won't have previews.

One might be able to get a better guage of how much to sharpen if you view your images at a 25% screen zoom in Photoshop, but it's not precise. Photoshop's screen views simply are not useful for real eval of sharpening. There are arguably 72 pixels/inch on a computer display. The Epson are arguably 360ppi (the real resolution is another debate). So, how ya gonna see 360ppi on a 72ppi display?

When you zoom into 100% the image is no longer anywhere near accurate to size. Look at an image at 200%? Pure science fiction. No printer could resolve what the screen is showing you.

The actual resolving power of the human eye with 20/20 vision is typically considered to be about one arc minute or 60 arc seconds. A rough approximation, at a distance of one foot, could be considered to be about 360 ppi for high contrast line pairs. The resolution requirements go down as the contrast reduces.

But a computer display, at the same distance is still 72ppi.

See the disconnect? See why you can't use a computer display to accurately evaluate sharpening?

So, what's a preview gonna do for you? Not a lot. . .

As far as Mitch matching PhotoKit Sharpener. . .well, he's tried to copy it right down to the terminology. But he still doesn't know the magic numbers. He hasn't done the extensive testing. On the other hand, Bruce Fraser, a noted and respected author and imaging expert HAS done the testing. . .perhaps that's worth something? His actions are error prone and his scripts are un-batchable from Bridge. So, workflow wise, you're still back to trying to delude yourself that you can "see" what you should do to an image on a display.


Flexibility. . .

I couldn't care less what the output sharpening looks like on the screen ... but I do for capture and creative (specifically haze reduction/local contrast). I've learnt to judge image sharpness with my old CRT second monitor at 50% ... something Bruce used to recommend ... but will probably have to modify things when it gets replaced by an LCD screen soon.

Personally, I'd rather see a positive response to my input (remember it's a $100 product you're selling) rather than bluster. But it's your business ...

Actually, that's 50% screen zoom for halftone reproduction, not ink jet printing. 50% is a similar pixel dither-4 image pixels to one display pixel as sending 2X the ppi to line screen ruling which works out 4 image pixels to one halftone screen dot

Bruce suggests 25% for higher resolution inkjet when printing at 1440 or 2880.

But again, looking at an image at 25% will not be a real accurate representation of the final output resolution. You need to judge the final print.

I'm evaluating overall image sharpness. I haven't yet decided what the output medium or size will be.

Pretty easy to illustrate this by doing eye-ball sharpening on a CRT and moving it over to an LCD (or vise versa). One or the other isn't going to look so hot...

Agreed, viewing and comparing prints is the ultimate evaluation of sharpening, and also for tonal/color results. In a traditional darkroom, a printer will make test strips for that purpose. Why someone has not come up with such an action or plugin for digital printing is beyond me. I think many will be more than glad to pay for it to save time and media cost.

Schewe, should you decide to take on this mission... (MI music)

I learned somewhere how to do this, not only for sharpening, but for many other editing.

Make sure all existing layer(s) are visible.
Create a new layer on top of existing layer(s). (If there is a single layer like Background, just duplicate it and ignore the next step.)
Hold down the alt key and merge visible.
Now you have a new layer to do whatever you want.

Cons:
Huge increase in file size.
Editing layers below the new layer no longer takes effect.

While quite useless for capture and output, I think previews might be handy for the creative sharpening step to get a general idea of what kind of effect each option will give.

The problems with previewing creative sharpening are

1.) It would take as long to calculate the preview as it would to actually execute the routine—a downsampled preview wouldn't be all that useful, so we have to run on the actual pixels. (This applies to the other sharpenings as well.)

2.) With the brushes, we don't know where in the image you're going to apply them, and what brush opacity you'll use to do so.

I've thought about letting the user select an area of interest, then previewing the selection, but then the question would arise as to why I don't let the user preview output sharpening (answer, because it may well scare the crap out of them). We aren't philosophically opposed to previews—we offer them in the new PhotoKit Color 2.0—but one of the things we're trying to do with PK Sharpener is to wean people from depending on the screen for making judgements that really need to be made from the print.

Some day, monitor profiles may include a sharpness parameter that would allow applications to render the same pixels with the same sharpness on all displays, just as we can now match color across displays, but that day is not even vaguely soon!

Meanwhile, if you really want to preview an area of the image, the most practical thing to do is make a selection, copy merged, paste to a new doc, and run the routine. Even if we implemented preview like we've done with PK Color, you wouldn't have access to the post-execution controls (layer opacities and mask brushing). For that, you need an actual doc open in Photoshop.

