Andre said:
I am not sure if this all comes down to eye of the beholder. Don't you
think that photo 2 has a green tint whilst photo 1 looks fresher? I
jsut feel that photo 2 lack oomph. Sorry can't be more arty with my
prose there!
Yes, this always comes down to individual perceptions, as well as to
individual preferences, and also monitor calibration. And what a dull
world it would be if Dali and Monet painted the same style.
As I see it, without reference to 'right vs. wrong', I would say that
Photo 1 is quite biased toward extra red over the entire range, with the
sky additionally being clipped to 255 in the blues. I see Photo 2 as
having the mid-tones and shadows also quite biased toward red, and also
somewhat less so biased toward magenta (lacking green), with the sky
pretty good on the right side, and too yellow (lacking blue) on the left
(I assume an effect from the angle of the sun, perhaps). As far as
'fresher and 'oomph', I guess those are pretty subjective and dependent
on personal interpretation. I would say that Photo 1 looks more
contrasty to me. Maybe that's 'oomph'?
How did you find that info in PS?
Two parts to this answer. First, I would suggest you keep the info
palette open. It will display the RGB values for any point you hover
your cursor over. Very informative. I use it constantly.
Second, it would have been clearer if I had said 'I set the BP to
20,..." & "I set the WP to 240,..." Essentially, there's some logic
that says if you make the blacks black and the whites white (and
sometimes make the grays gray), then everything will fall pretty much
into place, and only minor adjustment will be remaining. You can find
the blackest area in your image by opening the levels dialog box (RGB
view) and holding down the Alt key while sliding the left of the 3
triangles, the whitest area by doing the same with the right triangle.
To set these areas to values you want (I do this in curves, but it can
be in levels, too), first set the values, then apply them. More
specifically, in either the levels or curves dialog box, look for the 3
eyedroppers at the lower right. Double click the left one (black) and
set the RGB values to 20,20,20 (or 15's, or 12's - your preference).
Set the right hand one (white) to 240,240 240 (or something else close,
but = values) This is a one time deal. PS will remember these until
you change them. Next, from levels or curves, click the black dropper
and then move your cursor over the blackest area of your image, which
you previously found. Check the info palette and move your cursor
around that area a bit to find the darkest pixels, or a pixel that seems
to be a good average of the range of darkest pixels, and click on it.
This will adjust the black point of your image to match what you've set.
Then click the white dropper and repeat for white. Your image will in
most cases just snap into pretty close color balance.
Agreed. A lot of my scans seem to blow the sky.
This is a philosophy thing. Many devices (read cameras, their firmware)
and scanner drivers and one-hour photo labs will clip the shadows and
highlights so they have a smaller range to fit into 255 degrees of
level. This makes the result look more contrasty (maybe oomph? or
snappier?) because if they discard the upper and lower tiers of
information, then they can put more space between what's left. Other
devices work to preserve all information, and must place it all into the
same 255 degrees of level, and it looks less contrasty (flatter?). Some
people prefer it one way, and some prefer it the other. The 'capture it
all' approach allows *you* to decide what you want to discard, and how
you want to discard it, via post processing. On the downside, it's
generally more work for you. Again, a personal preference thing, not a
right/wrong thing.
Sorry can you explain what 4 panes of histograms. Are you in PS? Do
you mean RGB, R, G and B? Sorry that is my lack of PS knowledge.
Yes, exactly, in PS. RGB, R, G, B. Sorry, my fault. 4 panes is not
'official' terminology. I was just trying to save words, and it was
confusing.
yes, I also feel there is more substance to photo 2 but I am not hit
by the vibrancy of it compared to photo 1.
Again, a personal preferences issue.
Photo 1 Dimage
Photo 2 vuescan.
I am left with the dilema. More detail in 2 but, to me at least, more
vibrancy in 1.
Andre
I can't help with this choice, but at the end of the day, you should use
the one which gives the visual interpretation, the 'message', that you
prefer. And it is often a trade-off.
As a possibility, you could try working with Photo 2 to see if you can
increase the vibrancy, thus ending up with the best of both. One way is
to change the Graph Low and Graph High points in VueScan to .35/.65, or
..4/.6, or something like that (more contrast) Another is to increase
the contrast of the image in Photoshop. If you aren't comfortable with
Curves, try using the Brightness/Contrast tool, though I can't really
condone using this tool, as it destroys data (it itself clips). If
you're willing to venture into curves, make a new curves adjustment
layer and on the RGB screen, set an anchor point right in the middle of
the diagonal line (just click there), then move up and right to the 3/4
point and click/grab the line and start dragging it up (make an 'S'
curve). This will increase contrast (making the curve 'steeper). As a
next step, you will probably notice that you have darkened the shadows
and lightened the highlights by doing this. You can play, if you want,
with painting out the areas that were 'damaged'. Select the adjustment
layer. Select a 100-200 pixel soft round brush (or whatever size fits
best), set the foreground color to black, and start painting away on
your image. Erase changes with white.
You could also play with Hue/Saturation.
And if all of this doesn't get you where you want to be, or if it
entails more effort than you'd prefer, then stick with the Photo 1
approach. As I've said, there's no right or wrong to any of this stuff.
Just go with what ultimately feels best for you.
Best of luck with your endeavors, and keep us posted on your results.
Regards,
Mike
