U of S | Mailing List Archive | alt-photo-process-l | RE: Two tricolor prints

RE: Two tricolor prints



Great, I was hoping somebody that does a living from photoshop would chime in. Thanks for suggesting such a simple solution. If I could only remember it a week from now.
Marek



Date: Wed, 26 Sep 2007 20:20:08 +0000
From: editor@magnachrom.com
Subject: Re: Two tricolor prints
To: alt-photo-process-l@usask.ca

no need to go to LAB to get the benefits of the underlying math. Try the following:
  • Make sure your data is 16-bit RGB
  • Create a "drastic" curve layer above the background layer to increase the contrast (using a classic "S curve for example)
  • Colors will probably change along with the contrast -- this is typical and usually not desired
  • Now change the curve layer properies from "Normal" to "Luminosity" -- saturation will remain unchanged from the original background layer while being independent of the contrast curve in the curve layer
The above delivers the same mathematical changes to the data as making mods to the Luminence layer in LAB.
 
Michael
 
 
 
-----Original Message-----
From: Marek Matusz [mailto:marekmatusz@hotmail.com]
Sent: Wednesday, September 26, 2007 02:38 PM
To: alt-photo-process-l@usask.ca
Subject: RE: Two tricolor prints

I am trying to digest all that.
Drastic curves, especially S shape can alter colors quite unpredictably. This is easily demonstrated with a jpeg (8bit)containing some pastel colors or greys. Just bring the curve dialog and play with it to see how all the sudden a pinkish color will chane to greensih. To avoid that work in 16 bit mode, appy curves in the LAB color mode, or separate channels and apply curves to separated channels. This is what I typically do as I wing my curves on the fly, basis the picture at hand. Of course starting and endpoint of the curve are fixed for the gum process, and printer settings
The other simpler reason might be that you simply have enough density to block it in one negative used to print blue, but by applying curve you have taken the density in the negative away and now you are able to print the unwanted blue.
This is something that I watch for making separations. Are there places in the print that require clean two color mixes and is the color separation negative of the third primary sufficently dense to block the third color layer.
 