Basically, we've thought about it and decided that the results wouldn't justify the considerable effort. But if anyone has answers that address the above issues, I'm all ears.

(One of the reasons we offer a 7 day demo and a 30-day money-back guarantee is that we recognize PK Sharpener doesn't work for everyone. Given my limited resources, I prefer to focus on making it work better for the people it does work for than to try to make it work for everyone. But I take all feedback seriously.)

Something that became quite apparent to me after viewing some of my images on my friends Cinema Display (or other LCD's for matter). The LCD makes things look too sharp, at least to me.

Look too sharp on the LCD display or too sharp when printing out?

This illustrates the pitfalls of trying to eval sharpening on a display...

Where Sharpener falls down is with its "just trust the experts" approach. It may be just fine for novices and possibly even produce "optimal" results for a given set of fixed parameters ... however there's plenty of circumstances where it breaks down. Just look at the contone example above, and the one I gave for "4x5 positive". To my mind, a better approach would be to make the software adaptive to individual circumstances. For example, provide a set of sample images for the user to print out, tell them what to look for (over-sharpening artifacts) and allow them to feed this back as a setting they can select for subsequent prints with the same output parameters. A slider is all that's required. A similar approach could be used for capture. At least let people use their own judgement in what works best for them, rather that just a bunch of one-size-fits-all presets. To my mind, the presets are merely starting points.

As for previews, I recognize this is problematic to implement, but it's something I'd expect with a $100 package. I was encouraged by what I saw in the updated Color module. The problem with lightweight implementations like that used for Sharpener it that it's open for anybody to simply codify the sequence as an action/script and make it freely available. I don't buy this argument that you can't trust the screen at all to judge image sharpness and overall quality. It has to be learnt, and re-learnt for a new display, but something I consider essential in my day to day work. For example, I'm asking myself how large I can print this file before it starts to break down. Can I sustain these out of focus areas etc. I'm sure you're making the same assessments yourself. For creative sharpening (haze reduction etc) I'd like to see in advance which setting will work best for a given image, rather than guessing and having to back it off. Both USM and Smart Sharpen manage previews (the latter only just) why can't Sharpener even if it's only for a selection portion of the image at 1:1, 1:2 in the dialog? It probably means having to go beyond the Automation plug-in level though.

While I don't expect you to rush out and modify the product solely to my requirements, moving the product forward to encompass wider and more serious use is probably the approach I'd take over bluster and belittling potential customers as used by others in your team.

Ah, but they're starting points based upon extensive testing by some pretty knowledgeable people-such as Bruce Fraser, Mac Holbert from Nash Editions, Bill Atkinson and people like Mike-the guy hosting this forum. There's another guy who really likes PhotoKit Output Sharpener (Michael knows him) but his day job prevents him (and us) from mentioning his name...he uses his own "Capture Sharpener".

The "presets" are based upon science such as human visual acuity, the resolving capability of halftone and ink jet printers...many hundreds (prolly thousands) of digital captures & scans have been tested (Bruce and I tailored the scan sharpeners for our Imacon 848 scanners BTW and I have a lot of 4x5 chromes that I've used the 4x5 Positive Film on) and many hundreds (prolly thousands) of prints have been made to arrive at these "one-size-fits-all presets". We also have thousands of users all over the world successfully using PhotoKit Sharpener-we get a kick out of finding out what our users have used Sharpener for.

We know what to look for in optimizing images and we've been pretty successful at arriving at these "presets". I personally use them for my own work and print shows and I'm pretty fussy. Mac Holbert uses them at Nash Editions.

So, these "presets" must have something going for them...even if they are only "a starting point"...

I've never responded well to this big-names-use-it-so-it-should-be-good-for-schmucks-like-you thing.

No one - aside from yourself - is implying you are a schmuck.

I've bent over backwards to be as straight forward and as factual as I know how. Nowhere have I implied that any user is considered by me as a "schmuck". Far from it. . .the members of PG go way out of our way to be accessible and supportive of our customers and the general digital imaging community. I merely pointed out that real expert users. . .users who I consider truly knowledgeable in the field and whose opinions I respect a great deal have helped us arrive at what you call "presets". I'm rather proud of those "presets" and the fact PhotoKit Sharpener seems to have satisfied some rather fussy people.

