From: Tetractys on
George Preddy wrote:

> In general, if you are sharpening at all then
> you are oversharpening.

This is an untrue statement. Sharpening is part
of any proper digital workflow. Look here:
http://www.luminous-landscape.com/techniques/process.shtml

> Global digital sharpening tends to flatten or even
> invert the 3D nature of an image, since contrast
> is strengthened pixel by pixel, without regard to
> subject depth or the optics at play. The picture
> has to be really soft or taken OOF to get benefits
> of digital sharpening above the cost watermarking
> the image "from digital."

Huh?

For a good primer on sharpening, look here:
http://www.luminous-landscape.com/tutorials/sharpness.shtml

or an alternative, try this:
http://www.luminous-landscape.com/tutorials/high-pass-sharpening.shtml

Here is another good tutorial on USM:
http://www.creativepro.com/story/feature/11242-1.html

Here is MS's version of the comment that Bayer filtering is
responsible for the necessity of USM in the first place:
http://www.microsoft.com/windowsxp/using/digitalphotography/learnmore/sharpening.mspx


From: Stacey on
Tetractys wrote:


>
> Here is MS's version of the comment that Bayer filtering is
> responsible for the necessity of USM in the first place:
>
http://www.microsoft.com/windowsxp/using/digitalphotography/learnmore/sharpening.mspx


Wow you stepped right into it, here's where George explains that's why Bayer
filtering is such a bad thing...

--

Stacey
From: Tetractys on
Stacey wrote:

> Wow you stepped right into it, here's where George
> explains that's why Bayer filtering is such a bad thing...

Not really. It's an issue that should be addressed
in any discussion of Bayer filtering -- the need for
post-process sharpening. You can't avoid it, and
the need is there.

George's problem is that he blames the Bayer filter
for global problems like entropy and unrest in South-
west Asia, and believes that a crappy camera based
on poorly-implemented Foveon technology is superior
despite overwhelming evidence and opinion to the
contrary.

So I'll step in it all day long, but call it what it is.
Bayer = water puddle;
Sigma = dog poop.


From: george_preddy on


C Wright wrote:
> On 5/30/05 9:16 AM, in article
> 1117462598.038828.135100(a)g44g2000cwa.googlegroups.com,
> "george_preddy(a)yahoo.com" <george_preddy(a)yahoo.com> wrote:
>
> >
> >
> > Celtic Boar wrote:
> >> Please take a look at the attached link. I am still trying to get the hang
> >> of this Unsharp Mask Thing. Are these oversharpened.
> >>
> >> Canon 20D - Raw - 75-300 IS Zoom
> >>
> >> Thanks.
> >
> > I couldn't find any oringals there so there is no way to evaluate the
> > images for sharpness. In general, if you are sharpening at all then
> > you are oversharpening.
> >
> > Global digital sharpening tends to flatten or even invert the 3D nature
> > of an image, since contrast is strengthened pixel by pixel, without
> > regard to subject depth or the optics at play. The picture has to be
> > really soft or taken OOF to get benefits of digital sharpening above
> > the cost watermarking the image "from digital."
> >
> Considering your reputation in this group

There are 17 different George Preddys posting here, which one are you?

> I hesitate to reply at all but for
> the sake of potential newer dlsr users I can't let this statement stand!
> Unless someone has optional in-camera sharpness cranked way up dlsr's are
> noted for producing soft images - this is by the manufacturer's design.

No, it is the Bayer design. Since all Bayer cameras (all digitals
except Foveon based) output images that are already upscaled by 400%.

P&S's have harsh built in digital sharpening that can't be changed.
DSLRs are no softer, they stop short of oversharpening everything
before you see the image.

My point stands. Digital sharpening ruins images because it plows
through every pixel blindly, instead of leaving sharpness/blur to
optics. That is why film lovers hate digital, the result is flat and
lifeless, often even inverted.

If you sharpnen digitally, you are simply exceeding the capabilites of
your camera at the expense of quality. The only proper way to handle
digital images that are breaking down is to print them at the size they
were intended. There are no acceptable algorithmic crutches that can
substitute for sensor count.

> Most all advanced amateurs and pros prefer it this way and will use some
> sort of post processing program and either leave their images soft or
> sharpen as desired. Many, many images benefit from sharpening and to say
> that "In general, if you are sharpening at all then your are
> oversharpening." is just plain wrong! As far as the rest of your statement
> about sharpening tending to tending to "flatten" or "invert the 3D nature"
> of an image - huh!

If your camera needs sharpening, the only quality solution is to
downsize the already upscaled image to what your camera's sensor count
actually supports.

Bayer images are properly viewed at their optical, not recorded,
resolution. That is, 25% of the recorded size, or a complete dedicated
RGB sensor triplet for every presented color pixel.

If you are printing on paper, this is done for you. At smaller print
dimensions the image will look very sharp due to the printer
automatically cramming all of the optical data within the choosen
dimensions. At the enlargement size that you notice unacceptable blur,
you should stop and buy a better camera. Digital oversharpening is
never a pro quality solution beyond that optical limit, though it can
certainly is an oft used, low quality approach.

From: george_preddy on


Tetractys wrote:

> Not really. It's an issue that should be addressed
> in any discussion of Bayer filtering -- the need for
> post-process sharpening. You can't avoid it, and
> the need is there.

Of course you can avoid it. Don't output at 400% the supported optical
resolution of the sensor. You need an dedicated RGB triple for every
ooutput pixel or obviously you are right, the recorded image will look
bad.

> George's problem is that he blames the Bayer filter
> for global problems like entropy and unrest in South-
> west Asia, and believes that a crappy camera based
> on poorly-implemented Foveon technology is superior
> despite overwhelming evidence and opinion to the
> contrary.

Like it or not, well focused Foveon images don't need sharpening in
post processing. Why? Because the images are recorded at the sensor's
optical resolution--one dedicated RGB tiple for every RGB recorded
pixel.

Bayers record a 400% upscale by default, since most use Bayers, people
are conditioned to think digital imagery always sucks. It doesn't, but
it isn't magic either, you need real live RGB sensors to see then
record RGB images. Not a sharpening algorithm applied long after the
shutter closed.

My 1DMkII, for example, is a 2MP camera. I knew that when I bought it.
The camera's 8MP, 400% upscaled output will obviously look too blurry
and too flat after sharpened into oblivion. How could an 8MP recording
not look horrid with only 2M pixels worth of actual full color optical
data? At 2MP, image quality is sharp, 3D, and life like.

People simply expect too much from low MP cameras. If you digitally
sharpen, you are in that group.