I've recommended this before but, for an in-depth understanding of
digital image sharpening as outlined here by Moose, you should read this
book by the late Bruce Fraser and Jeff Schewe.
<http://www.amazon.com/World-Sharpening-Photoshop-Camera-Lightroom/dp/0321637550>
I don't profess to be anywhere near as good at sharpening as Moose is or
to take the pains of sharpening parts of the image differently under
masks. But I do sharpen images differently based on size and intended
usage. The best sharpening advice I've ever gotten comes from Fraser's
book. It's just a rule of thumb and I use it because it's easy to remember.
If you're sharpening for a display screen you can simply display the
image at the intended screen size and sharpen until it looks right. But
sharpening for prints is different because of the difference in
resolution between the display screen and print. Your display screen
probably has a resolution of about 90-100 pixels/inch. My own Dell 24"
display has a vertical resolution of 1080 pixels and vertical height of
about 11-1/4" giving a resolution of 96 pixels/inch. Prints, on the
other hand, are usually made at resolutions of 250-300 dpi with higher
resolutions used for smaller prints (viewed at reading distance of about
10") and lower resolutions used for larger prints viewed from afar.
Fraser's rule of thumb recommendation for sharpening prints is to first
size the image for the intended size print size and resolution. For
example, a 5x7 print at 300 dpi would be cropped at 5x7 ratio and sized
at 1500 x 2100 pixels. Then, before sharpening begins, you adjust the
size of the image on screen to compensate for the difference between
screen and print resolution by displaying the image at 96 ppi / 300 dpi
= 0.32 or 32% of a 1:1 pixel view. Had we intended to print at 250 dpi
we'd have 96 ppi / 250 dpi = 0.38 or 38 % of the full pixel view.
Fraser's own examples in the book give the same computation but using
older screens with lower resolution giving a guideline range of 25-30%.
I don't think there's a lot of precision required here and rounding to
the nearest 5 percent is probably fine.
Once you have the image cropped, sized and display size adjusted you are
advised to sharpen, not until it looks just right, but until it looks
just slightly "crunchy"... that is, just a bit over-sharpened. The
reason for the slight over-sharpening is that the printing process
itself tends to soften the final result. Unless slightly over-sharpened
the final print will look less sharp than desired. So, what is slightly
"crunchy"? That you have to determine by experiment but, if you're
doing your own printing you can readily determine that with a couple of
test prints. Once you've seen the degree of "crunchiness" you'll
recognize it in future. If you're doing a large and expensive print you
might want to make a small test print by cropping a small section (with
important detail) out of the main image and printing that at the same
resolution before committing to the large print.
There's a lot more than that to sharpening (including noise reduction
and "capture" sharpening vs "creative" sharpening vs "output"
sharpening) but if you know only that simple rule of thumb (for output
sharpening) it may already make an improvement to your prints.
Chuck Norcutt
On 3/30/2013 11:54 PM, Moose wrote:
> On 3/29/2013 4:06 AM, Brian Swale wrote:
>
>> I see Mike used an Olympus SP350 - a camera I had not previously heard
>> of. An 8 MP camera as well.
>>
>> Timing and light are critical components which Mike has nailed, but it is
>> also
>> sharp.
>
> Digital sampling ALWAYS softens edges - ALWAYS. The original capture of a
> subject, scanning an analog source and
> resizing an image are all cases of this. I wrote at some length about this,
> with a thought experiment and examples, on
> Zone-10 over five years ago. If you read this and the following three pages,
> perhaps you will understand what is
> happening.
> <http://zone-10.com/cmsm/index.php?option=com_content&task=view&id=149&Itemid=1&limit=1&limitstart=1>
>
> Mike uses post downsampling (re)sharpening. He's done a good job here. My
> eagle eye sees just the tiniest signs of
> sharpening artifacts, just about perfect for the display size (i.e. nobody
> else will likely notice them, but they will
> find the image 'sharp'.) (For printing, more sharpening would be better.)
>
> (Brian, pay particular attention where I talk about the decisions the JPEG
> engine must make, and why they may not be
> ideal for all uses.)
>
>> I doubt that my E-3 could give that resolution.
>
> Resolution at the camera level is only loosely associated with a sharp
> appearing web image (or print, for that matter).
> I have posted images from rather soft original Raw files that look tack sharp
> on the web. Other people post web images
> from very high detail originals that don't reflect the quality of those
> originals.
>
>> I'm getting increasingly dissatisfied with my device.
>
> If you were doing scientific work, your equipment regularly produced less
> accurate results than identical equipment used
> by others, what would you conclude? There are two leading candidates:
>
> 1. Your equipment is faulty, in need of repair or calibration.
>
> 2. There is a flaw or flaws in your technique in using the equipment.
>
> Yet you have proposed a third, logically inconsistent alternative:
>
> 3. The basic design and/or manufacture of the equipment is at fault.
>
> Now I happen to think the E-3 and E-5 were underwhelming cameras, but because
> of too little too late, relative to the
> competition. They are not incompetent in their basic functions. Tests when
> they came out and use by many, many people
> have shown that they are capable of first class, sharp images.*
>
> You are really stuck with 1 & 2 above, not 3. Now, if you want to update to a
> newer camera, that's just fine, there's
> some great equipment out there. However, if no. 2 is true, you are headed to
> disappointment with new equipment, as well.
>
> I've railed at you about software before, and you make excuses** or just
> ignore me (Which is fine, but why are you then
> back over and over again, complaining about the same problems?)
>
> So let me say it straight. Back in the film days, shooting slides, you got
> the best you could get when the film came
> back from processing. That is simply no longer true. Post processing is how
> the images you envy come to be the way they are.
>
> Your insistence on shooting JPEG only, processing those images with free,
> second rate software and not spending time to
> learn how to use software to best advantage for your images is holding you
> back, and will continue to do so, new camera
> or not.
>
> Back to Mike's image and your dissatisfaction with yours. Down sampling an
> image so that it retains much of the
> character of the large original is a tricky business. Using free software to
> do so, without comparing it to the best, is
> just asking to end up with poor apparent sharpness in web size images.
> Finding a good combination of down sampling and
> resharpening software and appropriate settings for them is crucial.
>
> Testy Moose
>
> * Send me the camera and a good lens, and I'll show you that it can easily
> take excellent, sharp images. ;-) Or turn
> on Raw, take 2-3 careful shots, and send me the ORFs.
>
> ** Often about money, and here you are, talking yourself into buying new
> equipment.
>
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|