You, on the other hand, seem to be reading entirely different posts than I've been writing. I don't know you from Adam and I have no idea what personal baggage you may be carrying, all I can say is I'm sorry if -ANYTHING- I may have said here is in the least bit insulting to you. I'll point out though that I'm not going to change my nature merely to suit you, I'm way too old and set in my ways.

:~)

Me neither, and I don't think that's really what Jeff was saying. While all our users are above average<s>, I'm not in the business of creating schmuckware or writing books for dummies.

There's no question that Sharpener isn't a fine product and has advanced the field. I just used it to sharpen a scan from a client's 6x7cm neg (I hate scanning negs) and not in a month of Sundays could I come up with the convoluted procedure that Sharpener uses. I do have some reservations about some other more generalized presets (or whatever you want to call them) that nobody has seen fit to address ... but frankly I've lost interest :-).

In a previous life I used to make a living out of selling software on the 'net and it used to piss me off when others steered potential customers to free alternatives ... inferior in many ways though they may be. So I understand where you're coming from. The difference is that I took this as a signal to really work on my software to add value ... features/elegance/simplicity/whatever ... to differentiate mine as much as possible. Everybody, myself included, gained from this. Whether you and your team want to take the preceding as constructive input is up to you. Being told that others have been well satisfied frankly means nothing to me when I can download and evaluate the software for myself to see how it fits into my workflow etc. As I would encourage others to do.

Cheers :-)

So how about a 400 page book on sharpening? I'd buy one ... seriously. I have an interest in the topic, which is behind my response to the original query.

Would you like a list of so called schmucks to contact ;-)

The demo runs fully functional for 7 days. There's nothing to keep either experts, big-names, or schmucks as you like to call them or anyone in between from making their own evaluations.

Real World Sharpening with Adobe Photoshop CS2 should hit stores late summer if I meet my deadlines, mid-fall in the worst-case scenario. It won't be 400 pages in the first edition but it won't cost $50 either.

The Capture Sharpening presets and Creative Sharpening presets really are just starting points. In Capture Sharpen, you can adjust the layer opacities of the light and dark contours, and in Expert mode you can also edit the masks. Unfortunately, that's not something that the Automate SDK lets us do in a preview—you need pixels open in Photoshop. When you combine that with the length of time it would take to calculate a preview (as long as it would take to run the actual routine), my feeling is that it makes more sense to make a richer set of presets that are tailored to specific camera types than to spend resources on generating a preview that can only be a starting point.

With the Creative Sharpeners, a preview makes even less sense. The brushes don't do anything until you brush, and the SDK simply doesn't allow brushing on previews. And again, the preview would take as long to calculate as the actual routine.

I'm acutely aware that the current Capture Sharpeners lump images into some very large buckets. For Sharpener 2, my main focus is to put the images in much smaller buckets automatically (we'll keep the old routines for those who find they serve them well). We plan to use metadata to identify the camera model, run an analysis routine to automate edge width selection, and tweak the mask generation to make it sensitive to working space gamma, among other things, but these should all help to make the presets much better starting points than they are today.

The Creative Sharpeners are what they are. If I ever find a way to automate creativity I know a fairly large number of people who'd want to kill me...

Other than adding some larger sizes for the Web/Multimedia output sharpeners, I haven't seen a reason to do anything to the output sharpeners. I've printed 10s of thousands of images through these, and I've only changed the layer opacities in a handful. Unlike all the other relationships we have to deal with in the capture and creative sharpeners, the relationship between input pixels and printed dots is fixed for any given print process, and hence is totally predictable.

I have no problem with people using or building free alternatives—in fact, I've published a great deal of the information people need to do so, both on creativepro.com and in Real World Photoshop—and I AM actively working on improving the software. But when the choice comes down to what I know I want vs what someone else thinks they want, I'm going to go with what I know I want. That's just the way we do things, and I'm not under the illusion that it will work for everyone. But I draw some small comfort from the demonstrated fact that when it works for me, it works for a fairly large number of others. If Jeff appeared to be blowing off the points you've raised, it's probably because they've been beaten to death by us internally, and he tends to forget that others weren't privy to those very long discussions, which often provoked a lot of testing that went nowhere.

Last but not least, if any of us considered you a schmuck, we wouldn't be writing at such length. I value your feedback, and it all gets factored into the mix. That doesn't necessarily mean that we'll build the product you want, but such is life. If I sell someone a piece of software that ends up being shelfware, I regard that as a failure rather than as a sale, which is why I tend to harp on the 7-day demo and 30-day MBG.