Marek

> Date: Wed, 26 Sep 2007 08:47:58 -0700
> From: kthayer@pacifier.com
> Subject: Re: Two tricolor prints
> To: alt-photo-process-l@usask.ca
>
> On Sep 26, 2007, at 6:25 AM, Marek Matusz wrote:
> > Katharine,
> > A great example that illustrates the point that I was trying to
> > make to Charles. Your muddy and dark print is a result of overall
> > low contrast of your negatives,
>
> Yes, but the overall low contrast is to some extent an inevitable
> result of setting the density of the negative to match the scale of
> the gum, which was part of my point. There's so much difference to
> be made up that it's very difficult to find a curve that will replace
> the contrast that you've lost, if that makes sense. If you have
> any slope through the midtones, then you sacrifice separation in the
> highlights or shadows or both. If you get separation in the
> highlights and shadows, then your midtones have to be flat. So
> far, I haven't found a curve that makes a better gum print than a
> diagonal line does (for my printer, light, negative media, etc) but I
> haven't given up yet.
>
> I would think maybe I was missing something (which is entirely
> possible, since while I'm generally thought of as a fairly
> intelligent person and mastered statistics with no trouble, this
> stuff just makes no sense to me at all) except that a number of
> people have thoughtfully sent me calibration charts and curves to
> show me how I "should" be calibrating my curves for gum, and these
> have shown the same difficulty I'm struggling with. For one thing
> they have been made with underpigmented mixes, which I find
> interesting and somewhat inexplicable, but the main thing is that the
> post-curve step tablets seem to follow the same pattern: say in a
> 101-step chart, the steps from 1-10 or so show nice gradation, and
> the steps from 85 or 90 to 100 show good gradation, but all the steps
> from 15-85 are the same tone. And the curve reflects that exactly: a
> slope in the highlights and shadows, but horizontal through the
> midtones. Or, they have to lighten the highlights too much in order
> to leave room for a slope in the midtones. No matter what you do,
> something has to give. So I'm still struggling with that
> fundamental difficulty with making curves for gum.
>
> But the problem here isn't so much with the curves as with the color;
> it's failing to block light where it should block it. This makes no
> sense to me because this color works for other kinds of prints; what
> should be paper white is paper white in all the other prints I've
> made using this color. These separations are the first place where
> I've seen this, gum printing where it shouldn't. It's really
> obvious when I print just the cyan layer, for example, using the two
> kinds of separation. The simple inversion separation keeps the blue
> off the poppies, leaving clean vibrant orange; the calibrated
> separation prints blue on the poppies, muddying the orange tone.
> There's no way to get rid of it but to brush it off, which isn't a
> kind of practice I want to get into. There's no logical reason for
> separations to be any different from other kinds of negative files,
> so I'm at a loss, but obviously if I'm going to use this calibration
> method for separations for tricolor gum, I'll need to go back to the
> drawing board and re-calibrate. I'm just really tired of fussing
> with the whole thing, was my point, and not yet convinced that it can
> yield good return on the investment of time, especially for tricolor
> separations.
>
>
> > so I would suggest correcting the curves not altering pigments.
>
> Hmm, a couple of different things have got confused here, it looks
> like. I daresay I created that confusion by connecting this to a
> discussion about altering pigment to relieve darkness in a print, but
> altering pigments isn't what I'm suggesting here, and thanks for
> pointing out the confusion. The connection to the other thing wasn't
> pigment alteration, but a more general connection: the difficulty of
> relying on individually calibrated layers for tricolor, when the
> final result is more a function of the interaction between the layers
> than of the calibration of the layers. If you're going to calibrate
> for tricolor, what you really need is a sort of three-dimensional
> calibration, which makes my head ache to think of.
>
> No, I put the blame for this problem squarely on the negatives;
> there's nothing wrong with the pigments, as is evident from the other
> print. But it's interesting that you characterize this as a "dark"
> print; to my eye, I wouldn't call this dark at all. I called it
> "light and muddy" in my initial post about it, and I think that's an
> accurate evaluation. A quick check of the histogram bears me out;
> the median value of the left print is 118; the median value of the
> right print is 79; in other words, the left print is lighter by 15%.
>
> My suggestion to reduce pigments for a dark print was intended for a
> very dark print, where the exposures for the individual layers are
> right but the tricolor is way too dark, as a result of
> overpigmentation of the layers. This print isn't remotely as dark
> as a print I would make that suggestion for.
>
> The print with the "punch" is printed from the same kind of simple
> separations I've always used for tricolor gum; easy, quick, you're
> there, which was another part of my point; I didn't do anything
> special to achieve that effect; it's the same effect I get in every
> tricolor I've ever made, except for the ones where I deliberately
> muted the colors by adjusting the pigment. I'm hoping this example
> will help convince people who are intimidated by the whole curves
> thing not to be afraid to just go ahead and print. I'd say to them,
> you'll learn a whole lot more from printing tricolor gum than you
> ever will from printing calibration charts.
>
>
> > Thanks for a great teaching example.
>
> You're very welcome; I'm awfully glad to see the list reviving and
> some actual discussion going on. And I take your points; yes,
> probably more fiddling in that direction might yield a better print,
> but I'm at the point of diminishing returns. I'd rather be printing
> gum.
>
> Katharine
>
>
>
>
>
>
> >
> > Marek
> > > Date: Tue, 25 Sep 2007 19:17:22 -0700
> > > From: kthayer@pacifier.com
> > > Subject: Two tricolor prints
> > > To: alt-photo-process-l@usask.ca
> > >
> > >
> > >
> > > On Sep 25, 2007, at 4:54 PM, Katharine Thayer wrote:
> > > >
> > > > I'm struggling with a similar issue, but finding that I have a
> > > > different problem than many people have reported. After happily
> > > > printing tricolor gum for decades by simply inverting the channels
> > > > and printing greyscale separations, I've been experimenting with a
> > > > more "sophisticated" approach, with less than satisfactory
> > > > results. But the problem isn't that the print is too dark when the
> > > > colors are layered over each other, it's that it's too light and
> > > > too muddy, compared to the same image printed with the same
> > > > emulsion by my old habitual method of simply inverting channels
> > >
> > > Here's a comparison of an image printed from the colorized and
> > > calibrated separations, and the same image printed from separations
> > > made by simply inverting channels and printing without further
> > > manipulation . The jpeg is rather small because I was intending to
> > > append it to a post on another forum, but since the question of
> > > calibrating tricolors came up here in the meantime, decided to just
> > > put it up on my site; at any rate it's not necessary to see detail;
> > > the difference in the vividness of the color, which is the issue of
> > > note here, should be quite apparent even in the small images.
> > >
> > > I named the page rather cockily "If it ain't broke..." but it is a
> > > little broke in fact; the greyscale separations can make a print
> > > that's too contrasty, which is the reason I decided to see if I
> > could
> > > get an improvement with colors and curves. But so far, I'll take the
> > > straight inversions even with blown highlights. This isn't just an
> > > anomaly; I did several prints with different pigments and
> > proportions
> > > and got the same cloudiness on all of them.
> > >
> > > I've been thinking for a while of switching to Prussian blue for
> > > tricolors, and these are my first prints with Prussian; it will
> > > definitely be my blue of choice for tricolor from now on. I'm still
> > > playing with pigments and proportions to balance the Prussian, but
> > > the combination I used here gives a fairly close approximation to
> > > the original colors.
> > >
> > > http://www.pacifier.com/~kthayer/html/tricolorcomp.html
> > >
> > > Katharine
> > >
> > >
> >
> >
> > Get news, entertainment and everything you care about at Live.com.
> > Check it out!
>



Connect to the next generation of MSN Messenger  Get it now!
Contacts
(Preferred e-mail)


Discover the new Windows Vista Learn more!