[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]
Re: [APML] A Curves Exercise [was: Histograms, curving and layers]
(Warning: a long one and no spell checker on this computer, ut'oh)
Hey being a contrarian (is that a word?),
How about an non-politically correct, blastphemious histogram, totally
non-curved version of the exercise of digitally enhancing Matt's excellent
exposure.
Actually, there are us poor bloaks out here in the hinterland using
"Photoshop LE". The LE stands for "Limited Edition" and acurately describes
some of its functionality.
I'm well aware of Chuck Vaughn's championing of the original clear
explanation of the use of the "histogram" as a guide of being sure one is
neither "creating" or "destroying" nebulosity while digitally "enhancing"
astrophotos and that this is best accomplished within "Curves". It makes
sense to me.
But on the other hand I might have some reservations about the described
process as being "actually what is there" as it exists in nature. I also
believe his methods with the histogram are even more valid with tri-color
Tech Pan work than any color film, each of which responds differently to
emitted spectral lines, but still, depending on the chemistry, method, and
time of deveolopment, the results with Tech Pan of "what is actually there"
may appear quite differently. Tech Pan and tri-color experience go well
over my head.
Actually, I believe the current hard emphesis on displaying a "proper"
histogram at the "end" of processing through Curves is in large part due to
there being so few of the many original real film "pros" (as I like to call
them) left on the APML. The Dark Side has claimed them.
If one goes snooping about on many of their still-up sites, quietly
looking at their sometimes utterly amazing results within a histogram, one
finds great departure from the "politically correct" form.
Now, as for me, I'm at a big disadvantage "software-wise".
Utilizing Photoshop LE one has absolutly no ability to access a histogram
under Curves. A histogram can only be viewed within "Levels".
Somehow it just chaps my hide to go out and pay, what is it, $600 or so,
or maybe $300 or $400 on Ebay for a full copy of Photoshop. When I bought
Photoshop LE 5.0 it cost about $110 a few years ago.
That said, none of my photos are enhanced under the restrictions of a
histogram. Rather, I depend on a good astro-film for the color response,
then I just enhance what the film has recorded, spectrally correct or not.
In other words, I use the film itself as a guide, that's really all you can
do anyway, unless you purposely try otherwise.
But, I do begin under "Levels" with a *correct* histogram. My aim is then
to "possibly destroy nebulosity" (easy to see) during processing but never
to "create it". Consequently, I end up with many different shapes of
histograms. The film and what it has recorded dictate what I end up with.
You guys using "Curves" have to be changing the shape of them too. Which
means we both are doing "art". All Art is subjective. With art, one really
can only make "subjective" comments as to their own taste for the result.
Actually, no "clearly objective" comments can be applicable. In the end,
one can really only claim to "like" or "dislike" the result, totally
subjectively.
In other words, one cannot clearly define a standard in a purely
scientific manner which is applicable in all instances of the *final*
result. It is just plain subjective. Many may like it (a particular style)
and even emulate it. But there is no objective standard.
So, I do it differently. And I've no immediate plans to give Adobe more
money. People with the full version of Photoshop and who know what they're
doing will likely out-do me most of the time. But, it is instructive to
have a look at some of the world's best astrophotographer's ending
histograms. Clip, Clip, Clip away!
Under Photoshop LE (it has other limitations like that brilliant AV
technique) I consider all the results I get under Curves to be inferior to
the combined use of all the other available tools.
Here's my totally Curve-less version of Matt's fine pic, along with his
Curved and un-Curved versions. I spent about 10 to 15 minutes on it and
probably could do better with more effort. For some reason the page loads
very slowly.
http://www.homestead.com/sky4u/ToCurveOrNotToCurve.html
My humble 2.7182818... ¢ worth,
Jim Vineyard
www.deepskyphotos.com
----- Original Message -----
From: "Matt BenDaniel" <matt@starmatt.com>
To: <astro-photo@seds.org>
Sent: Thursday, December 05, 2002 10:37 AM
Subject: RE: [APML] A Curves Exercise [was: Histograms, curving and layers]
> Greg,
>
> The levels tool is useful but not as versatile as curves.
>
> Yes, my image is somewhat clipped on the low end. I did that to suppress
grain noise. Perhaps now that I know how to use SGBNR, I should fix the
clipping and deal with the grain by smoothing it away.
>
> To load a curves file, open a curves dialog in PS (Image->Adjust Curves
or Layer->New Adjustment Layer->Curves) and click on the Load button.
>
> Matt
>
> At 13:28 05-12-02 -0500, Greg Hartke wrote:
>
> >This is an excellent exercise to PS neophytes like me and I really
> >appreciate your effort in helping us, Matt.