With that, I'd like to return the thread to its original topic, which was sharpening on a layer...

Great to hear. I can guarantee you a sale.

For sure I'll be trying v2 when it comes out. There's no metadata in my scans though but it looks like you're doing some other interesting things. The capture "presets" have always been my main concern. Also, when you offer a range of "Sharpener Effects" it makes sense to provide a visual means of distinguishing between them ... at least on a local level ... but I do understand the constraints you're working under and with the amount of processing required.

Thank you for the time and care you've taken to respond.

I concur with your sentiment 100%. The presets of a sharpening tool is analogous to a camera's auto exposure mode: they work well for most but not all situations. A set of sample images and a description of how a tool's settings can influence their sharpening can go a long way in helping the users. Because:

1. Images with different contents will require different sharpening settings for best results.

2. The different intents of the final products (web displays, print sizes, etc.) also calls for different sharpening settings.

3. Knowing what settings work well with a particular kind of image is a good starting point. It will cut down on the sharpen/print/change_settings interative process.

4. The set of images can be used to compare and evaluate different tools. It will make shoot-out or bake-off contests more meaningful. Currently they typically rely on a single image.

Come to think of it, perhaps reason number 4 is why sharpening tool makers DON'T want to have such a set of images available. Hope Bruce's new book will cover this.

A couple questions should be asked while criticizing the preset nature of PK Sharpener. First, can the variations of sources mentioned as examples previously really make a big enough of a difference to matter in print? And if they can, are the differences large enough that they can't be compensated for by adjusting the sharpening opacity?

Considering capture sharpening is a subtle thing to begin with, the differences between different sources might not be as big of an issue as you are making it out to be. Until you download the demo, try it out and make some test prints to compare, these objections are no more than armchair-quarterbacking.

Like you two, I was skeptical of PK Sharpener. The idea of a sharpening workflow was sound, and the science behind output sharpening was sound but I still had my reservations of how effective it could actually be due to the lack of control compared to tools like FocalBlade. After trying it, I bought it (twice since I switched platforms) because I found it to be the most effective sharpening tool available at the time (and it still is). The anal-retentive control allowed by other sharpening tools isn't really needed for great results and is only helpful in some off-cases of creative sharpening.

Bruce, all well said and very pleased to hear that you will be keeping the old markers for those of us who use your product and who have learned to use some of the presets (that could just as well be called generic nos 1, 2.., although that would raise a different set of howls, and the visual cues are helpful) as good starting points that we build from based on the shot, outpout, etc., and experience using PK.

And also happy to hear you will be continuing to focus in on making a product that works well for you, because it works brilliantly for many of us. Look forward to downloading Version 2 and giving it a spin.

Thanks for the kind words.

Those of you who would rather roll your own may find this column helpful.

http://www.creativepro.com:80/story/feature/20357-1.html

Bruce

Bruce

I'd be interested in hearing your opinions on the Smart Sharpen filter. Do you feel it's of much use? And if so, when? Have you evaluated it in it the context of that article?

John

I've struggled with Smart Sharpen since well before it was released. I think it has some potential as a sharpener for web and other screen-rez uses, but it's extremely slow, and I haven't yet been able to get it to do anything that I can't do quicker using a masked luminosity layer with USM. All in all, other than making images look nice on screen (hence web sharpening) I've found it disappointing

Just for the hell of it, I went back and play with NIK a little bit, specifically the Epson Output sharpener. Wow, if you want to see GNARLY on the screen, check it out. PK looks nowhere near as crunchy on the screen as NIK. I'd have to say I much prefer PK's output as well.

Bruce, you might be reluctant to do this, but do you have any comments on how NIK goes about doing it business? I DO like their RAW Presharpener module, and have used that in the past, and PK for output.

For a variety of reasons, no comment....

Okay, I can live with that. Pretty much tells me what I need to know anyway

Given that some of the gentlemen are here on this thread and i can't participate on the PK forum unless I buy it...

I've been doing some extensive testing with PK Sharpener as well as running to and from the lab with print orders as I try to get my head round the program before my time runs out!

Couple of questions:

Firstly, it has been explained extensively including on this thread how it is impossible to use the screen to verify sharpening. How then is it possible to use creative sharpening? I just can't see or predict what the results will be based on how much I'm painting in.

