Chuck Norcutt wrote:
> All I did was display Chris's full pixel image sample in PhotoShop at 31%.
> To my eye the grain was nearly gone but not completely so.
>
I don't understand this. The process PS uses to give a quick sample on
screen isn't the same that will be applied to actual image processing.
Viewing at an odd rez like 31% forces the screen renderer to do some odd
things tha pretty much always create visual artifacts that degrade/blur
detail.I think you are fussing out the grain in a was that actual
resampling to 31% would not do the same.
There are occasional threads here, including a recent one, on the theme
"How do I reduce a bunch of images for web display quickly and with
minimum effort/cost. People then dutifully mention a few free, quick,
easy to use apps. Nobody talks about possible quality differences in the
results.
I quit using the PS Prepare for Web process long ago, because I found it
made poorer images than other alternatives. The fact that the
alternatives also don't strip out EXIF data was a bonus.
I'm concerned with trying to make the reduced size web image reflect the
original as well as possible. Some processes work better at that than
others.
C.H.Ling wrote:
> I remember I have seen someone mentioned here that resize does not reduce
> noise, I'm not sure if it is true.
I've certainly said here that it doesn't "necessarily" reduce noise. In
practice, my experience is that sometimes it does reduce apparent noise,
sometimes not, and occasionally, it seems to make it more objectionable.
I've no theoretical knowledge, but have noticed the effect of down
sampling on noise seems to vary with the nature of the noise and the
amount by which the image is being down sampled.
There are so many variables involved, especially with film scanning, tha
most of us have only a limited sample of the real world variation. You
and I both scan @ 4000 dpi, but with both different scanners and
different software. Even with the same piece of film, some
characteristics of both grain and scanner noise will differ.
With 35mm film, I often fins that NR at full size stronger than I would
use for a print of large electronic image does a wonerful job of
eliminating grain/noise in a reduced size version.
Fernando was kind enough to send me a full size sample of his recent
image. Here's what my NI application to it looks like.
<http://www.moosemystic.net/Gallery/Others/FGonzalezG/BahiaSamp.htm>
One difficulty with NR like this is that our eyes, at least mine, tend
to interpret the noise reduction as a loss of fine detail. The edges of
the grain tend to make things look sharper. In this case, even if a tiny
bit is lost, it doesn't matter, as the whole point is to prepare for
down sampling.
Here's the downsized sample inserted in the full image. Pleas, please,
please, pay no attention the the gross differences in colors, etc. and
concentrate only on the grain/noise. Trying to match the processed image
would be lots of low return work.
<http://www.moosemystic.net/Gallery/Others/FGonzalezG/bahiadeM3.htm>
Notice how all the odd granularity of the original post is simply gone,
with much more detailed and subtle cloud and tonal detail replacing it.
ow much of the differences is due to the pre-resizing NR and how much to
better down sampling processing, I don't know.
I do know detail is better and grain gone.
Moose
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|