Inkwell: Authors and Artists

    
(replying to # 24) 

All of the above < smile >.  Some of my methods were definitely
problem-driven.  The tarnish repair is a prime example.  I didn't read
about that trick somewhere else.  I was just messing around trying to
figure out some easy way to repair tarnish and wondered what would
happen if I combined color selection tools with a curves adjustment. 
It worked far better than I expected.  I thought it would, maybe, cut
my work in half, but it almost completely solves the problem.  It is
the hack I'm proudest of.  

But, in preparing to write the book, I did a lot of research.  In
fact, the book took a full year to right, which was twice as long as I
had planned.  That's because I ended up having to explore a lot more
different options then I thought I would, just so I could present the
reader with a variety of different ways to get to where they wanted to
go.  

The first thing I did was to read Martin Evening's Adobe Photoshop for
Photographers book cover-to-cover, with a pad of post-it stickies next
to me.  Every time I hit something in his book that I didn't know
about (or, at least, hadn't previously realized would be useful for
restoration) I put in a sticky.  When I was done I had the better part
of 100 stickies in the book! That's why recommend his book so strongly
in my introduction.  I got a LOT of good ideas for restoration tricks
that way.  

Nowadays, I'm kind of coasting.  Every so often I think of a new twist
on an old problem and I make note of it.  Also, readers regularly send
me questions and information about techniques and plug-ins that I
haven't tried yet.  So they're my guide to what new things I should be
learning now.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
========================================= 
  

    
(replying to #25)

OK, Diane, I pulled down that JPEG from the flicker site.  That's a
good scan; there is nothing blocked up at either end of the tonal
scale.  There isn't really all that much discoloration, so I threw away
the color information by using Photoshop's Channel Mixer to combine
50% red and 50% green to make a new monochrome image.  I got rid of the
yellow channel entirely, because it doesn't really contribute anything
useful (it's very dark) and it most strongly shows the few rust stains
in the photograph.  If there were a lot of rust stains, I'd use the
difference between the yellow channel and the other channels to create
a mask to select for them (much the same way I selected for the
speckles in the slide in Example 4, Chapter 10 (page 381).  But there
really aren't that many rust spots, so it isn't worth the effort. 
It'll be pretty fast just cleaning them out with the clone tool or
using the dust and scratches filter combined with the history brush (
Chapter 8, page 268).  

Next I used the Shadow/ Highlight tool with the shadow amount set to
25% and the midtone contrast kicked up by 15 points.  That brought out
a lot more detail in the fish tank and the shadows.

I cleaned up the small damage by running the Dust & Scratches filter
with a radius of 4, assigning that to the history brush, reverting to
the previous history state, and painting out the spots.  When I hit a
spot that was a little too big for the history brush, I used the clone
tool to fill in.  Only took five minutes to clean up all the little
stuff.

The lower missing pieces are easy to fix.  There's nothing of
importance there and it is mostly in the shadows, so just clone some
nearby content and paint it in.  

The missing bits along the top and right margins take a little more
work but not much.  There isn't much detail in there.  Just make sure
you're cloning lines up with the features that there are.  For
instance, when you're cloning the part that has the diagonal roof
beams, make sure you sample and click a long the direction of the beam,
so that when you paint in the cloned material, it will just extend the
beams nicely.  When you get to a vertical strut, click a new sample
and align it with the vertical. 10 minutes work, maybe, to clean it all
up. 

Which leaves the really nasty part. There's gonna be no way to fix
this that does involve some pain. And some completely imaginary
painting. My approach would be to get rid of as much of the crack as I
can in the background by cloning nearby parts of the walls and beams.
It's tedious work. You should work in from both sides of the tear. When
you get to the middle the tones will not match, so set the clone tool
back to 30% strength and paint over that area to blend the left and
right sides together. With some work (anywhere from 15 minutes to half
an hour) you'll be able to get rid of everything except the missing
forehead and wrist. You're just going to have to paint those in, using
the clone tool like a brush to sample from nearby areas to get the
tones and textures right, but you'll have to decide where the hairline
is and where the cuff actually starts, and there's just no way around
it. Like I said, painful. 

Once you've done all of that, you're going to want to increase the
contrast in the highlights to get better tonal separation and eliminate
the "copy print" look. Use a curve like the one in Figure 5-26. That
will get you an acceptable result. 