> >
> >I downloaded both the uncurved and curved image and did some diddling. I
> >confess that after spending some time playing with the curves adjustment
to
> >the individual color channels, I got very frustrated at my inability to
get
> >a good image. I proceeded to go in a direction more familiar to me,
working
> >with levels. This may be the wrong approach but I was curious what would
> >happen. I suspect it's less versatile than curves but I don't enough to
say
> >that for sure.
> >
> >I looked at the information in the individual color channels and adjusted
> >the sliders on the low and high ends so that the range was more
compressed.
> >I moved them on either end such that they were right at the end of the
> >significant parts of the curves. I then adjusted the black point to 15,
15,
> >22 and looked at the result. After comparison to Matt's curved version, I
> >adjusted the center red slider for a little more red. The result was much
> >more like Matt's image, though not identical.
> >
> >I'm sure this isn't the right way, it's just what I could do. An
interesting
> >consequence was that the resulting histograms were all clearly quantized.
I
> >would assume that's the result of a compressed range and working with
only 8
> >bits of info.
> >
> >The curved version had interesting features. Am I mistaken or are the
blue
> >and green histograms clipped at the low end? My efforts resulted in
similar
> >clipping but I thought we were trying to avoid this?
> >
> >I really wanted to look at Matt's curves file but I'm embarrassed to say
I
> >couldn't figure out how to import and look at it. This must be simple but
I
> >couldn't manage it. Matt would you please elucidate? Sigh. Somebody point
me
> >at a command line Unix system so I can grep 'em to death. ;)
> >
> >Thanks.
> >
> >Greg Hartke
> >Sykesville, MD
> >
> >
> >
> >> -----Original Message-----
> >> From: owner-astro-photo@seds.org [mailto:owner-astro-photo@seds.org]On
> >> Behalf Of Matt BenDaniel
> >> Sent: Wednesday, December 04, 2002 9:39 PM
> >> To: astro-photo@seds.org
> >> Subject: [APML] A Curves Exercise [was: Histograms, curving and layers]
> >>
> >>
> >> Alan,
> >>
> >> The most important skill to master in image processing is curves.
> >> By far, you'll get more bang out of curves than any other tool.
> >> Both Photoshop and PictureWindow have good curves tools.
> >>
> >> I made an exercise for you (and anyone else who might be interested).
> >>
> >> I place an uncurved image of Central Auriga on my web site. Save
> >> the uncurved image onto your disk. Using solely a curves layer,
> >> try to make the color balance and contrast of the uncurved
> >> picture look like my curved one. All you need concern yourself
> >> with at this point is adjusting the R, G, and B curves in a
> >> single adjustment layer. That's it. I already took care of
> >> stacking, AV-ing and star shaping the uncurved image, so you
> >> don't need to worry about any of that with this image.
> >>
> >> Now, I don't claim that this exercise is easy. But if you do it
> >> right, you should be able to make the uncurved image look just
> >> like the curved one on my web page. The link to the uncurved
> >> image is at the bottom of the page:
> >>
> >> http://starmatt.com/gallery/astro/central_auriga.html
> >>
> >> I hope this exercise is helpful to you and others. Please let me
> >> know if you have questions, problems or have results to share.
> >>
> >> Matt
> >>
> >> At 17:45 04-12-02 -0800, Alan Voetsch wrote:
> >> >Hi Matt,
> >> >
> >> >--- Matt BenDaniel <matt@starmatt.com> wrote:
> >> >> Thanks. I'm always looking for suggestions for more
> >> >> articles.
> >> >
> >> >How about: making a copy to work on, setting black
> >> >points (whatever that is), making or using layers. All
> >> >I feel partially proficient at is adjusting contrast,
> >> >lightness and watching how much the image changes as I
> >> >move the curves slider around.
> >> >
> >> >> The order matters a lot. If one layer clips the
> >> >> image, nothing above it in the layer stack will
> >> >> rescue the lost info. I find that an AV layer works
> >> >> best if it is right above the source layer. When
> >> >> doing AV, I'll alternate adjustments to the AV layer
> >> >> and to the main curves layer. Their adjustment is
> >> >> intertwined.
> >> >
> >> >I understand the concept, but this sounds like there
> >> >are a lot of ways to learn a lesson the hard way.
> >> >Maybe another article on the sequential steps to use
> >> >throughout the whole process.
> >> >
> >> >Thanks Matt,
> >> >Alan
> >>
> >> --
> >> Matt BenDaniel
> >> matt@starmatt.com
> >> http://starmatt.com
>
>
> -- APML Archives at <http://astro.umsystem.edu/apml/> ---
> Unsubscribe at <majordomo@seds.org>
>
-- APML Archives at <http://astro.umsystem.edu/apml/> ---
Unsubscribe at <majordomo@seds.org>