For example, I know that I get a sharp print at 15X10" (300DPI on a fuji frontier) using USM 250,1,0 for a regular 5D RAW file. I know that I need a Smart Sharpen setting of 250, 1.2 for the same result (albeit better defined). Those look sharp on screen and look sharp in print.

But the PK approach, though sensible and certainly suprisingly good in print (looks worse than horrible on screen, artifacts etc), leaves me without a 'reference point' for sharpening.
Is it that capture sharpen + Output sharpen = standard sharp print and any creative is adding more sharpness or less to specific parts? Or does capture sharpen + my choice of creative = standard sharp and the output is on top of that and should be ignored? Meaning should I work blind through all stages or should I sharpen to 'look on screen' at the capture and creative stages and stop looking the moment I hit output? If it is the first then how do I apply creative on screen or even know what the heck I'm doing?

Couple more questions if I may.

I like to give at least some sharpening to all parts of facial features and certainly want to sharpen even non edged areas on landscape photos such as the face of a rock, etc. Having no sharpening at all doesn't look 'right' on areas of skin, especially when it jumps to super sharpened eyes and mouth. All of the sharpening in both capture and creative seems to ignore the 'non edges' completely. How would I set up a standard minimum of sharpening over the entire frame without having to work individual masks which is very time consuming?

The sharpening seems to change the colour somewhat, especially the output sharpening. How do I stop this happening? Eyes change colour as does hair, they seem to take on a lighter tone which is no longer accurate and I don't see with either USM or SS. This is shown in print of course.

I've seen that hair seem to take on a 'ropey' appearance even in print, thicker than it should be or is with equivelent levels of normal methods of sharpening. Any comments?

I want to buy this program, hell, just in testing it I've spent almost the buying price on prints! But I'm still confused as to how to use it and the other points above.

If you can answer my questions and especially set me straight vis a vis first point I would be most grateful.

I have not tried PK sharpening, and cannot comment on how painting area for selective sharping works. But with another tool, painting is done after sharpening is applied and the results are observable. But you seem to wish to be able to preview before appling sharpening, which would be difficult with any tool.

There is no sharpening 'reference point' (as opposed to a gray scale for tonal evaluation, for example). The amount of sharpening and where you want to sharpen is subjective. The sharpening tools do not offer a preview before sharpening. Evaluating a sharpened image on a monitor is at best an approximation of what the final print will look like. All these are facts, and therefore the ultimate sharpening evaluation is by looking at the prints. But this can be a long and iterative process.

Hence my suggestions for the sharpening tool developers to offer two features to alleviate the process:

- Provide a set of images as 'reference points'. Being able to sharpen these images well with different settings will be great starting points for similar images.

- Provide a test strip printing action so that a portion of an image with different sharpening amounts can be printed on a single sheet for evaluation.

The very rough rules of thumb:

1.) Capture Sharpening + Output Sharpening (with no creative sharpening) will produce an acceptably sharp print. It may not be as sharp as it could be, but it should produce better results than one pass of sharpening. This is what I think of as full auto mode. (It can be completely automated using Actions.) The choice of edge width for the Capture sharpener is fairly critical—if you apply wide edge sharpening to a high-frequency image, you'll wipe out detail, if you apply narrow edge sharpening to a head shot, you'll get moonscapes on the skin tones.

2.) If you use Capture+Creative sharpening to make the image look as sharp as you want it on screen, then apply Output sharpening before print, you'll get a better result than 1.

3.) Beyond that, displays vary enormously in how they render sharpness, and you simply have to learn the behavior of yours. Some day in the far-off future, display profiles may contain a sharpness parameter to which the display architecture responds, but I'm not holding my breath.... That makes the notion of a "standard sharpen" for the capture and creative phases problematic. (Output sharpening is different because the pixels always get turned into dots in the same way.)

Capture Sharpening always applies some sharpening to the entire image, though we try to minimize it on large flat areas like skies. Some of the creative sharpeners ignore non-edges, but most don't—they just sharpen whatever you point them at. If you can give me more details I can probably answer the question better.

Sharpening always involves tonal shifts, since it's a localized contrast boost. That's why we return two layers—sometimes a high-frequency component takes on an objectinable 'glitter' that can be mitigated by reducing the Light Contour layer opacity.

I haven't seen Output Sharpening introduce color shifts. You can send me examples at bruceatpixelgenius.com, replacing the at with @ (to fool the spambots).