You might want to do a little dodging on the fish tank to bring out a
bit more of the details, but the Shadow/ Highlight tool pulled in most
of what you'd want.  

Finally, I'd run Focus Magic with a one pixel Blur Width to restore
fine detail.  It definitely perks up the photograph.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    
So....initial work on tones & desaturation; small fill-in; major
cloning if need be; then touch up with curves?

And I really should be able to clone in that entire nibble?   I would
not have thought to be so bold.

And is this a case where blacking out the missing area and torn edges
by covering with matte black paper is going to help the quality of the
scan, by avoiding any glare at the edges, or is that only applicable to
glass plates?
  

    
Well, y'know, that's the wonderful thing about digital.It's all
undoable. So what do you have to lose by trying to clone in the
background? Worst that happens is you throw away the revisions and
wasted some time. 

Based on your scan, I can't see any flare from the surrounding white
areas. But you could always run the experiment, scanning with and
without a black background sheet and see if it makes any difference.

pax / Ctein
  

    
Wow, that's a booger of a project to take on, Diane.  Me, I got some
red-shifted family photos I'm looking at.  Which seems pretty
straightforward.

That Martin Evening book is terrific; I've read the last few editions,
I think, and I learn something great stuff from each one.  

Your other big recommendation was for Katrin Eismann's book on
restoration, right? What makes that book a good combination with yours?
  

    
(replying to #30) 

Truthfully, I don't know of any other books devoted to restoration
besides mine and Katrin's.  Her overall approach is very different from
mine; she comes from a tradition of photo retouching; she's much more
involved with recreating and repainting a photograph, if necessary,
than I am.  That's not my forte - I'm into bit-twiddling and data
extraction.  When faced with wholly-missing material, like in Dianne's
photo, Katrin has a lot more to say on the subject than I do.  

I don't think she and I cover very much of the same material.  That's
partially by intent; I didn't see any value in having our books
duplicate each other, especially since I had a whole book's worth of
ideas of my own to write about.  

Besides, she's the established expert in the field; if my book
couldn't do more than echo hers, no one would've bought mine! 


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 


 
  

    
Speaking of areas of expertise, I did want to talk a bit more outside
the scope of this book, about dye transfer and the traditional ("wet"?)
darkoom.  You mentioned earlier that some people are trying to produce
their own materials; how's that working? And how big is the dye
transfer community, anyway?
  

    

Once you've answered David's question I have one for you.  In chapter 4 you
focus on scanning slides, negs, and prints.  I like how you demonstrate how
to use the scanner software to get a full tonal range brought into the
computer at the start, rather than optimize the histogram in Photoshop.

When you address the practical decisions people face in scanning images,
though, I was surprised to see your "Resolution Decisions" section not
explicitly address the relationship between total pixels, pixels-per-inch
(ppi) and desired output size. For isntance with a 3000 pixel image you can
get an 8x10 at 300 ppi.  But if you only want to print a 4x6, and if 300 ppi
will be good enough, then you don't need 3000 pixels.  You can scan a
smaller file and still get the desired 300 ppi at 4x6 without having to
resample down from, say, 3000 pixels. Were you thinking that people would
simply want to print the restoration at the same document dimensions as the
original?
  

    
(replying to #32) 

Well, the dye transfer community is much larger than I would've
expected, 13 years after Kodak pulled the the plug on us all.  Michael
Garelick has been keeping a list of people he's pretty sure are still
doing dye transfer printing, and it numbers about three dozen.  He says
there may be a couple of dozen more that he hasn't confirmed.  I find
this hard to believe, but he's the one doing the research, not me.  

Most people printing are doing so off of stockpiles of Kodak supplies.
 Every so often someone comes out of the woodwork and says something
like, " Hey, I used to do dye transfer printing and I've got some
supplies in my closet; does anybody want them?" They get snatched up
instantly.  That's happened to me several times; the last time was only
about two months ago. Remarkably enough, it turned out to be someone
who only lives two blocks away from me.  What are the odds?!  There are
a lot of people out there , it seems, with small stashes of stuff.  So
all of us have been able to stretch our supplies much further than we
expected.  

