On 5/17/2016 4:25 PM, bj@xxxxxxxxxxxxxxx wrote:
I am somewhat dissatisfied with the resolution of my images.
Please see a sample of 3 - view large; at
http://zone-10.com/tope2/main.php?g2_itemId=20133
. . . I really expected a lot better.
What do you think?
I think we've done this dance before. Do you know the story of the man and his
dog?
"I told him to sit, and he wouldn't"
"So I kicked him, and he still wouldn't sit"
"So I denied him food; even after a couple of weeks, he still wouldn't sit"
"So I beat him, and he still wouldn't sit"
"So I beat him harder, with a bigger stick, and now all he'll do is lie down and
play dead."
A little harsh? I don't know. I've written this stuff before, and you persist in mistreating your light, and then
somehow expecting the results to be different the next time. These images are like so many you have posted before,
suffering from just the same ills.
1. You refuse, or perhaps are constitutionally unable (?), to use exposure compensation with scenes like these, with
high DR and significant areas of very bright yellow/orange/red. If you won't bracket exposures, SET EV TO -1.0 for such
scenes.
2. You casually mention a filter. Is this one you have tested? Gary's tests showed that even a high end brand may
produce a filter that significantly reduces sharpness.* For those of us without a vertical auto collimeter, A/B tests
are the only way to tell. If concerned about sharpness DO NOT USE A FILTER, unless/until you have done A/B tests with it
and confirmed that it doesn't degrade IQ.
3. You insist/persist in using either JPEGs or a second rate Raw converter. I don't know about the E-3, but I've taken
almost 14,000 shots with an E-M5 (~2,500 with 12-50) and processed at least hundreds from Raw. There is a great deal of
recoverable highlight detail in these Raw files. But you can't get at it with FS or Oly Viewer.
4. You insist/persist in using second rate sharpening. The USM sharpening in FS increases contrast at edges, often
creating artifacts, but does nothing to actually recover detail lost to AA filters and Bayer array color interpolation.
Deconvolution, although on the surface perhaps appearing to act like USM, is actually quite different and CAN recover
visible detail.
I'm sorry you lost you house, and dog, to a fire. I'm sorry about your troubles with your main outboard engine. But
neither of those need affect these technical aspects of your photography.
"Why is he beating this dead horse of highlights, again?", you may well ask. "Because if affects detail resolution, the
subject of your post!", I holler back. I've explained how/why many times, both in response to your posted images and
those of others, and it seems to done little good (except for Jim). For now, IT JUST DOES.
Looking at sharpness alone, a detail from image #2.
<http://www.moosemystic.net/Gallery/Others/BSwale/Leaves.htm>
You can see that there's a lot more fine detail available than you've found/shown. If you did any USM sharpening in FS,
that is probably slightly degrading the results I get with deconvolution using Focus Magic.
========================
On 5/18/2016 6:49 AM, Chuck Norcutt wrote:
What I think it that you worry too much. Your images look just fine to me. Take the image with the sheep, for
example (an E-M5 image). There are thousands and thousands of blades of grass in the foreground. Even though they
can only be a pixel or two wide they're well resolved at 100%.
I disagree, strongly.
<http://www.moosemystic.net/Gallery/Others/BSwale/Sheep_Grass.htm>
========================
This is trickier; it would be easier and more effective without the blown highlights. They first have to be dealt with,
then 'sharpening" has to be done without re-blowing highlights.
<http://www.moosemystic.net/Gallery/Others/BSwale/Leaves2.htm>
This has even worse blown highlights. There are several leaves which have gone completely white.
<http://www.moosemystic.net/Gallery/Others/BSwale/Leaves3.htm>
When all channels clip, the pixel becomes 255,255,255, pure white. Nothing to be done about that at this stage. -1.0 or
-1.3 EV and highlight recovery in Raw conversion would likely retain them. As you can see, blown and highly compressed
highlights create something like blobs of detail-less color. To the extent that un-compressing recreates something like
the original tonal relationships, visible detail is recovered.
Yes, "underexposing" to avoid highlight clipping does indeed make for a darker image. BUT, especially with subjects like
these, shot at base ISO, the midpoint may be easily restored, perhaps most easily with a Levels tool, without noise
problems (especially if one uses a capable NR app.)
Tough Love Moose
* "Contrast was slightly lower with the filter, but
lower resolution was the most important factor in image deterioration.
Please note that this particular filter is not indicative of Vivitar or
Vivitar VMC filters, in general. It just tested as a poor sample. Other
filter makers, even the most highly regarded, have been found to have
poor samples in selections taken from used and new stocks of filters.
The use of the term "poor" means star test images, viewed on
a vertical auto collimeter, which show images that are: multiple and
overlapping, fuzzy, off center, and images which rotate when the lens
is rotated. More often than not, only one of these faults are found in
an examined filter. These filters (including the test filter) often look
perfectly good when examined without the aid of instrumentation!"
--
What if the Hokey Pokey *IS* what it's all about?
|