On 11/21/2014 5:55 AM, Sawyer, Edward wrote:
Resolution and contrast are the same thing, just measured at different
frequencies.
Nope, I disagree. Resolution is the ability to render different brightness values across sharp brightness gradients
without blur/softening the gradient. Contrast is the ability to retain the original actual values of the brightness
differences. With film, and conventional photography, the interactions between the two are fairly subtle, and the tools
for handling them after exposure relatively weak and imprecise.
With digital capture and editing, that's no longer true.
Resolution - With adequate data about lens characteristics, almost all loss of resolution to lens aberrations may be
recovered. "Post processing can¹t add information. If it¹s not there ..." But it is there, in fact, with a proper PSF
and deconvolution. More general purpose deconvolution can't do as well, but can actually recover apparently 'lost' data.
As MikeG pointed out; in theory, deconvolution can't recover diffraction losses. In practice, as I at least partially
demonstrated with this example, a fair amount of diffraction loss may be recovered at the level of actual sensor
resolutions we are working with today.
<http://www.moosemystic.net/Gallery/tech/DiffractionContrast/diffracted_sweep.htm> In practical work with real images, I
have found this to be true.
Contrast. In the case of normal photographic lenses, all lost contrast may be recovered in post. I offer below some
examples far worse than anything a lens will do as proof.
It¹s not like you can¹t have both, there¹s no inherent tradeoff/balancing
necessary per se.
Perhaps not in theory, but in the practice of lens design and manufacture, yes,
there is.
There¹s also no such thing as ³something that may be corrected without
loss in post² Post processing can¹t add information. If it¹s not there
due to low contrast, nothing in post production is going to help. Low
contrast = low quality, bottom line.
I disagree, diametrically. Perhaps this will illustrate what I'm talking about.
<cid:part2.02030800.02010900@gmail.com>
1. Imagine the Line is one, high resolution, part of a lens test target seen at 100%. First the target, then as rendered
on sensor by high resolution, very low contrast lens. This is more extreme than any real world lens. I've set the
brightness of the black portion to 127 of 256 levels of brightness and the white part to 128.
Obviously, it is possible to completely recover the data. As long as the lens accurately records the v. sharp switch
between black and white = resolution.
2. So, how about subtle gradients? Here I've taken a gradient from pure black to pure white, reduced contrast by 90%, so
the human eye can barely see the change in brightness across the image. I've done it in 16 bit, so that there should be
plenty of values in the middle range to retain the tonal detail.
The posterization in both the original and the recovered versions is an artifact of conversion from 16 to 8 bit for a
web JPEG. It's not there in PS before conversion, but change of Mode to 8 bit creates it. I don't think it should be
there. Perhaps imperfection in the gradient tool I used?
In any case, recovery is the same as original.
3. And a high contrast pattern. Contrast reduced by 96%, then more than
recovered.
In Contrast Moose
--
What if the Hokey Pokey *IS* what it's all about?
--
_________________________________________________________________
Options: http://lists.thomasclausen.net/mailman/listinfo/olympus
Archives: http://lists.thomasclausen.net/mailman/private/olympus/
Themed Olympus Photo Exhibition: http://www.tope.nl/
|