There are three major components to dye transfer printing: the matrix
film (essentially, the printing plates one uses), the dyes one soaks
the matrix film in, and receiving paper one transfers the dye to.  The
matrix film is the stuff that is hardest to come by.  One of the people
in our group, Jim Browning, commissioned Efke in Europe to do a
special run of matrix film.  That has been sold out; not clear at this
time if there will be another run.  Dyes are pretty readily available;
people are compounding their own all the time.  In all likelihood,
they're better than the dyes that Kodak was selling.  The receiving
paper is a little bit tricky, but people are having moderate success
taking regular photographic paper and treating it with chemicals to
make good dye transfer receiving paper.  

It's hard for me to believe, but I think dye transfer may survive as a
small cottage industry for some remarkably long time.  I certainly
don't think, any longer, that I will be the last of the dye transfer
printers.  If truth be told, I am not the least bit passionate about
the process.  I do it simply because it is in certain respects
unsurpassed by any other printing method.  But it's entirely possible
that digital printing will improve so much that it will make dye
transfer largely irrelevant to me, at which time I will happily give it
up.  I'm in it for the print quality, not because I'm in love with
making prints that way. 


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]]
  

    
(replying to #33) 

I didn't address that because I think output size is irrelevant to the
input process.  

First off, you are correct in guessing that most people simply want an
accurate reproduction of the original, at the same size as the
original.  More often than not, though, if they want a difference size,
they want a BIGGER print.  

Second, the sole goal of my advice is to insure that you capture all
of the information that is in the original print.  What you do with
that information when you print it out is another matter.  But the
essence of restoration is to throw away as little photographic
information as possible.

Third, tools like Focus Magic do a better job when they have extra
pixels to deal with. Even if the original photo only has 150-200 ppi
worth of fine detail in it, Focus Magic does a much better job of
sharpening it up from a 600 ppi scan than from a 300 ppi scan.  

Fourth, as I discuss in Chapters 4 , 7, and 8, there are times when
it's extremely important to have very high-resolution information about
the photograph, to make it easier to isolate cracks and scratches. 
Sometimes that can be a real killer.  

The photograph in figure 7-27 is a supreme example of how bad the pain
can be.  The original is about 5 by 7 in.  It's obvious from the
highly-faded condition that it needs to be scanned at 16-bit depth in
color to capture what little tonal information there is in the
highlights and to give spatial and color information that the repair
methods I use can grab onto.  On top of that, the scratches are so
terribly fine (and the photo is sufficiently sharp) that I need to scan
at 1200 ppi in order to sharply capture the scratches and be able to
deal with them without smearing out detail in the photograph.  

This is really the worst of all possible worlds; I wind up with a 250
MB file, and that's before I add any layers to it.  It's painful on a
system like mine that can only efficiently address 2 GB of memory.  I
haven't investigated what Photoshop CS3 will do under Mac OS X or
Windows Vista, but I'm sure hoping it breaks the 4 GB memory barrier. 
This is the kind of nightmare file that should really to have 8 GB of
RAM to work in.  

Fortunately, things are almost never this bad.  This is by far the
worst (in terms of file size) image I've had to deal with.  Most big
photographs aren't very sharp and most photographs that need
ultra-high-resolution scans are rather small.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    

Ctein, perhaps this is too basic a question for this discussion, but maybe
there are other people out there scratching their heads like I'm scratching
mine.

What IS dye transfer? Why would I want to have a dye transfer print? 
  

    
A very good question.  And while Ctein is thinking about that
question, I'll throw a related one in.  

It seems like in many ways the *capture* side of things is much higher
quality in digital, finally -- in a sort of absolute, higher
resolution, not taking into account the aesthetic qualities of
something like film grain way.

And as you said earlier, digital restoration has made traditional
restoration obsolete, simply because the quality is so much higher.

And we can get very good output digitally these days, but definitely
not as good as dye transfer (foreshadowing!)

Have you seen any digital printing technologies that you think will
finally get us to higher quality than dye transfer or other older
technologies?
  

    
<drift>

sorry i haven't been able to keep up with the conversation this week, it's
been a very busy week for me at work where our new foundation has been
presenting mental health awareness seminars at a bunch of local highschools,
and we had 800+ people to an event Monday night.

anyway, i've been been reading the book, and see there is a ton of thing
i'll be able to try out with my current software and hardware.

and looking forward to coming back here later and reading the previous 30+
responses after work.
  

    
(replying to No. 36) 

Well, for people who want the long story, read this article: 

http://ctein.com/dyetrans.htm 

For people who want a shorter version, dye transfer prints have a
longer tonal range and wider color gamut, by a substantial margin, than
any other printing process ever devised.  They are also extremely
permanent for photographic prints; their life on display is a fair
number of decades; their life in the dark is some 300 years.  A dye
transfer print, just sitting in a box on someone's shelf, made 50 years
ago will look the same today as it did then.  You can't say that for
most color photographs! 

Dye transfer printing also offers more degrees of control over how the
image looks than any other darkroom process.  Anyone here who was
played with digital printing with programs like Photoshop knows how
incredibly important sophisticated controls are; you can do amazing
things to make a photograph look just so when you can accurately
control tone and color placement.  Dye transfer has much more extensive
tools for doing that than any other photographic process.  

As the article explains, it's very time consuming and hardly anybody
does it any more; a few dozen people, and there are only five of us
that I know of on the whole planet who offer printing services to
others.  And you can bet that we get paid accordingly for our services
< greedy grin >.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    
(replying to # 37) 

OK-- two-part answer.  

On the photography side of things, it is solely about money.  Digital
has been able to match or surpass silver halide photography for many
years.  The only question has been how much money you were willing to
throw at the problem.  I designed my first modern digital camera back
in 1970 or 71.  My design looks pretty much  like what you get in a
digital camera today, but I knew back then it wouldn't be buildable for
about 20 years. It didn't hit the consumer price point I was aiming at
for 30. 

Circa 1987, I was bragging on the PhotoForum on CompuServe that I
could build a digital camera that would match anything you could do in
medium format photography, if people were willing to pay the price.  Of
course no one was; you were talking high six figures, maybe low seven
figures.  But if I'd had any takers, I could have provided them with a
camera.  

Stephen Johnson, with his Digital National Parks project, back in the
mid 90's, produced photographs with vastly better resolution that
captured a vastly longer luminance range than you could ever do with
film.  He was using equipment that cost in the high five figures. 

 On the printing side of things, it's about motivation.  There's
nothing exotic about dye transfer prints: they use fairly ordinary acid
fixing dyes that are absorbed by a simple photographic-type paper
(gelatin emulsion on top of fiber base).  The materials are not
remotely as high-tech as in inkjet printers; there are individuals out
there making their own papers and dyes.  

About 10 years ago, just for amusement, I took an empty ink cartridge
on my H-P DeskJet printer and refilled it with dye transfer dyes and
tried running a piece of dye transfer receiving paper through the
printer.  It worked! The print quality was lousy; nothing was optimized
for putting dye transfer dyes on dye transfer paper.  But there was no
reason you couldn't do that.  In fact if someone were to give me a
hundred thousand dollars or so right now, I could build myself an
inkjet printer that would use dye transfer materials and produce prints
that would run rings around anything I could make now, either on the
computer in the darkroom.  

But would anyone make such a printer? Not likely.  Dye transfer didn't
die just because of Kodak's mendacity.  99.9% of the printing needs in
the world are completely satisfied by ordinary Ektacolor-type print
quality. So the real question is whether the combination of marketing,
steady incremental improvements, and the "horsepower" race between
printer makers ever drives the quality to to the dye transfer level. It
certainly isn't a requirement for the printer business; it is whether
that level of extreme quality is enough of a selling point to drive the
manufacturers to develop and market the products.  

One would also need 16-bit printer drivers and much better profiles to
make proper use of such a printer, but those are solvable problems.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    
I saw some of Stephen Johnston's early prints at a conference in
Monterey in the mid-90's.  (I think Katrin Eisman spoke at that
conference, too, now that I think about it)  They were pretty intense;
my immediate response was that there was almost too much detail, that I
was used to the softening that even 8x10 film applies to reality,
whereas with these images every single leaf on every single plant was
crystal clear, and the overall experience was overwhelming.

It seems like if one could convince the printer manufacturers that
they'd sell a lot more $100 an ounce ink, they'd do anything. :-)

I certainly find myself that my black and white prints (with an Epson
r2400 on Hahnemuhle Fine Art Pearl) are as good, and often better, than
anything I could do in the traditional darkroom -- at a normal viewing
distance.  If I "pixel peep" I can see the inkjet spots.  And I
suspect that micro-fine detail does have an effect on the final viewing
experience.

After the first time I zoomed in to actual pixels on an image to dodge
a highlight on someone's eye, I was ready to be done with the wet
darkroom -- I just had to wait for output to catch up.
  

    
(reply to # 41) 

Stephen's photographs struck me as "odd" because of their tonal
qualities.  For those who haven't seen his work, he was working with a
scanning digital back that could easily reproduce a dozen stops of
luminance range with very high linearity.  His camera saw things a lot
more like the human eye sees them than like photographic film does.  As
a result, those photographs had light and color qualities that were
more reminiscent of paintings than of photographs, but they were
combined with undeniably-photographic fine detail.  It was a startling
and peculiar mix and it surprised me to discover how much I had linked
tone and detail to each other in paintings and photographs.  

While I must say that I'm surprised at how much I like matte-finish
inkjet prints, seeing as I hate that surface in the darkroom, I don't
think the Epson R 2400 B&W prints are terribly great.  There's still a
distinct lack of tonal separation compared to a darkroom print (call it
gradation, if you will) and the D-max of the inkjet print is far
inferior to what one can get in a darkroom.  

Compensating for that is the huge amount of control that one has over
tonal placement on the computer.  In aggregate, inkjet prints are
frequently better than darkroom prints precisely because of that
control.  

That's true in color also; my Epson 2200/2400 color prints are
frequently just as artistically satisfying as my dye transfer prints of
the same negatives.  Considering the technical inferiority of the
medium, that's quite astonishing.  It's a result of having such superb
control over the precise placement of color and tone when working
digitally.  Just as dye transfer trumps conventional printing in the
darkroom because of the light-years-better control tools, Photoshop
trumps dye transfer.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    
(Just a little aside - Stephen Johnson's presenting at the Apple Store 
in SF Monday... to see what Ctein's talking about in that post!
    http://www.apple.com/retail/sanfrancisco/week/20070513.html )
  

    
I'll take your word on the quality of inkjet BW vs. darkroom; I'm a
pretty good darkroom printer, but I'm not a master by any means.

This does bring up a question for me, though -- Is there anything
about traditional restoration work that gives a better result than
digital work?  Or is the control so much better in digital that it's no
contest?
  

    
I'm not sure I'm competent to address this question as I have never
done traditional restoration.  

I think I'm going to take to pass on this one.  Anybody in this
conference a traditional restorer? 


                                 pax / Ctein 
  

    

Thanks for your answer about dye transfer prints, Ctein. I hadn't realized 
the difference in photographic longevity between dye transfer and 
traditional print work.

I see on your web site (www.ctein.com) that you're going to a special
Guest of Honor a Baycon over Memorial Day weekend this year. 
Congratulations! Are you a bit nervous about it? Are you doing any
special preparation? Did you buy a new suit? 
  

    
New?
  

    
Yes, the longevity is a very significant factor.  In fact, until
conventional color prints improved so much starting around 1990, dye
transfer prints were by far the best print for display as well as dark
storage.  Other materials finally caught up for display and have done
even better, but dye transfer is still amazingly good in overall
keeping characteristics.  And it has the advantage of having been
around a very long time, so we actually have real-world data on prints
that are 50-70 years old.  None of this business of having to rely
solely on highly-accelerated tests to estimate print lifetimes.  

Nervous about being a Guest of Honor? Guffaw! I'm an artist; that
makes me an attention junkie.  Large numbers of people want to pay
attention to me and listen to me lecture on topics of my choosing and
look at my work? I can somehow live with that.  It's the fourth or
fifth time I've done this kind of thing, anyway.  Basically I get to be
the center of attention, give them dog and pony shows, and the
convention pays for my weekend and gives me a suite, to boot.  It's a
really tough life.  

I don't have to do very much prep for it.  What does need to be done
it is pretty much already in the can.  So I just go down and be myself.
 

I will go out and buy a suit if I'm ever on trial for murder.  For any
lesser offense, I can't see going to the expense < grin >.  


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]] 
  

    

heh, I had a feeling "a new suit" wasn't on your to-do list, Ctein.

In addition to your talent in the world of photography, you have a 
degree in physics, and you're a technical writer whose work covers
everything from electro-optics to computer languages. Which of your
many areas of expertise interest you the most? 
  

    
Oh, that's an easy one! The photography.  Considering that doing
almost anything else would earn me a better living than pursuing my
art, it's pretty clear to me where my heart lies.  

But for intellectual hobbies? Quantum mechanics, environmental
science, and planetary geology, in that order. 


                                 pax / Ctein 
[[ Please excuse any word-salad. ViaVoice in training! ]]